5 * This file has definitions for some important file table
9 #include <linux/limits.h>
10 #include <linux/ioctl.h>
11 #include <linux/types.h>
14 * It's silly to have NR_OPEN bigger than NR_FILE, but you can change
15 * the file limit at runtime and only root can increase the per-process
16 * nr_file rlimit, so it's safe to set up a ridiculously high absolute
17 * upper limit on files-per-process.
19 * Some programs (notably those using select()) may have to be
20 * recompiled to take full advantage of the new limits..
23 /* Fixed constants first: */
25 #define INR_OPEN_CUR 1024 /* Initial setting for nfile rlimits */
26 #define INR_OPEN_MAX 4096 /* Hard limit for nfile rlimits */
28 #define BLOCK_SIZE_BITS 10
29 #define BLOCK_SIZE (1<<BLOCK_SIZE_BITS)
31 #define SEEK_SET 0 /* seek relative to beginning of file */
32 #define SEEK_CUR 1 /* seek relative to current file position */
33 #define SEEK_END 2 /* seek relative to end of file */
34 #define SEEK_DATA 3 /* seek to the next data */
35 #define SEEK_HOLE 4 /* seek to the next hole */
36 #define SEEK_MAX SEEK_HOLE
44 /* And dynamically-tunable limits and defaults: */
45 struct files_stat_struct {
46 unsigned long nr_files; /* read only */
47 unsigned long nr_free_files; /* read only */
48 unsigned long max_files; /* tunable */
51 struct inodes_stat_t {
54 int dummy[5]; /* padding for sysctl ABI compatibility */
58 #define NR_FILE 8192 /* this can well be larger on a larger system */
60 #define MAY_EXEC 0x00000001
61 #define MAY_WRITE 0x00000002
62 #define MAY_READ 0x00000004
63 #define MAY_APPEND 0x00000008
64 #define MAY_ACCESS 0x00000010
65 #define MAY_OPEN 0x00000020
66 #define MAY_CHDIR 0x00000040
67 /* called from RCU mode, don't block */
68 #define MAY_NOT_BLOCK 0x00000080
71 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
72 * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
75 /* file is open for reading */
76 #define FMODE_READ ((__force fmode_t)0x1)
77 /* file is open for writing */
78 #define FMODE_WRITE ((__force fmode_t)0x2)
79 /* file is seekable */
80 #define FMODE_LSEEK ((__force fmode_t)0x4)
81 /* file can be accessed using pread */
82 #define FMODE_PREAD ((__force fmode_t)0x8)
83 /* file can be accessed using pwrite */
84 #define FMODE_PWRITE ((__force fmode_t)0x10)
85 /* File is opened for execution with sys_execve / sys_uselib */
86 #define FMODE_EXEC ((__force fmode_t)0x20)
87 /* File is opened with O_NDELAY (only set for block devices) */
88 #define FMODE_NDELAY ((__force fmode_t)0x40)
89 /* File is opened with O_EXCL (only set for block devices) */
90 #define FMODE_EXCL ((__force fmode_t)0x80)
91 /* File is opened using open(.., 3, ..) and is writeable only for ioctls
92 (specialy hack for floppy.c) */
93 #define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
94 /* 32bit hashes as llseek() offset (for directories) */
95 #define FMODE_32BITHASH ((__force fmode_t)0x200)
96 /* 64bit hashes as llseek() offset (for directories) */
97 #define FMODE_64BITHASH ((__force fmode_t)0x400)
100 * Don't update ctime and mtime.
102 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
103 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
105 #define FMODE_NOCMTIME ((__force fmode_t)0x800)
107 /* Expect random access pattern */
108 #define FMODE_RANDOM ((__force fmode_t)0x1000)
110 /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
111 #define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
113 /* File is opened with O_PATH; almost nothing can be done with it */
114 #define FMODE_PATH ((__force fmode_t)0x4000)
116 /* File was opened by fanotify and shouldn't generate fanotify events */
117 #define FMODE_NONOTIFY ((__force fmode_t)0x1000000)
120 * Flag for rw_copy_check_uvector and compat_rw_copy_check_uvector
121 * that indicates that they should check the contents of the iovec are
122 * valid, but not check the memory that the iovec elements
125 #define CHECK_IOVEC_ONLY -1
131 /* public flags for file_system_type */
132 #define FS_REQUIRES_DEV 1
133 #define FS_BINARY_MOUNTDATA 2
134 #define FS_HAS_SUBTYPE 4
135 #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
136 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move()
137 * during rename() internally.
141 * These are the fs-independent mount-flags: up to 32 flags are supported
143 #define MS_RDONLY 1 /* Mount read-only */
144 #define MS_NOSUID 2 /* Ignore suid and sgid bits */
145 #define MS_NODEV 4 /* Disallow access to device special files */
146 #define MS_NOEXEC 8 /* Disallow program execution */
147 #define MS_SYNCHRONOUS 16 /* Writes are synced at once */
148 #define MS_REMOUNT 32 /* Alter flags of a mounted FS */
149 #define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
150 #define MS_DIRSYNC 128 /* Directory modifications are synchronous */
151 #define MS_NOATIME 1024 /* Do not update access times. */
152 #define MS_NODIRATIME 2048 /* Do not update directory access times */
156 #define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
157 MS_VERBOSE is deprecated. */
158 #define MS_SILENT 32768
159 #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
160 #define MS_UNBINDABLE (1<<17) /* change to unbindable */
161 #define MS_PRIVATE (1<<18) /* change to private */
162 #define MS_SLAVE (1<<19) /* change to slave */
163 #define MS_SHARED (1<<20) /* change to shared */
164 #define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */
165 #define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
166 #define MS_I_VERSION (1<<23) /* Update inode I_version field */
167 #define MS_STRICTATIME (1<<24) /* Always perform atime updates */
168 #define MS_NOSEC (1<<28)
169 #define MS_BORN (1<<29)
170 #define MS_ACTIVE (1<<30)
171 #define MS_NOUSER (1<<31)
174 * Superblock flags that can be altered by MS_REMOUNT
176 #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK|MS_I_VERSION)
179 * Old magic mount flag and mask
181 #define MS_MGC_VAL 0xC0ED0000
182 #define MS_MGC_MSK 0xffff0000
184 /* Inode flags - they have nothing to superblock flags now */
186 #define S_SYNC 1 /* Writes are synced at once */
187 #define S_NOATIME 2 /* Do not update access times */
188 #define S_APPEND 4 /* Append-only file */
189 #define S_IMMUTABLE 8 /* Immutable file */
190 #define S_DEAD 16 /* removed, but still open directory */
191 #define S_NOQUOTA 32 /* Inode is not counted to quota */
192 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
193 #define S_NOCMTIME 128 /* Do not update file c/mtime */
194 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
195 #define S_PRIVATE 512 /* Inode is fs-internal */
196 #define S_IMA 1024 /* Inode has an associated IMA struct */
197 #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
198 #define S_NOSEC 4096 /* no suid or xattr security attributes */
201 * Note that nosuid etc flags are inode-specific: setting some file-system
202 * flags just means all the inodes inherit those flags by default. It might be
203 * possible to override it selectively if you really wanted to with some
204 * ioctl() that is not currently implemented.
206 * Exception: MS_RDONLY is always applied to the entire file system.
208 * Unfortunately, it is possible to change a filesystems flags with it mounted
209 * with files in use. This means that all of the inodes will not have their
210 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
211 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
213 #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
215 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
216 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
217 ((inode)->i_flags & S_SYNC))
218 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
219 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
220 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
221 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
222 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
224 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
225 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
226 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
227 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
229 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
230 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
231 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
232 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
233 #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
234 #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
235 #define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
237 /* the read-only stuff doesn't really belong here, but any other place is
238 probably as bad and I don't want to create yet another include file. */
240 #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
241 #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
242 #define BLKRRPART _IO(0x12,95) /* re-read partition table */
243 #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
244 #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
245 #define BLKRASET _IO(0x12,98) /* set read ahead for block device */
246 #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
247 #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
248 #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
249 #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
250 #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
251 #define BLKSSZGET _IO(0x12,104)/* get block device sector size */
253 #define BLKPG _IO(0x12,105)/* See blkpg.h */
255 /* Some people are morons. Do not use sizeof! */
257 #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
258 #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
259 /* This was here just to show that the number is taken -
260 probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
262 /* A jump here: 108-111 have been used for various private purposes. */
263 #define BLKBSZGET _IOR(0x12,112,size_t)
264 #define BLKBSZSET _IOW(0x12,113,size_t)
265 #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
266 #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
267 #define BLKTRACESTART _IO(0x12,116)
268 #define BLKTRACESTOP _IO(0x12,117)
269 #define BLKTRACETEARDOWN _IO(0x12,118)
270 #define BLKDISCARD _IO(0x12,119)
271 #define BLKIOMIN _IO(0x12,120)
272 #define BLKIOOPT _IO(0x12,121)
273 #define BLKALIGNOFF _IO(0x12,122)
274 #define BLKPBSZGET _IO(0x12,123)
275 #define BLKDISCARDZEROES _IO(0x12,124)
276 #define BLKSECDISCARD _IO(0x12,125)
277 #define BLKROTATIONAL _IO(0x12,126)
278 #define BLKZEROOUT _IO(0x12,127)
280 #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
281 #define FIBMAP _IO(0x00,1) /* bmap access */
282 #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
283 #define FIFREEZE _IOWR('X', 119, int) /* Freeze */
284 #define FITHAW _IOWR('X', 120, int) /* Thaw */
285 #define FITRIM _IOWR('X', 121, struct fstrim_range) /* Trim */
287 #define FS_IOC_GETFLAGS _IOR('f', 1, long)
288 #define FS_IOC_SETFLAGS _IOW('f', 2, long)
289 #define FS_IOC_GETVERSION _IOR('v', 1, long)
290 #define FS_IOC_SETVERSION _IOW('v', 2, long)
291 #define FS_IOC_FIEMAP _IOWR('f', 11, struct fiemap)
292 #define FS_IOC32_GETFLAGS _IOR('f', 1, int)
293 #define FS_IOC32_SETFLAGS _IOW('f', 2, int)
294 #define FS_IOC32_GETVERSION _IOR('v', 1, int)
295 #define FS_IOC32_SETVERSION _IOW('v', 2, int)
298 * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS)
300 #define FS_SECRM_FL 0x00000001 /* Secure deletion */
301 #define FS_UNRM_FL 0x00000002 /* Undelete */
302 #define FS_COMPR_FL 0x00000004 /* Compress file */
303 #define FS_SYNC_FL 0x00000008 /* Synchronous updates */
304 #define FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
305 #define FS_APPEND_FL 0x00000020 /* writes to file may only append */
306 #define FS_NODUMP_FL 0x00000040 /* do not dump file */
307 #define FS_NOATIME_FL 0x00000080 /* do not update atime */
308 /* Reserved for compression usage... */
309 #define FS_DIRTY_FL 0x00000100
310 #define FS_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
311 #define FS_NOCOMP_FL 0x00000400 /* Don't compress */
312 #define FS_ECOMPR_FL 0x00000800 /* Compression error */
313 /* End compression flags --- maybe not all used */
314 #define FS_BTREE_FL 0x00001000 /* btree format dir */
315 #define FS_INDEX_FL 0x00001000 /* hash-indexed directory */
316 #define FS_IMAGIC_FL 0x00002000 /* AFS directory */
317 #define FS_JOURNAL_DATA_FL 0x00004000 /* Reserved for ext3 */
318 #define FS_NOTAIL_FL 0x00008000 /* file tail should not be merged */
319 #define FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
320 #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
321 #define FS_EXTENT_FL 0x00080000 /* Extents */
322 #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */
323 #define FS_NOCOW_FL 0x00800000 /* Do not cow file */
324 #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */
326 #define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
327 #define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
330 #define SYNC_FILE_RANGE_WAIT_BEFORE 1
331 #define SYNC_FILE_RANGE_WRITE 2
332 #define SYNC_FILE_RANGE_WAIT_AFTER 4
336 #include <linux/linkage.h>
337 #include <linux/wait.h>
338 #include <linux/kdev_t.h>
339 #include <linux/dcache.h>
340 #include <linux/path.h>
341 #include <linux/stat.h>
342 #include <linux/cache.h>
343 #include <linux/list.h>
344 #include <linux/radix-tree.h>
345 #include <linux/rbtree.h>
346 #include <linux/init.h>
347 #include <linux/pid.h>
348 #include <linux/bug.h>
349 #include <linux/mutex.h>
350 #include <linux/capability.h>
351 #include <linux/semaphore.h>
352 #include <linux/fiemap.h>
353 #include <linux/rculist_bl.h>
354 #include <linux/atomic.h>
355 #include <linux/shrinker.h>
356 #include <linux/migrate_mode.h>
357 #include <linux/uidgid.h>
358 #include <linux/lockdep.h>
359 #include <linux/percpu-rwsem.h>
360 #include <linux/blk_types.h>
362 #include <asm/byteorder.h>
364 struct export_operations;
370 struct pipe_inode_info;
371 struct poll_table_struct;
373 struct vm_area_struct;
376 struct swap_info_struct;
378 extern void __init inode_init(void);
379 extern void __init inode_init_early(void);
380 extern void __init files_init(unsigned long);
382 extern struct files_stat_struct files_stat;
383 extern unsigned long get_max_files(void);
384 extern int sysctl_nr_open;
385 extern struct inodes_stat_t inodes_stat;
386 extern int leases_enable, lease_break_time;
387 extern int sysctl_protected_symlinks;
388 extern int sysctl_protected_hardlinks;
391 typedef int (get_block_t)(struct inode *inode, sector_t iblock,
392 struct buffer_head *bh_result, int create);
393 typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
394 ssize_t bytes, void *private, int ret,
398 * The below are the various read and write types that we support. Some of
399 * them include behavioral modifiers that send information down to the
400 * block layer and IO scheduler. Terminology:
402 * The block layer uses device plugging to defer IO a little bit, in
403 * the hope that we will see more IO very shortly. This increases
404 * coalescing of adjacent IO and thus reduces the number of IOs we
405 * have to send to the device. It also allows for better queuing,
406 * if the IO isn't mergeable. If the caller is going to be waiting
407 * for the IO, then he must ensure that the device is unplugged so
408 * that the IO is dispatched to the driver.
