5 * This file has definitions for some important file table
9 #include <linux/limits.h>
10 #include <linux/ioctl.h>
11 #include <linux/blk_types.h>
12 #include <linux/types.h>
15 * It's silly to have NR_OPEN bigger than NR_FILE, but you can change
16 * the file limit at runtime and only root can increase the per-process
17 * nr_file rlimit, so it's safe to set up a ridiculously high absolute
18 * upper limit on files-per-process.
20 * Some programs (notably those using select()) may have to be
21 * recompiled to take full advantage of the new limits..
24 /* Fixed constants first: */
26 #define INR_OPEN_CUR 1024 /* Initial setting for nfile rlimits */
27 #define INR_OPEN_MAX 4096 /* Hard limit for nfile rlimits */
29 #define BLOCK_SIZE_BITS 10
30 #define BLOCK_SIZE (1<<BLOCK_SIZE_BITS)
32 #define SEEK_SET 0 /* seek relative to beginning of file */
33 #define SEEK_CUR 1 /* seek relative to current file position */
34 #define SEEK_END 2 /* seek relative to end of file */
35 #define SEEK_DATA 3 /* seek to the next data */
36 #define SEEK_HOLE 4 /* seek to the next hole */
37 #define SEEK_MAX SEEK_HOLE
45 /* And dynamically-tunable limits and defaults: */
46 struct files_stat_struct {
47 unsigned long nr_files; /* read only */
48 unsigned long nr_free_files; /* read only */
49 unsigned long max_files; /* tunable */
52 struct inodes_stat_t {
55 int dummy[5]; /* padding for sysctl ABI compatibility */
59 #define NR_FILE 8192 /* this can well be larger on a larger system */
61 #define MAY_EXEC 0x00000001
62 #define MAY_WRITE 0x00000002
63 #define MAY_READ 0x00000004
64 #define MAY_APPEND 0x00000008
65 #define MAY_ACCESS 0x00000010
66 #define MAY_OPEN 0x00000020
67 #define MAY_CHDIR 0x00000040
68 /* called from RCU mode, don't block */
69 #define MAY_NOT_BLOCK 0x00000080
72 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
73 * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
76 /* file is open for reading */
77 #define FMODE_READ ((__force fmode_t)0x1)
78 /* file is open for writing */
79 #define FMODE_WRITE ((__force fmode_t)0x2)
80 /* file is seekable */
81 #define FMODE_LSEEK ((__force fmode_t)0x4)
82 /* file can be accessed using pread */
83 #define FMODE_PREAD ((__force fmode_t)0x8)
84 /* file can be accessed using pwrite */
85 #define FMODE_PWRITE ((__force fmode_t)0x10)
86 /* File is opened for execution with sys_execve / sys_uselib */
87 #define FMODE_EXEC ((__force fmode_t)0x20)
88 /* File is opened with O_NDELAY (only set for block devices) */
89 #define FMODE_NDELAY ((__force fmode_t)0x40)
90 /* File is opened with O_EXCL (only set for block devices) */
91 #define FMODE_EXCL ((__force fmode_t)0x80)
92 /* File is opened using open(.., 3, ..) and is writeable only for ioctls
93 (specialy hack for floppy.c) */
94 #define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
95 /* 32bit hashes as llseek() offset (for directories) */
96 #define FMODE_32BITHASH ((__force fmode_t)0x200)
97 /* 64bit hashes as llseek() offset (for directories) */
98 #define FMODE_64BITHASH ((__force fmode_t)0x400)
101 * Don't update ctime and mtime.
103 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
104 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
106 #define FMODE_NOCMTIME ((__force fmode_t)0x800)
108 /* Expect random access pattern */
109 #define FMODE_RANDOM ((__force fmode_t)0x1000)
111 /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
112 #define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
114 /* File is opened with O_PATH; almost nothing can be done with it */
115 #define FMODE_PATH ((__force fmode_t)0x4000)
117 /* File was opened by fanotify and shouldn't generate fanotify events */
118 #define FMODE_NONOTIFY ((__force fmode_t)0x1000000)
121 * The below are the various read and write types that we support. Some of
122 * them include behavioral modifiers that send information down to the
123 * block layer and IO scheduler. Terminology:
125 * The block layer uses device plugging to defer IO a little bit, in
126 * the hope that we will see more IO very shortly. This increases
127 * coalescing of adjacent IO and thus reduces the number of IOs we
128 * have to send to the device. It also allows for better queuing,
129 * if the IO isn't mergeable. If the caller is going to be waiting
130 * for the IO, then he must ensure that the device is unplugged so
131 * that the IO is dispatched to the driver.
133 * All IO is handled async in Linux. This is fine for background
134 * writes, but for reads or writes that someone waits for completion
135 * on, we want to notify the block layer and IO scheduler so that they
136 * know about it. That allows them to make better scheduling
137 * decisions. So when the below references 'sync' and 'async', it
138 * is referencing this priority hint.
140 * With that in mind, the available types are:
142 * READ A normal read operation. Device will be plugged.
143 * READ_SYNC A synchronous read. Device is not plugged, caller can
144 * immediately wait on this read without caring about
146 * READA Used for read-ahead operations. Lower priority, and the
147 * block layer could (in theory) choose to ignore this
148 * request if it runs into resource problems.
149 * WRITE A normal async write. Device will be plugged.
150 * WRITE_SYNC Synchronous write. Identical to WRITE, but passes down
151 * the hint that someone will be waiting on this IO
152 * shortly. The write equivalent of READ_SYNC.
153 * WRITE_ODIRECT Special case write for O_DIRECT only.
154 * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
155 * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
156 * non-volatile media on completion.
157 * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
158 * by a cache flush and data is guaranteed to be on
159 * non-volatile media on completion.