410 * All IO is handled async in Linux. This is fine for background
411 * writes, but for reads or writes that someone waits for completion
412 * on, we want to notify the block layer and IO scheduler so that they
413 * know about it. That allows them to make better scheduling
414 * decisions. So when the below references 'sync' and 'async', it
415 * is referencing this priority hint.
417 * With that in mind, the available types are:
419 * READ A normal read operation. Device will be plugged.
420 * READ_SYNC A synchronous read. Device is not plugged, caller can
421 * immediately wait on this read without caring about
423 * READA Used for read-ahead operations. Lower priority, and the
424 * block layer could (in theory) choose to ignore this
425 * request if it runs into resource problems.
426 * WRITE A normal async write. Device will be plugged.
427 * WRITE_SYNC Synchronous write. Identical to WRITE, but passes down
428 * the hint that someone will be waiting on this IO
429 * shortly. The write equivalent of READ_SYNC.
430 * WRITE_ODIRECT Special case write for O_DIRECT only.
431 * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
432 * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
433 * non-volatile media on completion.
434 * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
435 * by a cache flush and data is guaranteed to be on
436 * non-volatile media on completion.
439 #define RW_MASK REQ_WRITE
440 #define RWA_MASK REQ_RAHEAD
443 #define WRITE RW_MASK
444 #define READA RWA_MASK
445 #define KERNEL_READ (READ|REQ_KERNEL)
446 #define KERNEL_WRITE (WRITE|REQ_KERNEL)
448 #define READ_SYNC (READ | REQ_SYNC)
449 #define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE)
450 #define WRITE_ODIRECT (WRITE | REQ_SYNC)
451 #define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
452 #define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
453 #define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
456 * Attribute flags. These should be or-ed together to figure out what
459 #define ATTR_MODE (1 << 0)
460 #define ATTR_UID (1 << 1)
461 #define ATTR_GID (1 << 2)
462 #define ATTR_SIZE (1 << 3)
463 #define ATTR_ATIME (1 << 4)
464 #define ATTR_MTIME (1 << 5)
465 #define ATTR_CTIME (1 << 6)
466 #define ATTR_ATIME_SET (1 << 7)
467 #define ATTR_MTIME_SET (1 << 8)
468 #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
469 #define ATTR_ATTR_FLAG (1 << 10)
470 #define ATTR_KILL_SUID (1 << 11)
471 #define ATTR_KILL_SGID (1 << 12)
472 #define ATTR_FILE (1 << 13)
473 #define ATTR_KILL_PRIV (1 << 14)
474 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
475 #define ATTR_TIMES_SET (1 << 16)
478 * This is the Inode Attributes structure, used for notify_change(). It
479 * uses the above definitions as flags, to know which values have changed.
480 * Also, in this manner, a Filesystem can look at only the values it cares
481 * about. Basically, these are the attributes that the VFS layer can
482 * request to change from the FS layer.
484 * Derek Atkins <warlord@MIT.EDU> 94-10-20
487 unsigned int ia_valid;
492 struct timespec ia_atime;
493 struct timespec ia_mtime;
494 struct timespec ia_ctime;
497 * Not an attribute, but an auxiliary info for filesystems wanting to
498 * implement an ftruncate() like method. NOTE: filesystem should
499 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
501 struct file *ia_file;
505 * Includes for diskquotas.
507 #include <linux/quota.h>
510 * enum positive_aop_returns - aop return codes with specific semantics
512 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
513 * completed, that the page is still locked, and
514 * should be considered active. The VM uses this hint
515 * to return the page to the active list -- it won't
516 * be a candidate for writeback again in the near
517 * future. Other callers must be careful to unlock
518 * the page if they get this return. Returned by
521 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
522 * unlocked it and the page might have been truncated.
523 * The caller should back up to acquiring a new page and
524 * trying again. The aop will be taking reasonable
525 * precautions not to livelock. If the caller held a page
526 * reference, it should drop it before retrying. Returned
529 * address_space_operation functions return these large constants to indicate
530 * special semantics to the caller. These are much larger than the bytes in a
531 * page to allow for functions that return the number of bytes operated on in a
535 enum positive_aop_returns {
536 AOP_WRITEPAGE_ACTIVATE = 0x80000,
537 AOP_TRUNCATED_PAGE = 0x80001,
540 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
541 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
542 #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
543 * helper code (eg buffer layer)
544 * to clear GFP_FS from alloc */
547 * oh the beauties of C type declarations.
550 struct address_space;
551 struct writeback_control;
554 const struct iovec *iov;
555 unsigned long nr_segs;
560 size_t iov_iter_copy_from_user_atomic(struct page *page,
561 struct iov_iter *i, unsigned long offset, size_t bytes);
562 size_t iov_iter_copy_from_user(struct page *page,
563 struct iov_iter *i, unsigned long offset, size_t bytes);
564 void iov_iter_advance(struct iov_iter *i, size_t bytes);
565 int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes);
566 size_t iov_iter_single_seg_count(struct iov_iter *i);
568 static inline void iov_iter_init(struct iov_iter *i,
569 const struct iovec *iov, unsigned long nr_segs,
570 size_t count, size_t written)
573 i->nr_segs = nr_segs;
575 i->count = count + written;
577 iov_iter_advance(i, written);
580 static inline size_t iov_iter_count(struct iov_iter *i)
586 * "descriptor" for what we're up to with a read.
587 * This allows us to use the same read code yet
588 * have multiple different users of the data that
589 * we read from a file.
591 * The simplest case just copies the data to user
604 typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
605 unsigned long, unsigned long);
607 struct address_space_operations {
608 int (*writepage)(struct page *page, struct writeback_control *wbc);
609 int (*readpage)(struct file *, struct page *);
611 /* Write back some dirty pages from this mapping. */
612 int (*writepages)(struct address_space *, struct writeback_control *);
614 /* Set a page dirty. Return true if this dirtied it */
615 int (*set_page_dirty)(struct page *page);
617 int (*readpages)(struct file *filp, struct address_space *mapping,
618 struct list_head *pages, unsigned nr_pages);
620 int (*write_begin)(struct file *, struct address_space *mapping,
621 loff_t pos, unsigned len, unsigned flags,
622 struct page **pagep, void **fsdata);
623 int (*write_end)(struct file *, struct address_space *mapping,
624 loff_t pos, unsigned len, unsigned copied,
625 struct page *page, void *fsdata);
627 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
628 sector_t (*bmap)(struct address_space *, sector_t);
629 void (*invalidatepage) (struct page *, unsigned long);
630 int (*releasepage) (struct page *, gfp_t);
631 void (*freepage)(struct page *);
632 ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
633 loff_t offset, unsigned long nr_segs);
634 int (*get_xip_mem)(struct address_space *, pgoff_t, int,
635 void **, unsigned long *);
637 * migrate the contents of a page to the specified target. If sync
638 * is false, it must not block.
640 int (*migratepage) (struct address_space *,
641 struct page *, struct page *, enum migrate_mode);
642 int (*launder_page) (struct page *);
643 int (*is_partially_uptodate) (struct page *, read_descriptor_t *,
645 int (*error_remove_page)(struct address_space *, struct page *);
647 /* swapfile support */
648 int (*swap_activate)(struct swap_info_struct *sis, struct file *file,
650 void (*swap_deactivate)(struct file *file);
653 extern const struct address_space_operations empty_aops;
656 * pagecache_write_begin/pagecache_write_end must be used by general code
657 * to write into the pagecache.
659 int pagecache_write_begin(struct file *, struct address_space *mapping,
660 loff_t pos, unsigned len, unsigned flags,
661 struct page **pagep, void **fsdata);
663 int pagecache_write_end(struct file *, struct address_space *mapping,
664 loff_t pos, unsigned len, unsigned copied,
665 struct page *page, void *fsdata);
667 struct backing_dev_info;
668 struct address_space {
669 struct inode *host; /* owner: inode, block_device */
670 struct radix_tree_root page_tree; /* radix tree of all pages */
671 spinlock_t tree_lock; /* and lock protecting it */
672 unsigned int i_mmap_writable;/* count VM_SHARED mappings */
673 struct rb_root i_mmap; /* tree of private and shared mappings */
674 struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
675 struct mutex i_mmap_mutex; /* protect tree, count, list */
676 /* Protected by tree_lock together with the radix tree */
677 unsigned long nrpages; /* number of total pages */
678 pgoff_t writeback_index;/* writeback starts here */
679 const struct address_space_operations *a_ops; /* methods */
680 unsigned long flags; /* error bits/gfp mask */
681 struct backing_dev_info *backing_dev_info; /* device readahead, etc */
682 spinlock_t private_lock; /* for use by the address_space */
683 struct list_head private_list; /* ditto */
684 struct address_space *assoc_mapping; /* ditto */
685 } __attribute__((aligned(sizeof(long))));
687 * On most architectures that alignment is already the case; but
688 * must be enforced here for CRIS, to let the least significant bit
689 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
691 struct request_queue;
693 struct block_device {
694 dev_t bd_dev; /* not a kdev_t - it's a search key */
696 struct inode * bd_inode; /* will die */
697 struct super_block * bd_super;
698 struct mutex bd_mutex; /* open/close mutex */
699 struct list_head bd_inodes;
703 bool bd_write_holder;
705 struct list_head bd_holder_disks;
707 struct block_device * bd_contains;
708 unsigned bd_block_size;
709 struct hd_struct * bd_part;
710 /* number of times partitions within this device have been opened. */
711 unsigned bd_part_count;
713 struct gendisk * bd_disk;
714 struct request_queue * bd_queue;
715 struct list_head bd_list;
717 * Private data. You must have bd_claim'ed the block_device
718 * to use this. NOTE: bd_claim allows an owner to claim
719 * the same device multiple times, the owner must take special
720 * care to not mess up bd_private for that case.
722 unsigned long bd_private;
724 /* The counter of freeze processes */
725 int bd_fsfreeze_count;
726 /* Mutex for freeze */
727 struct mutex bd_fsfreeze_mutex;
728 /* A semaphore that prevents I/O while block size is being changed */
729 struct percpu_rw_semaphore bd_block_size_semaphore;
733 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
736 #define PAGECACHE_TAG_DIRTY 0
737 #define PAGECACHE_TAG_WRITEBACK 1
738 #define PAGECACHE_TAG_TOWRITE 2
740 int mapping_tagged(struct address_space *mapping, int tag);
743 * Might pages of this file be mapped into userspace?
745 static inline int mapping_mapped(struct address_space *mapping)
747 return !RB_EMPTY_ROOT(&mapping->i_mmap) ||
748 !list_empty(&mapping->i_mmap_nonlinear);
752 * Might pages of this file have been modified in userspace?
753 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
754 * marks vma as VM_SHARED if it is shared, and the file was opened for
755 * writing i.e. vma may be mprotected writable even if now readonly.
757 static inline int mapping_writably_mapped(struct address_space *mapping)
759 return mapping->i_mmap_writable != 0;
763 * Use sequence counter to get consistent i_size on 32-bit processors.
765 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
766 #include <linux/seqlock.h>
767 #define __NEED_I_SIZE_ORDERED
768 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
770 #define i_size_ordered_init(inode) do { } while (0)
774 #define ACL_NOT_CACHED ((void *)(-1))
776 #define IOP_FASTPERM 0x0001
777 #define IOP_LOOKUP 0x0002
778 #define IOP_NOFOLLOW 0x0004
781 * Keep mostly read-only and often accessed (especially for
782 * the RCU path lookup and 'stat' data) fields at the beginning
783 * of the 'struct inode'
787 unsigned short i_opflags;
790 unsigned int i_flags;
792 #ifdef CONFIG_FS_POSIX_ACL
793 struct posix_acl *i_acl;
794 struct posix_acl *i_default_acl;
797 const struct inode_operations *i_op;
798 struct super_block *i_sb;
799 struct address_space *i_mapping;
801 #ifdef CONFIG_SECURITY
805 /* Stat data, not accessed from path walking */
808 * Filesystems may only read i_nlink directly. They shall use the
809 * following functions for modification:
811 * (set|clear|inc|drop)_nlink
812 * inode_(inc|dec)_link_count
815 const unsigned int i_nlink;
816 unsigned int __i_nlink;
820 struct timespec i_atime;
821 struct timespec i_mtime;
822 struct timespec i_ctime;
823 spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
824 unsigned short i_bytes;
825 unsigned int i_blkbits;
828 #ifdef __NEED_I_SIZE_ORDERED
829 seqcount_t i_size_seqcount;
833 unsigned long i_state;
834 struct mutex i_mutex;
836 unsigned long dirtied_when; /* jiffies of first dirtying */
838 struct hlist_node i_hash;
839 struct list_head i_wb_list; /* backing dev IO list */
840 struct list_head i_lru; /* inode LRU list */
841 struct list_head i_sb_list;
843 struct hlist_head i_dentry;
844 struct rcu_head i_rcu;
848 atomic_t i_dio_count;
849 atomic_t i_writecount;
850 const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
851 struct file_lock *i_flock;
852 struct address_space i_data;
854 struct dquot *i_dquot[MAXQUOTAS];
856 struct list_head i_devices;
858 struct pipe_inode_info *i_pipe;
859 struct block_device *i_bdev;
865 #ifdef CONFIG_FSNOTIFY
866 __u32 i_fsnotify_mask; /* all events this inode cares about */
867 struct hlist_head i_fsnotify_marks;
871 atomic_t i_readcount; /* struct files open RO */
873 void *i_private; /* fs or device private pointer */
876 static inline int inode_unhashed(struct inode *inode)
878 return hlist_unhashed(&inode->i_hash);
882 * inode->i_mutex nesting subclasses for the lock validator:
884 * 0: the object of the current VFS operation
889 * The locking order between these classes is
890 * parent -> child -> normal -> xattr -> quota
892 enum inode_i_mutex_lock_class
902 * NOTE: in a 32bit arch with a preemptable kernel and
903 * an UP compile the i_size_read/write must be atomic
904 * with respect to the local cpu (unlike with preempt disabled),
905 * but they don't need to be atomic with respect to other cpus like in
906 * true SMP (so they need either to either locally disable irq around
907 * the read or for example on x86 they can be still implemented as a
908 * cmpxchg8b without the need of the lock prefix). For SMP compiles
909 * and 64bit archs it makes no difference if preempt is enabled or not.
911 static inline loff_t i_size_read(const struct inode *inode)
913 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
918 seq = read_seqcount_begin(&inode->i_size_seqcount);
919 i_size = inode->i_size;
920 } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
922 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
926 i_size = inode->i_size;
930 return inode->i_size;
935 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
936 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
937 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
939 static inline void i_size_write(struct inode *inode, loff_t i_size)
941 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
942 write_seqcount_begin(&inode->i_size_seqcount);
943 inode->i_size = i_size;
944 write_seqcount_end(&inode->i_size_seqcount);
945 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
947 inode->i_size = i_size;
950 inode->i_size = i_size;
954 /* Helper functions so that in most cases filesystems will
955 * not need to deal directly with kuid_t and kgid_t and can
956 * instead deal with the raw numeric values that are stored
959 static inline uid_t i_uid_read(const struct inode *inode)
961 return from_kuid(&init_user_ns, inode->i_uid);
964 static inline gid_t i_gid_read(const struct inode *inode)
966 return from_kgid(&init_user_ns, inode->i_gid);
969 static inline void i_uid_write(struct inode *inode, uid_t uid)
971 inode->i_uid = make_kuid(&init_user_ns, uid);
974 static inline void i_gid_write(struct inode *inode, gid_t gid)
976 inode->i_gid = make_kgid(&init_user_ns, gid);
979 static inline unsigned iminor(const struct inode *inode)
981 return MINOR(inode->i_rdev);
984 static inline unsigned imajor(const struct inode *inode)
986 return MAJOR(inode->i_rdev);
989 extern struct block_device *I_BDEV(struct inode *inode);
992 rwlock_t lock; /* protects pid, uid, euid fields */
993 struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
994 enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
995 kuid_t uid, euid; /* uid/euid of process setting the owner */
996 int signum; /* posix.1b rt signal to be delivered on IO */
1000 * Track a single file's readahead state
1002 struct file_ra_state {
1003 pgoff_t start; /* where readahead started */
1004 unsigned int size; /* # of readahead pages */
1005 unsigned int async_size; /* do asynchronous readahead when
1006 there are only # of pages ahead */
1008 unsigned int ra_pages; /* Maximum readahead window */
1009 unsigned int mmap_miss; /* Cache miss stat for mmap accesses */
1010 loff_t prev_pos; /* Cache last read() position */
1014 * Check if @index falls in the readahead windows.
1016 static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
1018 return (index >= ra->start &&
1019 index < ra->start + ra->size);
1022 #define FILE_MNT_WRITE_TAKEN 1
1023 #define FILE_MNT_WRITE_RELEASED 2
1027 * fu_list becomes invalid after file_free is called and queued via
1028 * fu_rcuhead for RCU freeing
1031 struct list_head fu_list;
1032 struct rcu_head fu_rcuhead;
1035 #define f_dentry f_path.dentry
1036 #define f_vfsmnt f_path.mnt
1037 const struct file_operations *f_op;
1040 * Protects f_ep_links, f_flags, f_pos vs i_size in lseek SEEK_CUR.
1041 * Must not be taken from IRQ context.
1047 atomic_long_t f_count;
1048 unsigned int f_flags;
1051 struct fown_struct f_owner;
1052 const struct cred *f_cred;
1053 struct file_ra_state f_ra;
1056 #ifdef CONFIG_SECURITY
1059 /* needed for tty driver, and maybe others */
1063 /* Used by fs/eventpoll.c to link all the hooks to this file */
1064 struct list_head f_ep_links;
1065 struct list_head f_tfile_llink;
1066 #endif /* #ifdef CONFIG_EPOLL */
1067 struct address_space *f_mapping;
1068 #ifdef CONFIG_DEBUG_WRITECOUNT
1069 unsigned long f_mnt_write_state;
1073 struct file_handle {
1076 /* file identifier */
1077 unsigned char f_handle[0];
1080 static inline struct file *get_file(struct file *f)
1082 atomic_long_inc(&f->f_count);
1085 #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
1086 #define file_count(x) atomic_long_read(&(x)->f_count)
1088 #ifdef CONFIG_DEBUG_WRITECOUNT
1089 static inline void file_take_write(struct file *f)
1091 WARN_ON(f->f_mnt_write_state != 0);
1092 f->f_mnt_write_state = FILE_MNT_WRITE_TAKEN;
1094 static inline void file_release_write(struct file *f)
1096 f->f_mnt_write_state |= FILE_MNT_WRITE_RELEASED;
1098 static inline void file_reset_write(struct file *f)
1100 f->f_mnt_write_state = 0;
1102 static inline void file_check_state(struct file *f)
1105 * At this point, either both or neither of these bits
1108 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN);
1109 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_RELEASED);
1111 static inline int file_check_writeable(struct file *f)
1113 if (f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN)
1115 printk(KERN_WARNING "writeable file with no "
1116 "mnt_want_write()\n");
1120 #else /* !CONFIG_DEBUG_WRITECOUNT */
1121 static inline void file_take_write(struct file *filp) {}
1122 static inline void file_release_write(struct file *filp) {}
1123 static inline void file_reset_write(struct file *filp) {}
1124 static inline void file_check_state(struct file *filp) {}
1125 static inline int file_check_writeable(struct file *filp)
1129 #endif /* CONFIG_DEBUG_WRITECOUNT */
1131 #define MAX_NON_LFS ((1UL<<31) - 1)
1133 /* Page cache limit. The filesystems should put that into their s_maxbytes
1134 limits, otherwise bad things can happen in VM. */
1135 #if BITS_PER_LONG==32
1136 #define MAX_LFS_FILESIZE (((loff_t)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
1137 #elif BITS_PER_LONG==64
1138 #define MAX_LFS_FILESIZE ((loff_t)0x7fffffffffffffffLL)
1143 #define FL_ACCESS 8 /* not trying to lock, just looking */
1144 #define FL_EXISTS 16 /* when unlocking, test for existence */
1145 #define FL_LEASE 32 /* lease held on this file */
1146 #define FL_CLOSE 64 /* unlock on close */
1147 #define FL_SLEEP 128 /* A blocking lock */
1148 #define FL_DOWNGRADE_PENDING 256 /* Lease is being downgraded */
1149 #define FL_UNLOCK_PENDING 512 /* Lease is being broken */
1152 * Special return value from posix_lock_file() and vfs_lock_file() for
1153 * asynchronous locking.