162 #define RW_MASK REQ_WRITE
163 #define RWA_MASK REQ_RAHEAD
166 #define WRITE RW_MASK
167 #define READA RWA_MASK
169 #define READ_SYNC (READ | REQ_SYNC)
170 #define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE)
171 #define WRITE_ODIRECT (WRITE | REQ_SYNC)
172 #define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
173 #define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
174 #define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
180 /* public flags for file_system_type */
181 #define FS_REQUIRES_DEV 1
182 #define FS_BINARY_MOUNTDATA 2
183 #define FS_HAS_SUBTYPE 4
184 #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
185 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move()
186 * during rename() internally.
190 * These are the fs-independent mount-flags: up to 32 flags are supported
192 #define MS_RDONLY 1 /* Mount read-only */
193 #define MS_NOSUID 2 /* Ignore suid and sgid bits */
194 #define MS_NODEV 4 /* Disallow access to device special files */
195 #define MS_NOEXEC 8 /* Disallow program execution */
196 #define MS_SYNCHRONOUS 16 /* Writes are synced at once */
197 #define MS_REMOUNT 32 /* Alter flags of a mounted FS */
198 #define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
199 #define MS_DIRSYNC 128 /* Directory modifications are synchronous */
200 #define MS_NOATIME 1024 /* Do not update access times. */
201 #define MS_NODIRATIME 2048 /* Do not update directory access times */
205 #define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
206 MS_VERBOSE is deprecated. */
207 #define MS_SILENT 32768
208 #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
209 #define MS_UNBINDABLE (1<<17) /* change to unbindable */
210 #define MS_PRIVATE (1<<18) /* change to private */
211 #define MS_SLAVE (1<<19) /* change to slave */
212 #define MS_SHARED (1<<20) /* change to shared */
213 #define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */
214 #define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
215 #define MS_I_VERSION (1<<23) /* Update inode I_version field */
216 #define MS_STRICTATIME (1<<24) /* Always perform atime updates */
217 #define MS_NOSEC (1<<28)
218 #define MS_BORN (1<<29)
219 #define MS_ACTIVE (1<<30)
220 #define MS_NOUSER (1<<31)
223 * Superblock flags that can be altered by MS_REMOUNT
225 #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK|MS_I_VERSION)
228 * Old magic mount flag and mask
230 #define MS_MGC_VAL 0xC0ED0000
231 #define MS_MGC_MSK 0xffff0000
233 /* Inode flags - they have nothing to superblock flags now */
235 #define S_SYNC 1 /* Writes are synced at once */
236 #define S_NOATIME 2 /* Do not update access times */
237 #define S_APPEND 4 /* Append-only file */
238 #define S_IMMUTABLE 8 /* Immutable file */
239 #define S_DEAD 16 /* removed, but still open directory */
240 #define S_NOQUOTA 32 /* Inode is not counted to quota */
241 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
242 #define S_NOCMTIME 128 /* Do not update file c/mtime */
243 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
244 #define S_PRIVATE 512 /* Inode is fs-internal */
245 #define S_IMA 1024 /* Inode has an associated IMA struct */
246 #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
247 #define S_NOSEC 4096 /* no suid or xattr security attributes */
250 * Note that nosuid etc flags are inode-specific: setting some file-system
251 * flags just means all the inodes inherit those flags by default. It might be
252 * possible to override it selectively if you really wanted to with some
253 * ioctl() that is not currently implemented.
255 * Exception: MS_RDONLY is always applied to the entire file system.
257 * Unfortunately, it is possible to change a filesystems flags with it mounted
258 * with files in use. This means that all of the inodes will not have their
259 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
260 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
262 #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
264 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
265 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
266 ((inode)->i_flags & S_SYNC))
267 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
268 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
269 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
270 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
271 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
273 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
274 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
275 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
276 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
278 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
279 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
280 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
281 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
282 #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
283 #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
284 #define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
286 /* the read-only stuff doesn't really belong here, but any other place is
287 probably as bad and I don't want to create yet another include file. */
289 #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
290 #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
291 #define BLKRRPART _IO(0x12,95) /* re-read partition table */
292 #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
293 #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
294 #define BLKRASET _IO(0x12,98) /* set read ahead for block device */
295 #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
296 #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
297 #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
298 #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
299 #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
300 #define BLKSSZGET _IO(0x12,104)/* get block device sector size */
302 #define BLKPG _IO(0x12,105)/* See blkpg.h */
304 /* Some people are morons. Do not use sizeof! */
306 #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
307 #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
308 /* This was here just to show that the number is taken -
309 probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
311 /* A jump here: 108-111 have been used for various private purposes. */
312 #define BLKBSZGET _IOR(0x12,112,size_t)
313 #define BLKBSZSET _IOW(0x12,113,size_t)
314 #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