1155 #define FILE_LOCK_DEFERRED 1
1158 * The POSIX file lock owner is determined by
1159 * the "struct files_struct" in the thread group
1160 * (or NULL for no owner - BSD locks).
1162 * Lockd stuffs a "host" pointer into this.
1164 typedef struct files_struct *fl_owner_t;
1166 struct file_lock_operations {
1167 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
1168 void (*fl_release_private)(struct file_lock *);
1171 struct lock_manager_operations {
1172 int (*lm_compare_owner)(struct file_lock *, struct file_lock *);
1173 void (*lm_notify)(struct file_lock *); /* unblock callback */
1174 int (*lm_grant)(struct file_lock *, struct file_lock *, int);
1175 void (*lm_break)(struct file_lock *);
1176 int (*lm_change)(struct file_lock **, int);
1179 struct lock_manager {
1180 struct list_head list;
1184 void locks_start_grace(struct net *, struct lock_manager *);
1185 void locks_end_grace(struct lock_manager *);
1186 int locks_in_grace(struct net *);
1188 /* that will die - we need it for nfs_lock_info */
1189 #include <linux/nfs_fs_i.h>
1192 struct file_lock *fl_next; /* singly linked list for this inode */
1193 struct list_head fl_link; /* doubly linked list of all locks */
1194 struct list_head fl_block; /* circular list of blocked processes */
1195 fl_owner_t fl_owner;
1196 unsigned int fl_flags;
1197 unsigned char fl_type;
1198 unsigned int fl_pid;
1199 struct pid *fl_nspid;
1200 wait_queue_head_t fl_wait;
1201 struct file *fl_file;
1205 struct fasync_struct * fl_fasync; /* for lease break notifications */
1206 /* for lease breaks: */
1207 unsigned long fl_break_time;
1208 unsigned long fl_downgrade_time;
1210 const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
1211 const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
1213 struct nfs_lock_info nfs_fl;
1214 struct nfs4_lock_info nfs4_fl;
1216 struct list_head link; /* link in AFS vnode's pending_locks list */
1217 int state; /* state of grant or error if -ve */
1222 /* The following constant reflects the upper bound of the file/locking space */
1224 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
1225 #define OFFSET_MAX INT_LIMIT(loff_t)
1226 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
1229 #include <linux/fcntl.h>
1231 extern void send_sigio(struct fown_struct *fown, int fd, int band);
1233 #ifdef CONFIG_FILE_LOCKING
1234 extern int fcntl_getlk(struct file *, struct flock __user *);
1235 extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
1236 struct flock __user *);
1238 #if BITS_PER_LONG == 32
1239 extern int fcntl_getlk64(struct file *, struct flock64 __user *);
1240 extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
1241 struct flock64 __user *);
1244 extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
1245 extern int fcntl_getlease(struct file *filp);
1248 void locks_free_lock(struct file_lock *fl);
1249 extern void locks_init_lock(struct file_lock *);
1250 extern struct file_lock * locks_alloc_lock(void);
1251 extern void locks_copy_lock(struct file_lock *, struct file_lock *);
1252 extern void __locks_copy_lock(struct file_lock *, const struct file_lock *);
1253 extern void locks_remove_posix(struct file *, fl_owner_t);
1254 extern void locks_remove_flock(struct file *);
1255 extern void locks_release_private(struct file_lock *);
1256 extern void posix_test_lock(struct file *, struct file_lock *);
1257 extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
1258 extern int posix_lock_file_wait(struct file *, struct file_lock *);
1259 extern int posix_unblock_lock(struct file *, struct file_lock *);
1260 extern int vfs_test_lock(struct file *, struct file_lock *);
1261 extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
1262 extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
1263 extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
1264 extern int __break_lease(struct inode *inode, unsigned int flags);
1265 extern void lease_get_mtime(struct inode *, struct timespec *time);
1266 extern int generic_setlease(struct file *, long, struct file_lock **);
1267 extern int vfs_setlease(struct file *, long, struct file_lock **);
1268 extern int lease_modify(struct file_lock **, int);
1269 extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
1270 extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
1271 extern void locks_delete_block(struct file_lock *waiter);
1272 extern void lock_flocks(void);
1273 extern void unlock_flocks(void);
1274 #else /* !CONFIG_FILE_LOCKING */
1275 static inline int fcntl_getlk(struct file *file, struct flock __user *user)
1280 static inline int fcntl_setlk(unsigned int fd, struct file *file,
1281 unsigned int cmd, struct flock __user *user)
1286 #if BITS_PER_LONG == 32
1287 static inline int fcntl_getlk64(struct file *file, struct flock64 __user *user)
1292 static inline int fcntl_setlk64(unsigned int fd, struct file *file,
1293 unsigned int cmd, struct flock64 __user *user)
1298 static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
1303 static inline int fcntl_getlease(struct file *filp)
1308 static inline void locks_init_lock(struct file_lock *fl)
1313 static inline void __locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1318 static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1323 static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
1328 static inline void locks_remove_flock(struct file *filp)
1333 static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
1338 static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
1339 struct file_lock *conflock)
1344 static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
1349 static inline int posix_unblock_lock(struct file *filp,
1350 struct file_lock *waiter)
1355 static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
1360 static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
1361 struct file_lock *fl, struct file_lock *conf)
1366 static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
1371 static inline int flock_lock_file_wait(struct file *filp,
1372 struct file_lock *request)
1377 static inline int __break_lease(struct inode *inode, unsigned int mode)
1382 static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
1387 static inline int generic_setlease(struct file *filp, long arg,
1388 struct file_lock **flp)
1393 static inline int vfs_setlease(struct file *filp, long arg,
1394 struct file_lock **lease)
1399 static inline int lease_modify(struct file_lock **before, int arg)
1404 static inline int lock_may_read(struct inode *inode, loff_t start,
1410 static inline int lock_may_write(struct inode *inode, loff_t start,
1416 static inline void locks_delete_block(struct file_lock *waiter)
1420 static inline void lock_flocks(void)
1424 static inline void unlock_flocks(void)
1428 #endif /* !CONFIG_FILE_LOCKING */
1431 struct fasync_struct {
1435 struct fasync_struct *fa_next; /* singly linked list */
1436 struct file *fa_file;
1437 struct rcu_head fa_rcu;
1440 #define FASYNC_MAGIC 0x4601
1442 /* SMP safe fasync helpers: */
1443 extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
1444 extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
1445 extern int fasync_remove_entry(struct file *, struct fasync_struct **);
1446 extern struct fasync_struct *fasync_alloc(void);
1447 extern void fasync_free(struct fasync_struct *);
1449 /* can be called from interrupts */
1450 extern void kill_fasync(struct fasync_struct **, int, int);
1452 extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
1453 extern int f_setown(struct file *filp, unsigned long arg, int force);
1454 extern void f_delown(struct file *filp);
1455 extern pid_t f_getown(struct file *filp);
1456 extern int send_sigurg(struct fown_struct *fown);
1464 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1465 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1466 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1467 #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1468 #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
1470 extern struct list_head super_blocks;
1471 extern spinlock_t sb_lock;
1473 /* Possible states of 'frozen' field */
1475 SB_UNFROZEN = 0, /* FS is unfrozen */
1476 SB_FREEZE_WRITE = 1, /* Writes, dir ops, ioctls frozen */
1477 SB_FREEZE_PAGEFAULT = 2, /* Page faults stopped as well */
1478 SB_FREEZE_FS = 3, /* For internal FS use (e.g. to stop
1479 * internal threads if needed) */
1480 SB_FREEZE_COMPLETE = 4, /* ->freeze_fs finished successfully */
1483 #define SB_FREEZE_LEVELS (SB_FREEZE_COMPLETE - 1)
1486 /* Counters for counting writers at each level */
1487 struct percpu_counter counter[SB_FREEZE_LEVELS];
1488 wait_queue_head_t wait; /* queue for waiting for
1489 writers / faults to finish */
1490 int frozen; /* Is sb frozen? */
1491 wait_queue_head_t wait_unfrozen; /* queue for waiting for
1493 #ifdef CONFIG_DEBUG_LOCK_ALLOC
1494 struct lockdep_map lock_map[SB_FREEZE_LEVELS];
1498 struct super_block {
1499 struct list_head s_list; /* Keep this first */
1500 dev_t s_dev; /* search index; _not_ kdev_t */
1501 unsigned char s_blocksize_bits;
1502 unsigned long s_blocksize;
1503 loff_t s_maxbytes; /* Max file size */
1504 struct file_system_type *s_type;
1505 const struct super_operations *s_op;
1506 const struct dquot_operations *dq_op;
1507 const struct quotactl_ops *s_qcop;
1508 const struct export_operations *s_export_op;
1509 unsigned long s_flags;
1510 unsigned long s_magic;
1511 struct dentry *s_root;
1512 struct rw_semaphore s_umount;
1515 #ifdef CONFIG_SECURITY
1518 const struct xattr_handler **s_xattr;
1520 struct list_head s_inodes; /* all inodes */
1521 struct hlist_bl_head s_anon; /* anonymous dentries for (nfs) exporting */
1523 struct list_head __percpu *s_files;
1525 struct list_head s_files;
1527 struct list_head s_mounts; /* list of mounts; _not_ for fs use */
1528 /* s_dentry_lru, s_nr_dentry_unused protected by dcache.c lru locks */
1529 struct list_head s_dentry_lru; /* unused dentry lru */
1530 int s_nr_dentry_unused; /* # of dentry on lru */
1532 /* s_inode_lru_lock protects s_inode_lru and s_nr_inodes_unused */
1533 spinlock_t s_inode_lru_lock ____cacheline_aligned_in_smp;
1534 struct list_head s_inode_lru; /* unused inode lru */
1535 int s_nr_inodes_unused; /* # of inodes on lru */
1537 struct block_device *s_bdev;
1538 struct backing_dev_info *s_bdi;
1539 struct mtd_info *s_mtd;
1540 struct hlist_node s_instances;
1541 struct quota_info s_dquot; /* Diskquota specific options */
1543 struct sb_writers s_writers;
1545 char s_id[32]; /* Informational name */
1546 u8 s_uuid[16]; /* UUID */
1548 void *s_fs_info; /* Filesystem private info */
1549 unsigned int s_max_links;
1552 /* Granularity of c/m/atime in ns.
1553 Cannot be worse than a second */
1557 * The next field is for VFS *only*. No filesystems have any business
1558 * even looking at it. You had been warned.
1560 struct mutex s_vfs_rename_mutex; /* Kludge */
1563 * Filesystem subtype. If non-empty the filesystem type field
1564 * in /proc/mounts will be "type.subtype"
1569 * Saved mount options for lazy filesystems using
1570 * generic_show_options()
1572 char __rcu *s_options;
1573 const struct dentry_operations *s_d_op; /* default d_op for dentries */
1576 * Saved pool identifier for cleancache (-1 means none)
1578 int cleancache_poolid;
1580 struct shrinker s_shrink; /* per-sb shrinker handle */
1582 /* Number of inodes with nlink == 0 but still referenced */
1583 atomic_long_t s_remove_count;
1585 /* Being remounted read-only */
1586 int s_readonly_remount;
1589 /* superblock cache pruning functions */
1590 extern void prune_icache_sb(struct super_block *sb, int nr_to_scan);
1591 extern void prune_dcache_sb(struct super_block *sb, int nr_to_scan);
1593 extern struct timespec current_fs_time(struct super_block *sb);
1596 * Snapshotting support.
1599 void __sb_end_write(struct super_block *sb, int level);
1600 int __sb_start_write(struct super_block *sb, int level, bool wait);
1603 * sb_end_write - drop write access to a superblock
1604 * @sb: the super we wrote to
1606 * Decrement number of writers to the filesystem. Wake up possible waiters
1607 * wanting to freeze the filesystem.
1609 static inline void sb_end_write(struct super_block *sb)
1611 __sb_end_write(sb, SB_FREEZE_WRITE);
1615 * sb_end_pagefault - drop write access to a superblock from a page fault
1616 * @sb: the super we wrote to
1618 * Decrement number of processes handling write page fault to the filesystem.
1619 * Wake up possible waiters wanting to freeze the filesystem.
1621 static inline void sb_end_pagefault(struct super_block *sb)
1623 __sb_end_write(sb, SB_FREEZE_PAGEFAULT);
1627 * sb_end_intwrite - drop write access to a superblock for internal fs purposes
1628 * @sb: the super we wrote to
1630 * Decrement fs-internal number of writers to the filesystem. Wake up possible
1631 * waiters wanting to freeze the filesystem.
1633 static inline void sb_end_intwrite(struct super_block *sb)
1635 __sb_end_write(sb, SB_FREEZE_FS);
1639 * sb_start_write - get write access to a superblock
1640 * @sb: the super we write to
1642 * When a process wants to write data or metadata to a file system (i.e. dirty
1643 * a page or an inode), it should embed the operation in a sb_start_write() -
1644 * sb_end_write() pair to get exclusion against file system freezing. This
1645 * function increments number of writers preventing freezing. If the file
1646 * system is already frozen, the function waits until the file system is
1649 * Since freeze protection behaves as a lock, users have to preserve
1650 * ordering of freeze protection and other filesystem locks. Generally,
1651 * freeze protection should be the outermost lock. In particular, we have:
1654 * -> i_mutex (write path, truncate, directory ops, ...)
1655 * -> s_umount (freeze_super, thaw_super)
1657 static inline void sb_start_write(struct super_block *sb)
1659 __sb_start_write(sb, SB_FREEZE_WRITE, true);
1662 static inline int sb_start_write_trylock(struct super_block *sb)
1664 return __sb_start_write(sb, SB_FREEZE_WRITE, false);
1668 * sb_start_pagefault - get write access to a superblock from a page fault
1669 * @sb: the super we write to
1671 * When a process starts handling write page fault, it should embed the
1672 * operation into sb_start_pagefault() - sb_end_pagefault() pair to get
1673 * exclusion against file system freezing. This is needed since the page fault
1674 * is going to dirty a page. This function increments number of running page
1675 * faults preventing freezing. If the file system is already frozen, the
1676 * function waits until the file system is thawed.
1678 * Since page fault freeze protection behaves as a lock, users have to preserve
1679 * ordering of freeze protection and other filesystem locks. It is advised to
1680 * put sb_start_pagefault() close to mmap_sem in lock ordering. Page fault
1681 * handling code implies lock dependency:
1684 * -> sb_start_pagefault
1686 static inline void sb_start_pagefault(struct super_block *sb)
1688 __sb_start_write(sb, SB_FREEZE_PAGEFAULT, true);
1692 * sb_start_intwrite - get write access to a superblock for internal fs purposes
1693 * @sb: the super we write to
1695 * This is the third level of protection against filesystem freezing. It is
1696 * free for use by a filesystem. The only requirement is that it must rank
1697 * below sb_start_pagefault.
1699 * For example filesystem can call sb_start_intwrite() when starting a
1700 * transaction which somewhat eases handling of freezing for internal sources
1701 * of filesystem changes (internal fs threads, discarding preallocation on file
1704 static inline void sb_start_intwrite(struct super_block *sb)
1706 __sb_start_write(sb, SB_FREEZE_FS, true);
1710 extern bool inode_owner_or_capable(const struct inode *inode);
1712 /* not quite ready to be deprecated, but... */
1713 extern void lock_super(struct super_block *);
1714 extern void unlock_super(struct super_block *);
1717 * VFS helper functions..
1719 extern int vfs_create(struct inode *, struct dentry *, umode_t, bool);
1720 extern int vfs_mkdir(struct inode *, struct dentry *, umode_t);
1721 extern int vfs_mknod(struct inode *, struct dentry *, umode_t, dev_t);
1722 extern int vfs_symlink(struct inode *, struct dentry *, const char *);
1723 extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
1724 extern int vfs_rmdir(struct inode *, struct dentry *);
1725 extern int vfs_unlink(struct inode *, struct dentry *);
1726 extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
1729 * VFS dentry helper functions.
1731 extern void dentry_unhash(struct dentry *dentry);
1734 * VFS file helper functions.
1736 extern void inode_init_owner(struct inode *inode, const struct inode *dir,
1739 * VFS FS_IOC_FIEMAP helper definitions.
1741 struct fiemap_extent_info {
1742 unsigned int fi_flags; /* Flags as passed from user */
1743 unsigned int fi_extents_mapped; /* Number of mapped extents */
1744 unsigned int fi_extents_max; /* Size of fiemap_extent array */
1745 struct fiemap_extent __user *fi_extents_start; /* Start of
1746 fiemap_extent array */
1748 int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
1749 u64 phys, u64 len, u32 flags);
1750 int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
1755 * NOTE! These match bits 12..15 of stat.st_mode
1756 * (ie "(i_mode >> 12) & 15").
1758 #define DT_UNKNOWN 0
1769 * This is the "filldir" function type, used by readdir() to let
1770 * the kernel specify what kind of dirent layout it wants to have.
1771 * This allows the kernel to read directories into kernel space or
1772 * to have different dirent layouts depending on the binary type.
1774 typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned);
1775 struct block_device_operations;
1777 /* These macros are for out of kernel modules to test that
1778 * the kernel supports the unlocked_ioctl and compat_ioctl
1779 * fields in struct file_operations. */
1780 #define HAVE_COMPAT_IOCTL 1
1781 #define HAVE_UNLOCKED_IOCTL 1
1783 struct file_operations {
1784 struct module *owner;
1785 loff_t (*llseek) (struct file *, loff_t, int);
1786 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1787 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
1788 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1789 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1790 int (*readdir) (struct file *, void *, filldir_t);
1791 unsigned int (*poll) (struct file *, struct poll_table_struct *);
1792 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1793 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1794 int (*mmap) (struct file *, struct vm_area_struct *);
1795 int (*open) (struct inode *, struct file *);
1796 int (*flush) (struct file *, fl_owner_t id);
1797 int (*release) (struct inode *, struct file *);
1798 int (*fsync) (struct file *, loff_t, loff_t, int datasync);
1799 int (*aio_fsync) (struct kiocb *, int datasync);
1800 int (*fasync) (int, struct file *, int);
1801 int (*lock) (struct file *, int, struct file_lock *);
1802 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1803 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1804 int (*check_flags)(int);
1805 int (*flock) (struct file *, int, struct file_lock *);
1806 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1807 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
1808 int (*setlease)(struct file *, long, struct file_lock **);
1809 long (*fallocate)(struct file *file, int mode, loff_t offset,
1813 struct inode_operations {
1814 struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int);
1815 void * (*follow_link) (struct dentry *, struct nameidata *);
1816 int (*permission) (struct inode *, int);
1817 struct posix_acl * (*get_acl)(struct inode *, int);
1819 int (*readlink) (struct dentry *, char __user *,int);
1820 void (*put_link) (struct dentry *, struct nameidata *, void *);
1822 int (*create) (struct inode *,struct dentry *, umode_t, bool);
1823 int (*link) (struct dentry *,struct inode *,struct dentry *);
1824 int (*unlink) (struct inode *,struct dentry *);
1825 int (*symlink) (struct inode *,struct dentry *,const char *);
1826 int (*mkdir) (struct inode *,struct dentry *,umode_t);
1827 int (*rmdir) (struct inode *,struct dentry *);
1828 int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t);
1829 int (*rename) (struct inode *, struct dentry *,
1830 struct inode *, struct dentry *);
1831 void (*truncate) (struct inode *);
1832 int (*setattr) (struct dentry *, struct iattr *);
1833 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1834 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1835 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1836 ssize_t (*listxattr) (struct dentry *, char *, size_t);
1837 int (*removexattr) (struct dentry *, const char *);
1838 int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
1840 int (*update_time)(struct inode *, struct timespec *, int);
1841 int (*atomic_open)(struct inode *, struct dentry *,
1842 struct file *, unsigned open_flag,
1843 umode_t create_mode, int *opened);
1844 } ____cacheline_aligned;
1848 ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
1849 unsigned long nr_segs, unsigned long fast_segs,
1850 struct iovec *fast_pointer,
1851 struct iovec **ret_pointer);
1853 extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1854 extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1855 extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1856 unsigned long, loff_t *);
1857 extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1858 unsigned long, loff_t *);
1860 struct super_operations {
1861 struct inode *(*alloc_inode)(struct super_block *sb);
1862 void (*destroy_inode)(struct inode *);
1864 void (*dirty_inode) (struct inode *, int flags);
1865 int (*write_inode) (struct inode *, struct writeback_control *wbc);
1866 int (*drop_inode) (struct inode *);
1867 void (*evict_inode) (struct inode *);
1868 void (*put_super) (struct super_block *);
1869 int (*sync_fs)(struct super_block *sb, int wait);
1870 int (*freeze_fs) (struct super_block *);
1871 int (*unfreeze_fs) (struct super_block *);
1872 int (*statfs) (struct dentry *, struct kstatfs *);
1873 int (*remount_fs) (struct super_block *, int *, char *);
1874 void (*umount_begin) (struct super_block *);
1876 int (*show_options)(struct seq_file *, struct dentry *);
1877 int (*show_devname)(struct seq_file *, struct dentry *);
1878 int (*show_path)(struct seq_file *, struct dentry *);
1879 int (*show_stats)(struct seq_file *, struct dentry *);
1881 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1882 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
1884 int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
1885 int (*nr_cached_objects)(struct super_block *);
1886 void (*free_cached_objects)(struct super_block *, int);
1890 * Inode state bits. Protected by inode->i_lock
1892 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1893 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1895 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1896 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1897 * various stages of removing an inode.