315 #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
316 #define BLKTRACESTART _IO(0x12,116)
317 #define BLKTRACESTOP _IO(0x12,117)
318 #define BLKTRACETEARDOWN _IO(0x12,118)
319 #define BLKDISCARD _IO(0x12,119)
320 #define BLKIOMIN _IO(0x12,120)
321 #define BLKIOOPT _IO(0x12,121)
322 #define BLKALIGNOFF _IO(0x12,122)
323 #define BLKPBSZGET _IO(0x12,123)
324 #define BLKDISCARDZEROES _IO(0x12,124)
325 #define BLKSECDISCARD _IO(0x12,125)
327 #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
328 #define FIBMAP _IO(0x00,1) /* bmap access */
329 #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
330 #define FIFREEZE _IOWR('X', 119, int) /* Freeze */
331 #define FITHAW _IOWR('X', 120, int) /* Thaw */
332 #define FITRIM _IOWR('X', 121, struct fstrim_range) /* Trim */
334 #define FS_IOC_GETFLAGS _IOR('f', 1, long)
335 #define FS_IOC_SETFLAGS _IOW('f', 2, long)
336 #define FS_IOC_GETVERSION _IOR('v', 1, long)
337 #define FS_IOC_SETVERSION _IOW('v', 2, long)
338 #define FS_IOC_FIEMAP _IOWR('f', 11, struct fiemap)
339 #define FS_IOC32_GETFLAGS _IOR('f', 1, int)
340 #define FS_IOC32_SETFLAGS _IOW('f', 2, int)
341 #define FS_IOC32_GETVERSION _IOR('v', 1, int)
342 #define FS_IOC32_SETVERSION _IOW('v', 2, int)
345 * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS)
347 #define FS_SECRM_FL 0x00000001 /* Secure deletion */
348 #define FS_UNRM_FL 0x00000002 /* Undelete */
349 #define FS_COMPR_FL 0x00000004 /* Compress file */
350 #define FS_SYNC_FL 0x00000008 /* Synchronous updates */
351 #define FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
352 #define FS_APPEND_FL 0x00000020 /* writes to file may only append */
353 #define FS_NODUMP_FL 0x00000040 /* do not dump file */
354 #define FS_NOATIME_FL 0x00000080 /* do not update atime */
355 /* Reserved for compression usage... */
356 #define FS_DIRTY_FL 0x00000100
357 #define FS_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
358 #define FS_NOCOMP_FL 0x00000400 /* Don't compress */
359 #define FS_ECOMPR_FL 0x00000800 /* Compression error */
360 /* End compression flags --- maybe not all used */
361 #define FS_BTREE_FL 0x00001000 /* btree format dir */
362 #define FS_INDEX_FL 0x00001000 /* hash-indexed directory */
363 #define FS_IMAGIC_FL 0x00002000 /* AFS directory */
364 #define FS_JOURNAL_DATA_FL 0x00004000 /* Reserved for ext3 */
365 #define FS_NOTAIL_FL 0x00008000 /* file tail should not be merged */
366 #define FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
367 #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
368 #define FS_EXTENT_FL 0x00080000 /* Extents */
369 #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */
370 #define FS_NOCOW_FL 0x00800000 /* Do not cow file */
371 #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */
373 #define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
374 #define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
377 #define SYNC_FILE_RANGE_WAIT_BEFORE 1
378 #define SYNC_FILE_RANGE_WRITE 2
379 #define SYNC_FILE_RANGE_WAIT_AFTER 4
383 #include <linux/linkage.h>
384 #include <linux/wait.h>
385 #include <linux/kdev_t.h>
386 #include <linux/dcache.h>
387 #include <linux/path.h>
388 #include <linux/stat.h>
389 #include <linux/cache.h>
390 #include <linux/list.h>
391 #include <linux/radix-tree.h>
392 #include <linux/prio_tree.h>
393 #include <linux/init.h>
394 #include <linux/pid.h>
395 #include <linux/mutex.h>
396 #include <linux/capability.h>
397 #include <linux/semaphore.h>
398 #include <linux/fiemap.h>
399 #include <linux/rculist_bl.h>
400 #include <linux/atomic.h>
401 #include <linux/shrinker.h>
403 #include <asm/byteorder.h>
405 struct export_operations;
411 struct pipe_inode_info;
412 struct poll_table_struct;
414 struct vm_area_struct;
418 extern void __init inode_init(void);
419 extern void __init inode_init_early(void);
420 extern void __init files_init(unsigned long);
422 extern struct files_stat_struct files_stat;
423 extern unsigned long get_max_files(void);
424 extern int sysctl_nr_open;
425 extern struct inodes_stat_t inodes_stat;
426 extern int leases_enable, lease_break_time;
429 typedef int (get_block_t)(struct inode *inode, sector_t iblock,
430 struct buffer_head *bh_result, int create);
431 typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
432 ssize_t bytes, void *private, int ret,
436 * Attribute flags. These should be or-ed together to figure out what
439 #define ATTR_MODE (1 << 0)
440 #define ATTR_UID (1 << 1)
441 #define ATTR_GID (1 << 2)
442 #define ATTR_SIZE (1 << 3)
443 #define ATTR_ATIME (1 << 4)
444 #define ATTR_MTIME (1 << 5)
445 #define ATTR_CTIME (1 << 6)
446 #define ATTR_ATIME_SET (1 << 7)
447 #define ATTR_MTIME_SET (1 << 8)
448 #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
449 #define ATTR_ATTR_FLAG (1 << 10)
450 #define ATTR_KILL_SUID (1 << 11)
451 #define ATTR_KILL_SGID (1 << 12)
452 #define ATTR_FILE (1 << 13)
453 #define ATTR_KILL_PRIV (1 << 14)
454 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
455 #define ATTR_TIMES_SET (1 << 16)
458 * This is the Inode Attributes structure, used for notify_change(). It
459 * uses the above definitions as flags, to know which values have changed.
460 * Also, in this manner, a Filesystem can look at only the values it cares
461 * about. Basically, these are the attributes that the VFS layer can
462 * request to change from the FS layer.
464 * Derek Atkins <warlord@MIT.EDU> 94-10-20
467 unsigned int ia_valid;
472 struct timespec ia_atime;
473 struct timespec ia_mtime;
474 struct timespec ia_ctime;
477 * Not an attribute, but an auxiliary info for filesystems wanting to
478 * implement an ftruncate() like method. NOTE: filesystem should
479 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
481 struct file *ia_file;
485 * Includes for diskquotas.
487 #include <linux/quota.h>
490 * Maximum number of layers of fs stack. Needs to be limited to
491 * prevent kernel stack overflow
493 #define FILESYSTEM_MAX_STACK_DEPTH 2
496 * enum positive_aop_returns - aop return codes with specific semantics
498 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
499 * completed, that the page is still locked, and
500 * should be considered active. The VM uses this hint
501 * to return the page to the active list -- it won't
502 * be a candidate for writeback again in the near
503 * future. Other callers must be careful to unlock
504 * the page if they get this return. Returned by
507 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
508 * unlocked it and the page might have been truncated.