1899 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1901 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1902 * fdatasync(). i_atime is the usual cause.
1903 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1904 * these changes separately from I_DIRTY_SYNC so that we
1905 * don't have to write inode on fdatasync() when only
1906 * mtime has changed in it.
1907 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1908 * I_NEW Serves as both a mutex and completion notification.
1909 * New inodes set I_NEW. If two processes both create
1910 * the same inode, one of them will release its inode and
1911 * wait for I_NEW to be released before returning.
1912 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1913 * also cause waiting on I_NEW, without I_NEW actually
1914 * being set. find_inode() uses this to prevent returning
1915 * nearly-dead inodes.
1916 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1917 * is zero. I_FREEING must be set when I_WILL_FREE is
1919 * I_FREEING Set when inode is about to be freed but still has dirty
1920 * pages or buffers attached or the inode itself is still
1922 * I_CLEAR Added by clear_inode(). In this state the inode is
1923 * clean and can be destroyed. Inode keeps I_FREEING.
1925 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1926 * prohibited for many purposes. iget() must wait for
1927 * the inode to be completely released, then create it
1928 * anew. Other functions will just ignore such inodes,
1929 * if appropriate. I_NEW is used for waiting.
1931 * I_SYNC Writeback of inode is running. The bit is set during
1932 * data writeback, and cleared with a wakeup on the bit
1933 * address once it is done. The bit is also used to pin
1934 * the inode in memory for flusher thread.
1936 * I_REFERENCED Marks the inode as recently references on the LRU list.
1938 * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit().
1940 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1942 #define I_DIRTY_SYNC (1 << 0)
1943 #define I_DIRTY_DATASYNC (1 << 1)
1944 #define I_DIRTY_PAGES (1 << 2)
1946 #define I_NEW (1 << __I_NEW)
1947 #define I_WILL_FREE (1 << 4)
1948 #define I_FREEING (1 << 5)
1949 #define I_CLEAR (1 << 6)
1951 #define I_SYNC (1 << __I_SYNC)
1952 #define I_REFERENCED (1 << 8)
1953 #define __I_DIO_WAKEUP 9
1954 #define I_DIO_WAKEUP (1 << I_DIO_WAKEUP)
1956 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1958 extern void __mark_inode_dirty(struct inode *, int);
1959 static inline void mark_inode_dirty(struct inode *inode)
1961 __mark_inode_dirty(inode, I_DIRTY);
1964 static inline void mark_inode_dirty_sync(struct inode *inode)
1966 __mark_inode_dirty(inode, I_DIRTY_SYNC);
1969 extern void inc_nlink(struct inode *inode);
1970 extern void drop_nlink(struct inode *inode);
1971 extern void clear_nlink(struct inode *inode);
1972 extern void set_nlink(struct inode *inode, unsigned int nlink);
1974 static inline void inode_inc_link_count(struct inode *inode)
1977 mark_inode_dirty(inode);
1980 static inline void inode_dec_link_count(struct inode *inode)
1983 mark_inode_dirty(inode);
1987 * inode_inc_iversion - increments i_version
1988 * @inode: inode that need to be updated
1990 * Every time the inode is modified, the i_version field will be incremented.
1991 * The filesystem has to be mounted with i_version flag
1994 static inline void inode_inc_iversion(struct inode *inode)
1996 spin_lock(&inode->i_lock);
1998 spin_unlock(&inode->i_lock);
2001 enum file_time_flags {
2008 extern void touch_atime(struct path *);
2009 static inline void file_accessed(struct file *file)
2011 if (!(file->f_flags & O_NOATIME))
2012 touch_atime(&file->f_path);
2015 int sync_inode(struct inode *inode, struct writeback_control *wbc);
2016 int sync_inode_metadata(struct inode *inode, int wait);
2018 struct file_system_type {
2021 struct dentry *(*mount) (struct file_system_type *, int,
2022 const char *, void *);
2023 void (*kill_sb) (struct super_block *);
2024 struct module *owner;
2025 struct file_system_type * next;
2026 struct hlist_head fs_supers;
2028 struct lock_class_key s_lock_key;
2029 struct lock_class_key s_umount_key;
2030 struct lock_class_key s_vfs_rename_key;
2031 struct lock_class_key s_writers_key[SB_FREEZE_LEVELS];
2033 struct lock_class_key i_lock_key;
2034 struct lock_class_key i_mutex_key;
2035 struct lock_class_key i_mutex_dir_key;
2038 extern struct dentry *mount_ns(struct file_system_type *fs_type, int flags,
2039 void *data, int (*fill_super)(struct super_block *, void *, int));
2040 extern struct dentry *mount_bdev(struct file_system_type *fs_type,
2041 int flags, const char *dev_name, void *data,
2042 int (*fill_super)(struct super_block *, void *, int));
2043 extern struct dentry *mount_single(struct file_system_type *fs_type,
2044 int flags, void *data,
2045 int (*fill_super)(struct super_block *, void *, int));
2046 extern struct dentry *mount_nodev(struct file_system_type *fs_type,
2047 int flags, void *data,
2048 int (*fill_super)(struct super_block *, void *, int));
2049 extern struct dentry *mount_subtree(struct vfsmount *mnt, const char *path);
2050 void generic_shutdown_super(struct super_block *sb);
2051 void kill_block_super(struct super_block *sb);
2052 void kill_anon_super(struct super_block *sb);
2053 void kill_litter_super(struct super_block *sb);
2054 void deactivate_super(struct super_block *sb);
2055 void deactivate_locked_super(struct super_block *sb);
2056 int set_anon_super(struct super_block *s, void *data);
2057 int get_anon_bdev(dev_t *);
2058 void free_anon_bdev(dev_t);
2059 struct super_block *sget(struct file_system_type *type,
2060 int (*test)(struct super_block *,void *),
2061 int (*set)(struct super_block *,void *),
2062 int flags, void *data);
2063 extern struct dentry *mount_pseudo(struct file_system_type *, char *,
2064 const struct super_operations *ops,
2065 const struct dentry_operations *dops,
2068 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
2069 #define fops_get(fops) \
2070 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
2071 #define fops_put(fops) \
2072 do { if (fops) module_put((fops)->owner); } while(0)
2074 extern int register_filesystem(struct file_system_type *);
2075 extern int unregister_filesystem(struct file_system_type *);
2076 extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
2077 #define kern_mount(type) kern_mount_data(type, NULL)
2078 extern void kern_unmount(struct vfsmount *mnt);
2079 extern int may_umount_tree(struct vfsmount *);
2080 extern int may_umount(struct vfsmount *);
2081 extern long do_mount(const char *, const char *, const char *, unsigned long, void *);
2082 extern struct vfsmount *collect_mounts(struct path *);
2083 extern void drop_collected_mounts(struct vfsmount *);
2084 extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
2086 extern int vfs_statfs(struct path *, struct kstatfs *);
2087 extern int user_statfs(const char __user *, struct kstatfs *);
2088 extern int fd_statfs(int, struct kstatfs *);
2089 extern int vfs_ustat(dev_t, struct kstatfs *);
2090 extern int freeze_super(struct super_block *super);
2091 extern int thaw_super(struct super_block *super);
2092 extern bool our_mnt(struct vfsmount *mnt);
2094 extern int current_umask(void);
2097 extern struct kobject *fs_kobj;
2099 #define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
2100 extern int rw_verify_area(int, struct file *, loff_t *, size_t);
2102 #define FLOCK_VERIFY_READ 1
2103 #define FLOCK_VERIFY_WRITE 2
2105 #ifdef CONFIG_FILE_LOCKING
2106 extern int locks_mandatory_locked(struct inode *);
2107 extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
2110 * Candidates for mandatory locking have the setgid bit set
2111 * but no group execute bit - an otherwise meaningless combination.