509 * The caller should back up to acquiring a new page and
510 * trying again. The aop will be taking reasonable
511 * precautions not to livelock. If the caller held a page
512 * reference, it should drop it before retrying. Returned
515 * address_space_operation functions return these large constants to indicate
516 * special semantics to the caller. These are much larger than the bytes in a
517 * page to allow for functions that return the number of bytes operated on in a
521 enum positive_aop_returns {
522 AOP_WRITEPAGE_ACTIVATE = 0x80000,
523 AOP_TRUNCATED_PAGE = 0x80001,
526 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
527 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
528 #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
529 * helper code (eg buffer layer)
530 * to clear GFP_FS from alloc */
533 * oh the beauties of C type declarations.
536 struct address_space;
537 struct writeback_control;
541 const struct iovec *iov;
542 unsigned long nr_segs;
547 size_t iov_iter_copy_from_user_atomic(struct page *page,
548 struct iov_iter *i, unsigned long offset, size_t bytes);
549 size_t iov_iter_copy_from_user(struct page *page,
550 struct iov_iter *i, unsigned long offset, size_t bytes);
551 void iov_iter_advance(struct iov_iter *i, size_t bytes);
552 int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes);
553 size_t iov_iter_single_seg_count(struct iov_iter *i);
555 static inline void iov_iter_init(struct iov_iter *i,
556 const struct iovec *iov, unsigned long nr_segs,
557 size_t count, size_t written)
560 i->nr_segs = nr_segs;
562 i->count = count + written;
564 iov_iter_advance(i, written);
567 static inline size_t iov_iter_count(struct iov_iter *i)
573 * "descriptor" for what we're up to with a read.
574 * This allows us to use the same read code yet
575 * have multiple different users of the data that
576 * we read from a file.
578 * The simplest case just copies the data to user
591 typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
592 unsigned long, unsigned long);
594 struct address_space_operations {
595 int (*writepage)(struct page *page, struct writeback_control *wbc);
596 int (*readpage)(struct file *, struct page *);
598 /* Write back some dirty pages from this mapping. */
599 int (*writepages)(struct address_space *, struct writeback_control *);
601 /* Set a page dirty. Return true if this dirtied it */
602 int (*set_page_dirty)(struct page *page);
604 int (*readpages)(struct file *filp, struct address_space *mapping,
605 struct list_head *pages, unsigned nr_pages);
607 int (*write_begin)(struct file *, struct address_space *mapping,
608 loff_t pos, unsigned len, unsigned flags,
609 struct page **pagep, void **fsdata);
610 int (*write_end)(struct file *, struct address_space *mapping,
611 loff_t pos, unsigned len, unsigned copied,
612 struct page *page, void *fsdata);
614 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
615 sector_t (*bmap)(struct address_space *, sector_t);
616 void (*invalidatepage) (struct page *, unsigned long);
617 int (*releasepage) (struct page *, gfp_t);
618 void (*freepage)(struct page *);
619 ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
620 loff_t offset, unsigned long nr_segs);
621 int (*get_xip_mem)(struct address_space *, pgoff_t, int,
622 void **, unsigned long *);
624 * migrate the contents of a page to the specified target. If sync
625 * is false, it must not block.
627 int (*migratepage) (struct address_space *,
628 struct page *, struct page *, enum migrate_mode);
629 int (*launder_page) (struct page *);
630 int (*is_partially_uptodate) (struct page *, read_descriptor_t *,
632 int (*error_remove_page)(struct address_space *, struct page *);
635 extern const struct address_space_operations empty_aops;
638 * pagecache_write_begin/pagecache_write_end must be used by general code
639 * to write into the pagecache.
641 int pagecache_write_begin(struct file *, struct address_space *mapping,
642 loff_t pos, unsigned len, unsigned flags,
643 struct page **pagep, void **fsdata);
645 int pagecache_write_end(struct file *, struct address_space *mapping,
646 loff_t pos, unsigned len, unsigned copied,
647 struct page *page, void *fsdata);
649 struct backing_dev_info;
650 struct address_space {
651 struct inode *host; /* owner: inode, block_device */
652 struct radix_tree_root page_tree; /* radix tree of all pages */
653 spinlock_t tree_lock; /* and lock protecting it */
654 atomic_t i_mmap_writable;/* count VM_SHARED mappings */
655 struct prio_tree_root i_mmap; /* tree of private and shared mappings */
656 struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
657 struct mutex i_mmap_mutex; /* protect tree, count, list */
658 /* Protected by tree_lock together with the radix tree */
659 unsigned long nrpages; /* number of total pages */
660 pgoff_t writeback_index;/* writeback starts here */
661 const struct address_space_operations *a_ops; /* methods */
662 unsigned long flags; /* error bits/gfp mask */
663 struct backing_dev_info *backing_dev_info; /* device readahead, etc */
664 spinlock_t private_lock; /* for use by the address_space */
665 struct list_head private_list; /* ditto */
666 struct address_space *assoc_mapping; /* ditto */
667 } __attribute__((aligned(sizeof(long))));
669 * On most architectures that alignment is already the case; but
670 * must be enforced here for CRIS, to let the least significant bit
671 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
674 struct block_device {
675 dev_t bd_dev; /* not a kdev_t - it's a search key */
677 struct inode * bd_inode; /* will die */
678 struct super_block * bd_super;
679 struct mutex bd_mutex; /* open/close mutex */
680 struct list_head bd_inodes;
684 bool bd_write_holder;
686 struct list_head bd_holder_disks;
688 struct block_device * bd_contains;
689 unsigned bd_block_size;
690 struct hd_struct * bd_part;
691 /* number of times partitions within this device have been opened. */
692 unsigned bd_part_count;
694 struct gendisk * bd_disk;
695 struct list_head bd_list;
697 * Private data. You must have bd_claim'ed the block_device
698 * to use this. NOTE: bd_claim allows an owner to claim
699 * the same device multiple times, the owner must take special
700 * care to not mess up bd_private for that case.
702 unsigned long bd_private;
704 /* The counter of freeze processes */
705 int bd_fsfreeze_count;
706 /* Mutex for freeze */
707 struct mutex bd_fsfreeze_mutex;
711 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
714 #define PAGECACHE_TAG_DIRTY 0
715 #define PAGECACHE_TAG_WRITEBACK 1
716 #define PAGECACHE_TAG_TOWRITE 2
718 int mapping_tagged(struct address_space *mapping, int tag);
721 * Might pages of this file be mapped into userspace?
723 static inline int mapping_mapped(struct address_space *mapping)
725 return !prio_tree_empty(&mapping->i_mmap) ||
726 !list_empty(&mapping->i_mmap_nonlinear);
730 * Might pages of this file have been modified in userspace?
731 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
732 * marks vma as VM_SHARED if it is shared, and the file was opened for
733 * writing i.e. vma may be mprotected writable even if now readonly.
735 * If i_mmap_writable is negative, no new writable mappings are allowed. You
736 * can only deny writable mappings, if none exists right now.
738 static inline int mapping_writably_mapped(struct address_space *mapping)
740 return atomic_read(&mapping->i_mmap_writable) > 0;
743 static inline int mapping_map_writable(struct address_space *mapping)
745 return atomic_inc_unless_negative(&mapping->i_mmap_writable) ?
749 static inline void mapping_unmap_writable(struct address_space *mapping)
751 atomic_dec(&mapping->i_mmap_writable);
754 static inline int mapping_deny_writable(struct address_space *mapping)
756 return atomic_dec_unless_positive(&mapping->i_mmap_writable) ?
760 static inline void mapping_allow_writable(struct address_space *mapping)
762 atomic_inc(&mapping->i_mmap_writable);
766 * Use sequence counter to get consistent i_size on 32-bit processors.
768 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
769 #include <linux/seqlock.h>
770 #define __NEED_I_SIZE_ORDERED
771 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
773 #define i_size_ordered_init(inode) do { } while (0)
777 #define ACL_NOT_CACHED ((void *)(-1))
779 #define IOP_FASTPERM 0x0001
780 #define IOP_LOOKUP 0x0002
781 #define IOP_NOFOLLOW 0x0004
784 * Keep mostly read-only and often accessed (especially for
785 * the RCU path lookup and 'stat' data) fields at the beginning
786 * of the 'struct inode'
790 unsigned short i_opflags;
793 unsigned int i_flags;
795 #ifdef CONFIG_FS_POSIX_ACL
796 struct posix_acl *i_acl;
797 struct posix_acl *i_default_acl;
800 const struct inode_operations *i_op;
801 struct super_block *i_sb;
802 struct address_space *i_mapping;
804 #ifdef CONFIG_SECURITY
808 /* Stat data, not accessed from path walking */
811 * Filesystems may only read i_nlink directly. They shall use the
812 * following functions for modification:
814 * (set|clear|inc|drop)_nlink
815 * inode_(inc|dec)_link_count
818 const unsigned int i_nlink;
819 unsigned int __i_nlink;
822 struct timespec i_atime;
823 struct timespec i_mtime;
824 struct timespec i_ctime;
825 spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
826 unsigned short i_bytes;
830 #ifdef __NEED_I_SIZE_ORDERED
831 seqcount_t i_size_seqcount;
835 unsigned long i_state;
836 struct mutex i_mutex;
838 unsigned long dirtied_when; /* jiffies of first dirtying */
840 struct hlist_node i_hash;
841 struct list_head i_wb_list; /* backing dev IO list */
842 struct list_head i_lru; /* inode LRU list */
843 struct list_head i_sb_list;
845 struct list_head i_dentry;
846 struct rcu_head i_rcu;
849 unsigned int i_blkbits;
851 atomic_t i_dio_count;
852 atomic_t i_writecount;
853 const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
854 struct file_lock *i_flock;
855 struct address_space i_data;
857 struct dquot *i_dquot[MAXQUOTAS];
859 struct list_head i_devices;
861 struct pipe_inode_info *i_pipe;
862 struct block_device *i_bdev;
868 #ifdef CONFIG_FSNOTIFY
869 __u32 i_fsnotify_mask; /* all events this inode cares about */
870 struct hlist_head i_fsnotify_marks;
874 atomic_t i_readcount; /* struct files open RO */
876 void *i_private; /* fs or device private pointer */
879 static inline int inode_unhashed(struct inode *inode)
881 return hlist_unhashed(&inode->i_hash);
885 * inode->i_mutex nesting subclasses for the lock validator:
887 * 0: the object of the current VFS operation
892 * The locking order between these classes is
893 * parent -> child -> normal -> xattr -> quota
895 enum inode_i_mutex_lock_class
905 * NOTE: in a 32bit arch with a preemptable kernel and
906 * an UP compile the i_size_read/write must be atomic
907 * with respect to the local cpu (unlike with preempt disabled),
908 * but they don't need to be atomic with respect to other cpus like in
909 * true SMP (so they need either to either locally disable irq around
910 * the read or for example on x86 they can be still implemented as a
911 * cmpxchg8b without the need of the lock prefix). For SMP compiles
912 * and 64bit archs it makes no difference if preempt is enabled or not.