2114 static inline int __mandatory_lock(struct inode *ino)
2116 return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
2120 * ... and these candidates should be on MS_MANDLOCK mounted fs,
2121 * otherwise these will be advisory locks
2124 static inline int mandatory_lock(struct inode *ino)
2126 return IS_MANDLOCK(ino) && __mandatory_lock(ino);
2129 static inline int locks_verify_locked(struct inode *inode)
2131 if (mandatory_lock(inode))
2132 return locks_mandatory_locked(inode);
2136 static inline int locks_verify_truncate(struct inode *inode,
2140 if (inode->i_flock && mandatory_lock(inode))
2141 return locks_mandatory_area(
2142 FLOCK_VERIFY_WRITE, inode, filp,
2143 size < inode->i_size ? size : inode->i_size,
2144 (size < inode->i_size ? inode->i_size - size
2145 : size - inode->i_size)
2150 static inline int break_lease(struct inode *inode, unsigned int mode)
2153 return __break_lease(inode, mode);
2156 #else /* !CONFIG_FILE_LOCKING */
2157 static inline int locks_mandatory_locked(struct inode *inode)
2162 static inline int locks_mandatory_area(int rw, struct inode *inode,
2163 struct file *filp, loff_t offset,
2169 static inline int __mandatory_lock(struct inode *inode)
2174 static inline int mandatory_lock(struct inode *inode)
2179 static inline int locks_verify_locked(struct inode *inode)
2184 static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
2190 static inline int break_lease(struct inode *inode, unsigned int mode)
2195 #endif /* CONFIG_FILE_LOCKING */
2200 const char *name; /* pointer to actual string */
2201 const __user char *uptr; /* original userland pointer */
2202 struct audit_names *aname;
2203 bool separate; /* should "name" be freed? */
2206 extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
2208 extern int do_fallocate(struct file *file, int mode, loff_t offset,
2210 extern long do_sys_open(int dfd, const char __user *filename, int flags,
2212 extern struct file *file_open_name(struct filename *, int, umode_t);
2213 extern struct file *filp_open(const char *, int, umode_t);
2214 extern struct file *file_open_root(struct dentry *, struct vfsmount *,
2216 extern struct file * dentry_open(const struct path *, int, const struct cred *);
2217 extern int filp_close(struct file *, fl_owner_t id);
2219 extern struct filename *getname(const char __user *);
2225 extern int finish_open(struct file *file, struct dentry *dentry,
2226 int (*open)(struct inode *, struct file *),
2228 extern int finish_no_open(struct file *file, struct dentry *dentry);
2232 extern int ioctl_preallocate(struct file *filp, void __user *argp);
2235 extern void __init vfs_caches_init_early(void);
2236 extern void __init vfs_caches_init(unsigned long);
2238 extern struct kmem_cache *names_cachep;
2240 extern void final_putname(struct filename *name);
2242 #define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL)
2243 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
2244 #ifndef CONFIG_AUDITSYSCALL
2245 #define putname(name) final_putname(name)
2247 extern void putname(struct filename *name);
2251 extern int register_blkdev(unsigned int, const char *);
2252 extern void unregister_blkdev(unsigned int, const char *);
2253 extern struct block_device *bdget(dev_t);
2254 extern struct block_device *bdgrab(struct block_device *bdev);
2255 extern void bd_set_size(struct block_device *, loff_t size);
2256 extern sector_t blkdev_max_block(struct block_device *bdev);
2257 extern void bd_forget(struct inode *inode);
2258 extern void bdput(struct block_device *);
2259 extern void invalidate_bdev(struct block_device *);
2260 extern void iterate_bdevs(void (*)(struct block_device *, void *), void *);
2261 extern int sync_blockdev(struct block_device *bdev);
2262 extern void kill_bdev(struct block_device *);
2263 extern struct super_block *freeze_bdev(struct block_device *);
2264 extern void emergency_thaw_all(void);
2265 extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
2266 extern int fsync_bdev(struct block_device *);
2268 static inline void bd_forget(struct inode *inode) {}
2269 static inline int sync_blockdev(struct block_device *bdev) { return 0; }
2270 static inline void kill_bdev(struct block_device *bdev) {}
2271 static inline void invalidate_bdev(struct block_device *bdev) {}
2273 static inline struct super_block *freeze_bdev(struct block_device *sb)
2278 static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
2283 static inline void iterate_bdevs(void (*f)(struct block_device *, void *), void *arg)
2287 extern int sync_filesystem(struct super_block *);
2288 extern const struct file_operations def_blk_fops;
2289 extern const struct file_operations def_chr_fops;
2290 extern const struct file_operations bad_sock_fops;
2291 extern const struct file_operations def_fifo_fops;
2293 extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
2294 extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
2295 extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
2296 extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
2297 extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
2299 extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
2301 extern int blkdev_put(struct block_device *bdev, fmode_t mode);
2303 extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
2304 extern void bd_unlink_disk_holder(struct block_device *bdev,
2305 struct gendisk *disk);
2307 static inline int bd_link_disk_holder(struct block_device *bdev,
2308 struct gendisk *disk)
2312 static inline void bd_unlink_disk_holder(struct block_device *bdev,
2313 struct gendisk *disk)
2320 #define CHRDEV_MAJOR_HASH_SIZE 255
2321 extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
2322 extern int register_chrdev_region(dev_t, unsigned, const char *);
2323 extern int __register_chrdev(unsigned int major, unsigned int baseminor,
2324 unsigned int count, const char *name,
2325 const struct file_operations *fops);
2326 extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
2327 unsigned int count, const char *name);
2328 extern void unregister_chrdev_region(dev_t, unsigned);
2329 extern void chrdev_show(struct seq_file *,off_t);
2331 static inline int register_chrdev(unsigned int major, const char *name,
2332 const struct file_operations *fops)
2334 return __register_chrdev(major, 0, 256, name, fops);
2337 static inline void unregister_chrdev(unsigned int major, const char *name)
2339 __unregister_chrdev(major, 0, 256, name);
2342 /* fs/block_dev.c */
2343 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
2344 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
2347 #define BLKDEV_MAJOR_HASH_SIZE 255
2348 extern const char *__bdevname(dev_t, char *buffer);
2349 extern const char *bdevname(struct block_device *bdev, char *buffer);
2350 extern struct block_device *lookup_bdev(const char *);
2351 extern void blkdev_show(struct seq_file *,off_t);
2354 #define BLKDEV_MAJOR_HASH_SIZE 0
2357 extern void init_special_inode(struct inode *, umode_t, dev_t);
2359 /* Invalid inode operations -- fs/bad_inode.c */
2360 extern void make_bad_inode(struct inode *);
2361 extern int is_bad_inode(struct inode *);
2363 extern const struct file_operations read_pipefifo_fops;
2364 extern const struct file_operations write_pipefifo_fops;
2365 extern const struct file_operations rdwr_pipefifo_fops;
2369 * return READ, READA, or WRITE
2371 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2374 * return data direction, READ or WRITE
2376 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
2378 extern void check_disk_size_change(struct gendisk *disk,
2379 struct block_device *bdev);
2380 extern int revalidate_disk(struct gendisk *);
2381 extern int check_disk_change(struct block_device *);
2382 extern int __invalidate_device(struct block_device *, bool);
2383 extern int invalidate_partition(struct gendisk *, int);
2385 unsigned long invalidate_mapping_pages(struct address_space *mapping,
2386 pgoff_t start, pgoff_t end);
2388 static inline void invalidate_remote_inode(struct inode *inode)
2390 if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
2391 S_ISLNK(inode->i_mode))
2392 invalidate_mapping_pages(inode->i_mapping, 0, -1);
2394 extern int invalidate_inode_pages2(struct address_space *mapping);
2395 extern int invalidate_inode_pages2_range(struct address_space *mapping,
2396 pgoff_t start, pgoff_t end);
2397 extern int write_inode_now(struct inode *, int);
2398 extern int filemap_fdatawrite(struct address_space *);
2399 extern int filemap_flush(struct address_space *);
2400 extern int filemap_fdatawait(struct address_space *);
2401 extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
2403 extern int filemap_write_and_wait(struct address_space *mapping);
2404 extern int filemap_write_and_wait_range(struct address_space *mapping,
2405 loff_t lstart, loff_t lend);
2406 extern int __filemap_fdatawrite_range(struct address_space *mapping,
2407 loff_t start, loff_t end, int sync_mode);
2408 extern int filemap_fdatawrite_range(struct address_space *mapping,
2409 loff_t start, loff_t end);
2411 extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
2413 extern int vfs_fsync(struct file *file, int datasync);
2414 extern int generic_write_sync(struct file *file, loff_t pos, loff_t count);
2415 extern void emergency_sync(void);
2416 extern void emergency_remount(void);
2418 extern sector_t bmap(struct inode *, sector_t);
2420 extern int notify_change(struct dentry *, struct iattr *);
2421 extern int inode_permission(struct inode *, int);
2422 extern int generic_permission(struct inode *, int);
2424 static inline bool execute_ok(struct inode *inode)
2426 return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
2430 * get_write_access() gets write permission for a file.
2431 * put_write_access() releases this write permission.
2432 * This is used for regular files.
2433 * We cannot support write (and maybe mmap read-write shared) accesses and
2434 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2435 * can have the following values:
2436 * 0: no writers, no VM_DENYWRITE mappings
2437 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2438 * > 0: (i_writecount) users are writing to the file.
2440 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2441 * except for the cases where we don't hold i_writecount yet. Then we need to
2442 * use {get,deny}_write_access() - these functions check the sign and refuse
2443 * to do the change if sign is wrong.
2445 static inline int get_write_access(struct inode *inode)
2447 return atomic_inc_unless_negative(&inode->i_writecount) ? 0 : -ETXTBSY;
2449 static inline int deny_write_access(struct file *file)
2451 struct inode *inode = file->f_path.dentry->d_inode;
2452 return atomic_dec_unless_positive(&inode->i_writecount) ? 0 : -ETXTBSY;
2454 static inline void put_write_access(struct inode * inode)
2456 atomic_dec(&inode->i_writecount);
2458 static inline void allow_write_access(struct file *file)
2461 atomic_inc(&file->f_path.dentry->d_inode->i_writecount);
2464 static inline void i_readcount_dec(struct inode *inode)
2466 BUG_ON(!atomic_read(&inode->i_readcount));
2467 atomic_dec(&inode->i_readcount);
2469 static inline void i_readcount_inc(struct inode *inode)
2471 atomic_inc(&inode->i_readcount);
2474 static inline void i_readcount_dec(struct inode *inode)
2478 static inline void i_readcount_inc(struct inode *inode)
2483 extern int do_pipe_flags(int *, int);
2485 extern int kernel_read(struct file *, loff_t, char *, unsigned long);
2486 extern struct file * open_exec(const char *);
2488 /* fs/dcache.c -- generic fs support functions */
2489 extern int is_subdir(struct dentry *, struct dentry *);
2490 extern int path_is_under(struct path *, struct path *);
2491 extern ino_t find_inode_number(struct dentry *, struct qstr *);
2493 #include <linux/err.h>
2495 /* needed for stackable file system support */
2496 extern loff_t default_llseek(struct file *file, loff_t offset, int origin);
2498 extern loff_t vfs_llseek(struct file *file, loff_t offset, int origin);
2500 extern int inode_init_always(struct super_block *, struct inode *);
2501 extern void inode_init_once(struct inode *);
2502 extern void address_space_init_once(struct address_space *mapping);
2503 extern void ihold(struct inode * inode);
2504 extern void iput(struct inode *);
2505 extern struct inode * igrab(struct inode *);
2506 extern ino_t iunique(struct super_block *, ino_t);
2507 extern int inode_needs_sync(struct inode *inode);
2508 extern int generic_delete_inode(struct inode *inode);
2509 static inline int generic_drop_inode(struct inode *inode)
2511 return !inode->i_nlink || inode_unhashed(inode);
2514 extern struct inode *ilookup5_nowait(struct super_block *sb,
2515 unsigned long hashval, int (*test)(struct inode *, void *),
2517 extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
2518 int (*test)(struct inode *, void *), void *data);
2519 extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
2521 extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
2522 extern struct inode * iget_locked(struct super_block *, unsigned long);
2523 extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
2524 extern int insert_inode_locked(struct inode *);
2525 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2526 extern void lockdep_annotate_inode_mutex_key(struct inode *inode);
2528 static inline void lockdep_annotate_inode_mutex_key(struct inode *inode) { };
2530 extern void unlock_new_inode(struct inode *);
2531 extern unsigned int get_next_ino(void);
2533 extern void __iget(struct inode * inode);
2534 extern void iget_failed(struct inode *);
2535 extern void clear_inode(struct inode *);
2536 extern void __destroy_inode(struct inode *);
2537 extern struct inode *new_inode_pseudo(struct super_block *sb);
2538 extern struct inode *new_inode(struct super_block *sb);
2539 extern void free_inode_nonrcu(struct inode *inode);