914 static inline loff_t i_size_read(const struct inode *inode)
916 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
921 seq = read_seqcount_begin(&inode->i_size_seqcount);
922 i_size = inode->i_size;
923 } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
925 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
929 i_size = inode->i_size;
933 return inode->i_size;
938 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
939 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
940 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
942 static inline void i_size_write(struct inode *inode, loff_t i_size)
944 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
946 write_seqcount_begin(&inode->i_size_seqcount);
947 inode->i_size = i_size;
948 write_seqcount_end(&inode->i_size_seqcount);
950 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
952 inode->i_size = i_size;
955 inode->i_size = i_size;
959 static inline unsigned iminor(const struct inode *inode)
961 return MINOR(inode->i_rdev);
964 static inline unsigned imajor(const struct inode *inode)
966 return MAJOR(inode->i_rdev);
969 extern struct block_device *I_BDEV(struct inode *inode);
972 rwlock_t lock; /* protects pid, uid, euid fields */
973 struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
974 enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
975 uid_t uid, euid; /* uid/euid of process setting the owner */
976 int signum; /* posix.1b rt signal to be delivered on IO */
980 * Track a single file's readahead state
982 struct file_ra_state {
983 pgoff_t start; /* where readahead started */
984 unsigned int size; /* # of readahead pages */
985 unsigned int async_size; /* do asynchronous readahead when
986 there are only # of pages ahead */
988 unsigned int ra_pages; /* Maximum readahead window */
989 unsigned int mmap_miss; /* Cache miss stat for mmap accesses */
990 loff_t prev_pos; /* Cache last read() position */
994 * Check if @index falls in the readahead windows.
996 static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
998 return (index >= ra->start &&
999 index < ra->start + ra->size);
1002 #define FILE_MNT_WRITE_TAKEN 1
1003 #define FILE_MNT_WRITE_RELEASED 2
1007 * fu_list becomes invalid after file_free is called and queued via
1008 * fu_rcuhead for RCU freeing
1011 struct list_head fu_list;
1012 struct rcu_head fu_rcuhead;
1015 #define f_dentry f_path.dentry
1016 #define f_vfsmnt f_path.mnt
1017 const struct file_operations *f_op;
1020 * Protects f_ep_links, f_flags, f_pos vs i_size in lseek SEEK_CUR.
1021 * Must not be taken from IRQ context.
1027 atomic_long_t f_count;
1028 unsigned int f_flags;
1031 struct fown_struct f_owner;
1032 const struct cred *f_cred;
1033 struct file_ra_state f_ra;
1036 #ifdef CONFIG_SECURITY
1039 /* needed for tty driver, and maybe others */
1043 /* Used by fs/eventpoll.c to link all the hooks to this file */
1044 struct list_head f_ep_links;
1045 struct list_head f_tfile_llink;
1046 #endif /* #ifdef CONFIG_EPOLL */
1047 struct address_space *f_mapping;
1048 #ifdef CONFIG_DEBUG_WRITECOUNT
1049 unsigned long f_mnt_write_state;
1053 struct file_handle {
1056 /* file identifier */
1057 unsigned char f_handle[0];
1060 static inline struct file *get_file(struct file *f)
1062 atomic_long_inc(&f->f_count);
1065 #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
1066 #define file_count(x) atomic_long_read(&(x)->f_count)
1068 #ifdef CONFIG_DEBUG_WRITECOUNT
1069 static inline void file_take_write(struct file *f)
1071 WARN_ON(f->f_mnt_write_state != 0);
1072 f->f_mnt_write_state = FILE_MNT_WRITE_TAKEN;
1074 static inline void file_release_write(struct file *f)
1076 f->f_mnt_write_state |= FILE_MNT_WRITE_RELEASED;
1078 static inline void file_reset_write(struct file *f)
1080 f->f_mnt_write_state = 0;
1082 static inline void file_check_state(struct file *f)
1085 * At this point, either both or neither of these bits
1088 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN);
1089 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_RELEASED);
1091 static inline int file_check_writeable(struct file *f)
1093 if (f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN)
1095 printk(KERN_WARNING "writeable file with no "
1096 "mnt_want_write()\n");
1100 #else /* !CONFIG_DEBUG_WRITECOUNT */
1101 static inline void file_take_write(struct file *filp) {}
1102 static inline void file_release_write(struct file *filp) {}
1103 static inline void file_reset_write(struct file *filp) {}
1104 static inline void file_check_state(struct file *filp) {}
1105 static inline int file_check_writeable(struct file *filp)
1109 #endif /* CONFIG_DEBUG_WRITECOUNT */
1111 #define MAX_NON_LFS ((1UL<<31) - 1)
1113 /* Page cache limit. The filesystems should put that into their s_maxbytes
1114 limits, otherwise bad things can happen in VM. */
1115 #if BITS_PER_LONG==32
1116 #define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
1117 #elif BITS_PER_LONG==64
1118 #define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
1123 #define FL_ACCESS 8 /* not trying to lock, just looking */
1124 #define FL_EXISTS 16 /* when unlocking, test for existence */
1125 #define FL_LEASE 32 /* lease held on this file */
1126 #define FL_CLOSE 64 /* unlock on close */
1127 #define FL_SLEEP 128 /* A blocking lock */
1128 #define FL_DOWNGRADE_PENDING 256 /* Lease is being downgraded */
1129 #define FL_UNLOCK_PENDING 512 /* Lease is being broken */
1132 * Special return value from posix_lock_file() and vfs_lock_file() for
1133 * asynchronous locking.
1135 #define FILE_LOCK_DEFERRED 1
1138 * The POSIX file lock owner is determined by
1139 * the "struct files_struct" in the thread group
1140 * (or NULL for no owner - BSD locks).
1142 * Lockd stuffs a "host" pointer into this.
1144 typedef struct files_struct *fl_owner_t;
1146 struct file_lock_operations {
1147 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
1148 void (*fl_release_private)(struct file_lock *);
1151 struct lock_manager_operations {
1152 int (*lm_compare_owner)(struct file_lock *, struct file_lock *);
1153 void (*lm_notify)(struct file_lock *); /* unblock callback */
1154 int (*lm_grant)(struct file_lock *, struct file_lock *, int);
1155 void (*lm_release_private)(struct file_lock *);
1156 void (*lm_break)(struct file_lock *);
1157 int (*lm_change)(struct file_lock **, int);
1160 struct lock_manager {
1161 struct list_head list;
1164 void locks_start_grace(struct lock_manager *);
1165 void locks_end_grace(struct lock_manager *);
1166 int locks_in_grace(void);
1168 /* that will die - we need it for nfs_lock_info */
1169 #include <linux/nfs_fs_i.h>
1172 struct file_lock *fl_next; /* singly linked list for this inode */
1173 struct list_head fl_link; /* doubly linked list of all locks */
1174 struct list_head fl_block; /* circular list of blocked processes */
1175 fl_owner_t fl_owner;
1176 unsigned int fl_flags;
1177 unsigned char fl_type;
1178 unsigned int fl_pid;
1179 struct pid *fl_nspid;
1180 wait_queue_head_t fl_wait;
1181 struct file *fl_file;
1185 struct fasync_struct * fl_fasync; /* for lease break notifications */
1186 /* for lease breaks: */
1187 unsigned long fl_break_time;
1188 unsigned long fl_downgrade_time;
1190 const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
1191 const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
1193 struct nfs_lock_info nfs_fl;
1194 struct nfs4_lock_info nfs4_fl;
1196 struct list_head link; /* link in AFS vnode's pending_locks list */
1197 int state; /* state of grant or error if -ve */
1202 /* The following constant reflects the upper bound of the file/locking space */
1204 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
1205 #define OFFSET_MAX INT_LIMIT(loff_t)
1206 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
1209 #include <linux/fcntl.h>
1211 extern void send_sigio(struct fown_struct *fown, int fd, int band);
1213 #ifdef CONFIG_FILE_LOCKING
1214 extern int fcntl_getlk(struct file *, struct flock __user *);
1215 extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
1216 struct flock __user *);
1218 #if BITS_PER_LONG == 32
1219 extern int fcntl_getlk64(struct file *, struct flock64 __user *);
1220 extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
1221 struct flock64 __user *);
1224 extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
1225 extern int fcntl_getlease(struct file *filp);
1228 void locks_free_lock(struct file_lock *fl);
1229 extern void locks_init_lock(struct file_lock *);
1230 extern struct file_lock * locks_alloc_lock(void);
1231 extern void locks_copy_lock(struct file_lock *, struct file_lock *);
1232 extern void __locks_copy_lock(struct file_lock *, const struct file_lock *);
1233 extern void locks_remove_posix(struct file *, fl_owner_t);
1234 extern void locks_remove_flock(struct file *);
1235 extern void locks_release_private(struct file_lock *);
1236 extern void posix_test_lock(struct file *, struct file_lock *);
1237 extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
1238 extern int posix_lock_file_wait(struct file *, struct file_lock *);
1239 extern int posix_unblock_lock(struct file *, struct file_lock *);
1240 extern int vfs_test_lock(struct file *, struct file_lock *);
1241 extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
1242 extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
1243 extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
1244 extern int __break_lease(struct inode *inode, unsigned int flags);
1245 extern void lease_get_mtime(struct inode *, struct timespec *time);
1246 extern int generic_setlease(struct file *, long, struct file_lock **);
1247 extern int vfs_setlease(struct file *, long, struct file_lock **);
1248 extern int lease_modify(struct file_lock **, int);
1249 extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
1250 extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
1251 extern void locks_delete_block(struct file_lock *waiter);
1252 extern void lock_flocks(void);
1253 extern void unlock_flocks(void);
1254 #else /* !CONFIG_FILE_LOCKING */
1255 static inline int fcntl_getlk(struct file *file, struct flock __user *user)
1260 static inline int fcntl_setlk(unsigned int fd, struct file *file,
1261 unsigned int cmd, struct flock __user *user)
1266 #if BITS_PER_LONG == 32
1267 static inline int fcntl_getlk64(struct file *file, struct flock64 __user *user)
1272 static inline int fcntl_setlk64(unsigned int fd, struct file *file,
1273 unsigned int cmd, struct flock64 __user *user)
1278 static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
1283 static inline int fcntl_getlease(struct file *filp)
1288 static inline void locks_init_lock(struct file_lock *fl)
1293 static inline void __locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1298 static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1303 static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
1308 static inline void locks_remove_flock(struct file *filp)
1313 static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
1318 static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
1319 struct file_lock *conflock)
1324 static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
1329 static inline int posix_unblock_lock(struct file *filp,
1330 struct file_lock *waiter)
1335 static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
1340 static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
1341 struct file_lock *fl, struct file_lock *conf)
1346 static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
1351 static inline int flock_lock_file_wait(struct file *filp,
1352 struct file_lock *request)
1357 static inline int __break_lease(struct inode *inode, unsigned int mode)
1362 static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
1367 static inline int generic_setlease(struct file *filp, long arg,
1368 struct file_lock **flp)