2540 extern int should_remove_suid(struct dentry *);
2541 extern int file_remove_suid(struct file *);
2543 extern void __insert_inode_hash(struct inode *, unsigned long hashval);
2544 static inline void insert_inode_hash(struct inode *inode)
2546 __insert_inode_hash(inode, inode->i_ino);
2549 extern void __remove_inode_hash(struct inode *);
2550 static inline void remove_inode_hash(struct inode *inode)
2552 if (!inode_unhashed(inode))
2553 __remove_inode_hash(inode);
2556 extern void inode_sb_list_add(struct inode *inode);
2559 extern void submit_bio(int, struct bio *);
2560 extern int bdev_read_only(struct block_device *);
2562 extern int set_blocksize(struct block_device *, int);
2563 extern int sb_set_blocksize(struct super_block *, int);
2564 extern int sb_min_blocksize(struct super_block *, int);
2566 extern int generic_file_mmap(struct file *, struct vm_area_struct *);
2567 extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
2568 extern int generic_file_remap_pages(struct vm_area_struct *, unsigned long addr,
2569 unsigned long size, pgoff_t pgoff);
2570 extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
2571 int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
2572 extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2573 extern ssize_t __generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long,
2575 extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2576 extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
2577 unsigned long *, loff_t, loff_t *, size_t, size_t);
2578 extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
2579 unsigned long, loff_t, loff_t *, size_t, ssize_t);
2580 extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
2581 extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
2582 extern int generic_segment_checks(const struct iovec *iov,
2583 unsigned long *nr_segs, size_t *count, int access_flags);
2585 /* fs/block_dev.c */
2586 extern ssize_t blkdev_aio_read(struct kiocb *iocb, const struct iovec *iov,
2587 unsigned long nr_segs, loff_t pos);
2588 extern ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
2589 unsigned long nr_segs, loff_t pos);
2590 extern int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
2592 extern void block_sync_page(struct page *page);
2595 extern ssize_t generic_file_splice_read(struct file *, loff_t *,
2596 struct pipe_inode_info *, size_t, unsigned int);
2597 extern ssize_t default_file_splice_read(struct file *, loff_t *,
2598 struct pipe_inode_info *, size_t, unsigned int);
2599 extern ssize_t generic_file_splice_write(struct pipe_inode_info *,
2600 struct file *, loff_t *, size_t, unsigned int);
2601 extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
2602 struct file *out, loff_t *, size_t len, unsigned int flags);
2603 extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
2604 size_t len, unsigned int flags);
2607 file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
2608 extern loff_t noop_llseek(struct file *file, loff_t offset, int origin);
2609 extern loff_t no_llseek(struct file *file, loff_t offset, int origin);
2610 extern loff_t generic_file_llseek(struct file *file, loff_t offset, int origin);
2611 extern loff_t generic_file_llseek_size(struct file *file, loff_t offset,
2612 int origin, loff_t maxsize, loff_t eof);
2613 extern int generic_file_open(struct inode * inode, struct file * filp);
2614 extern int nonseekable_open(struct inode * inode, struct file * filp);
2616 #ifdef CONFIG_FS_XIP
2617 extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
2619 extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
2620 extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
2621 size_t len, loff_t *ppos);
2622 extern int xip_truncate_page(struct address_space *mapping, loff_t from);
2624 static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
2631 typedef void (dio_submit_t)(int rw, struct bio *bio, struct inode *inode,
2632 loff_t file_offset);
2635 /* need locking between buffered and direct access */
2638 /* filesystem does not support filling holes */
2639 DIO_SKIP_HOLES = 0x02,
2642 void dio_end_io(struct bio *bio, int error);
2644 ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
2645 struct block_device *bdev, const struct iovec *iov, loff_t offset,
2646 unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
2647 dio_submit_t submit_io, int flags);
2649 static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
2650 struct inode *inode, const struct iovec *iov, loff_t offset,
2651 unsigned long nr_segs, get_block_t get_block)
2653 return __blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
2654 offset, nr_segs, get_block, NULL, NULL,
2655 DIO_LOCKING | DIO_SKIP_HOLES);
2659 void inode_dio_wait(struct inode *inode);
2660 void inode_dio_done(struct inode *inode);
2662 extern const struct file_operations generic_ro_fops;
2664 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2666 extern int vfs_readlink(struct dentry *, char __user *, int, const char *);
2667 extern int vfs_follow_link(struct nameidata *, const char *);
2668 extern int page_readlink(struct dentry *, char __user *, int);
2669 extern void *page_follow_link_light(struct dentry *, struct nameidata *);
2670 extern void page_put_link(struct dentry *, struct nameidata *, void *);
2671 extern int __page_symlink(struct inode *inode, const char *symname, int len,
2673 extern int page_symlink(struct inode *inode, const char *symname, int len);
2674 extern const struct inode_operations page_symlink_inode_operations;
2675 extern int generic_readlink(struct dentry *, char __user *, int);
2676 extern void generic_fillattr(struct inode *, struct kstat *);
2677 extern int vfs_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2678 void __inode_add_bytes(struct inode *inode, loff_t bytes);
2679 void inode_add_bytes(struct inode *inode, loff_t bytes);
2680 void inode_sub_bytes(struct inode *inode, loff_t bytes);
2681 loff_t inode_get_bytes(struct inode *inode);
2682 void inode_set_bytes(struct inode *inode, loff_t bytes);
2684 extern int vfs_readdir(struct file *, filldir_t, void *);
2686 extern int vfs_stat(const char __user *, struct kstat *);
2687 extern int vfs_lstat(const char __user *, struct kstat *);
2688 extern int vfs_fstat(unsigned int, struct kstat *);
2689 extern int vfs_fstatat(int , const char __user *, struct kstat *, int);
2691 extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
2693 extern int __generic_block_fiemap(struct inode *inode,
2694 struct fiemap_extent_info *fieinfo,
2695 loff_t start, loff_t len,
2696 get_block_t *get_block);
2697 extern int generic_block_fiemap(struct inode *inode,
2698 struct fiemap_extent_info *fieinfo, u64 start,
2699 u64 len, get_block_t *get_block);
2701 extern void get_filesystem(struct file_system_type *fs);
2702 extern void put_filesystem(struct file_system_type *fs);
2703 extern struct file_system_type *get_fs_type(const char *name);
2704 extern struct super_block *get_super(struct block_device *);
2705 extern struct super_block *get_super_thawed(struct block_device *);
2706 extern struct super_block *get_active_super(struct block_device *bdev);
2707 extern void drop_super(struct super_block *sb);
2708 extern void iterate_supers(void (*)(struct super_block *, void *), void *);
2709 extern void iterate_supers_type(struct file_system_type *,
2710 void (*)(struct super_block *, void *), void *);
2712 extern int dcache_dir_open(struct inode *, struct file *);
2713 extern int dcache_dir_close(struct inode *, struct file *);
2714 extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
2715 extern int dcache_readdir(struct file *, void *, filldir_t);
2716 extern int simple_setattr(struct dentry *, struct iattr *);
2717 extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2718 extern int simple_statfs(struct dentry *, struct kstatfs *);
2719 extern int simple_open(struct inode *inode, struct file *file);
2720 extern int simple_link(struct dentry *, struct inode *, struct dentry *);
2721 extern int simple_unlink(struct inode *, struct dentry *);
2722 extern int simple_rmdir(struct inode *, struct dentry *);
2723 extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
2724 extern int noop_fsync(struct file *, loff_t, loff_t, int);
2725 extern int simple_empty(struct dentry *);
2726 extern int simple_readpage(struct file *file, struct page *page);
2727 extern int simple_write_begin(struct file *file, struct address_space *mapping,
2728 loff_t pos, unsigned len, unsigned flags,
2729 struct page **pagep, void **fsdata);
2730 extern int simple_write_end(struct file *file, struct address_space *mapping,
2731 loff_t pos, unsigned len, unsigned copied,
2732 struct page *page, void *fsdata);
2734 extern struct dentry *simple_lookup(struct inode *, struct dentry *, unsigned int flags);
2735 extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
2736 extern const struct file_operations simple_dir_operations;
2737 extern const struct inode_operations simple_dir_inode_operations;
2738 struct tree_descr { char *name; const struct file_operations *ops; int mode; };
2739 struct dentry *d_alloc_name(struct dentry *, const char *);
2740 extern int simple_fill_super(struct super_block *, unsigned long, struct tree_descr *);
2741 extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
2742 extern void simple_release_fs(struct vfsmount **mount, int *count);
2744 extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
2745 loff_t *ppos, const void *from, size_t available);
2746 extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
2747 const void __user *from, size_t count);
2749 extern int generic_file_fsync(struct file *, loff_t, loff_t, int);
2751 extern int generic_check_addressable(unsigned, u64);
2753 #ifdef CONFIG_MIGRATION
2754 extern int buffer_migrate_page(struct address_space *,
2755 struct page *, struct page *,
2758 #define buffer_migrate_page NULL
2761 extern int inode_change_ok(const struct inode *, struct iattr *);
2762 extern int inode_newsize_ok(const struct inode *, loff_t offset);
2763 extern void setattr_copy(struct inode *inode, const struct iattr *attr);
2765 extern int file_update_time(struct file *file);
2767 extern int generic_show_options(struct seq_file *m, struct dentry *root);
2768 extern void save_mount_options(struct super_block *sb, char *options);
2769 extern void replace_mount_options(struct super_block *sb, char *options);
2771 static inline ino_t parent_ino(struct dentry *dentry)
2776 * Don't strictly need d_lock here? If the parent ino could change
2777 * then surely we'd have a deeper race in the caller?
2779 spin_lock(&dentry->d_lock);
2780 res = dentry->d_parent->d_inode->i_ino;
2781 spin_unlock(&dentry->d_lock);
2785 /* Transaction based IO helpers */
2788 * An argresp is stored in an allocated page and holds the
2789 * size of the argument or response, along with its content
2791 struct simple_transaction_argresp {
2796 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2798 char *simple_transaction_get(struct file *file, const char __user *buf,
2800 ssize_t simple_transaction_read(struct file *file, char __user *buf,
2801 size_t size, loff_t *pos);
2802 int simple_transaction_release(struct inode *inode, struct file *file);
2804 void simple_transaction_set(struct file *file, size_t n);
2807 * simple attribute files
2809 * These attributes behave similar to those in sysfs:
2811 * Writing to an attribute immediately sets a value, an open file can be
2812 * written to multiple times.
2814 * Reading from an attribute creates a buffer from the value that might get
2815 * read with multiple read calls. When the attribute has been read
2816 * completely, no further read calls are possible until the file is opened
2819 * All attributes contain a text representation of a numeric value
2820 * that are accessed with the get() and set() functions.
2822 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2823 static int __fops ## _open(struct inode *inode, struct file *file) \
2825 __simple_attr_check_format(__fmt, 0ull); \
2826 return simple_attr_open(inode, file, __get, __set, __fmt); \
2828 static const struct file_operations __fops = { \
2829 .owner = THIS_MODULE, \
2830 .open = __fops ## _open, \
2831 .release = simple_attr_release, \
2832 .read = simple_attr_read, \
2833 .write = simple_attr_write, \
2834 .llseek = generic_file_llseek, \
2837 static inline __printf(1, 2)
2838 void __simple_attr_check_format(const char *fmt, ...)
2840 /* don't do anything, just let the compiler check the arguments; */
2843 int simple_attr_open(struct inode *inode, struct file *file,
2844 int (*get)(void *, u64 *), int (*set)(void *, u64),
2846 int simple_attr_release(struct inode *inode, struct file *file);
2847 ssize_t simple_attr_read(struct file *file, char __user *buf,
2848 size_t len, loff_t *ppos);
2849 ssize_t simple_attr_write(struct file *file, const char __user *buf,
2850 size_t len, loff_t *ppos);
2853 int proc_nr_files(struct ctl_table *table, int write,
2854 void __user *buffer, size_t *lenp, loff_t *ppos);
2855 int proc_nr_dentry(struct ctl_table *table, int write,
2856 void __user *buffer, size_t *lenp, loff_t *ppos);
2857 int proc_nr_inodes(struct ctl_table *table, int write,
2858 void __user *buffer, size_t *lenp, loff_t *ppos);
2859 int __init get_filesystem_list(char *buf);
2861 #define __FMODE_EXEC ((__force int) FMODE_EXEC)
2862 #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
2864 #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
2865 #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
2866 (flag & __FMODE_NONOTIFY)))
2868 static inline int is_sxid(umode_t mode)
2870 return (mode & S_ISUID) || ((mode & S_ISGID) && (mode & S_IXGRP));
2873 static inline void inode_has_no_xattr(struct inode *inode)
2875 if (!is_sxid(inode->i_mode) && (inode->i_sb->s_flags & MS_NOSEC))
2876 inode->i_flags |= S_NOSEC;
2879 #endif /* __KERNEL__ */
2880 #endif /* _LINUX_FS_H */