1373 static inline int vfs_setlease(struct file *filp, long arg,
1374 struct file_lock **lease)
1379 static inline int lease_modify(struct file_lock **before, int arg)
1384 static inline int lock_may_read(struct inode *inode, loff_t start,
1390 static inline int lock_may_write(struct inode *inode, loff_t start,
1396 static inline void locks_delete_block(struct file_lock *waiter)
1400 static inline void lock_flocks(void)
1404 static inline void unlock_flocks(void)
1408 #endif /* !CONFIG_FILE_LOCKING */
1411 struct fasync_struct {
1415 struct fasync_struct *fa_next; /* singly linked list */
1416 struct file *fa_file;
1417 struct rcu_head fa_rcu;
1420 #define FASYNC_MAGIC 0x4601
1422 /* SMP safe fasync helpers: */
1423 extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
1424 extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
1425 extern int fasync_remove_entry(struct file *, struct fasync_struct **);
1426 extern struct fasync_struct *fasync_alloc(void);
1427 extern void fasync_free(struct fasync_struct *);
1429 /* can be called from interrupts */
1430 extern void kill_fasync(struct fasync_struct **, int, int);
1432 extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
1433 extern int f_setown(struct file *filp, unsigned long arg, int force);
1434 extern void f_delown(struct file *filp);
1435 extern pid_t f_getown(struct file *filp);
1436 extern int send_sigurg(struct fown_struct *fown);
1442 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1443 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1444 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1445 #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1446 #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
1449 #define SB_I_MULTIROOT 0x00000008 /* Multiple roots to the dentry tree */
1451 extern struct list_head super_blocks;
1452 extern spinlock_t sb_lock;
1454 struct super_block {
1455 struct list_head s_list; /* Keep this first */
1456 dev_t s_dev; /* search index; _not_ kdev_t */
1457 unsigned char s_dirt;
1458 unsigned char s_blocksize_bits;
1459 unsigned long s_blocksize;
1460 loff_t s_maxbytes; /* Max file size */
1461 struct file_system_type *s_type;
1462 const struct super_operations *s_op;
1463 const struct dquot_operations *dq_op;
1464 const struct quotactl_ops *s_qcop;
1465 const struct export_operations *s_export_op;
1466 unsigned long s_flags;
1467 unsigned long s_iflags; /* internal SB_I_* flags */
1468 unsigned long s_magic;
1469 struct dentry *s_root;
1470 struct rw_semaphore s_umount;
1471 struct mutex s_lock;
1474 #ifdef CONFIG_SECURITY
1477 const struct xattr_handler **s_xattr;
1479 struct list_head s_inodes; /* all inodes */
1480 struct hlist_bl_head s_anon; /* anonymous dentries for (nfs) exporting */
1482 struct list_head __percpu *s_files;
1484 struct list_head s_files;
1486 /* s_dentry_lru, s_nr_dentry_unused protected by dcache.c lru locks */
1487 struct list_head s_dentry_lru; /* unused dentry lru */
1488 int s_nr_dentry_unused; /* # of dentry on lru */
1490 /* s_inode_lru_lock protects s_inode_lru and s_nr_inodes_unused */
1491 spinlock_t s_inode_lru_lock ____cacheline_aligned_in_smp;
1492 struct list_head s_inode_lru; /* unused inode lru */
1493 int s_nr_inodes_unused; /* # of inodes on lru */
1495 struct block_device *s_bdev;
1496 struct backing_dev_info *s_bdi;
1497 struct mtd_info *s_mtd;
1498 struct list_head s_instances;
1499 struct quota_info s_dquot; /* Diskquota specific options */
1502 wait_queue_head_t s_wait_unfrozen;
1504 char s_id[32]; /* Informational name */
1505 u8 s_uuid[16]; /* UUID */
1507 void *s_fs_info; /* Filesystem private info */
1510 /* Granularity of c/m/atime in ns.
1511 Cannot be worse than a second */
1515 * The next field is for VFS *only*. No filesystems have any business
1516 * even looking at it. You had been warned.
1518 struct mutex s_vfs_rename_mutex; /* Kludge */
1521 * Filesystem subtype. If non-empty the filesystem type field
1522 * in /proc/mounts will be "type.subtype"
1527 * Saved mount options for lazy filesystems using
1528 * generic_show_options()
1530 char __rcu *s_options;
1531 const struct dentry_operations *s_d_op; /* default d_op for dentries */
1534 * Saved pool identifier for cleancache (-1 means none)
1536 int cleancache_poolid;
1538 struct shrinker s_shrink; /* per-sb shrinker handle */
1541 * Indicates how deep in a filesystem stack this SB is
1546 /* superblock cache pruning functions */
1547 extern void prune_icache_sb(struct super_block *sb, int nr_to_scan);
1548 extern void prune_dcache_sb(struct super_block *sb, int nr_to_scan);
1550 extern struct timespec current_fs_time(struct super_block *sb);
1553 * Snapshotting support.
1557 SB_FREEZE_WRITE = 1,
1558 SB_FREEZE_TRANS = 2,
1561 #define vfs_check_frozen(sb, level) \
1562 wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
1565 * until VFS tracks user namespaces for inodes, just make all files
1566 * belong to init_user_ns
1568 extern struct user_namespace init_user_ns;
1569 #define inode_userns(inode) (&init_user_ns)
1570 extern bool inode_owner_or_capable(const struct inode *inode);
1572 /* not quite ready to be deprecated, but... */
1573 extern void lock_super(struct super_block *);
1574 extern void unlock_super(struct super_block *);
1577 * VFS helper functions..
1579 extern int vfs_create(struct inode *, struct dentry *, int, struct nameidata *);
1580 extern int vfs_mkdir(struct inode *, struct dentry *, int);
1581 extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t);
1582 extern int vfs_symlink(struct inode *, struct dentry *, const char *);
1583 extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
1584 extern int vfs_rmdir(struct inode *, struct dentry *);
1585 extern int vfs_unlink(struct inode *, struct dentry *);
1586 extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
1589 * VFS dentry helper functions.
1591 extern void dentry_unhash(struct dentry *dentry);
1594 * VFS file helper functions.
1596 extern void inode_init_owner(struct inode *inode, const struct inode *dir,
1599 * VFS FS_IOC_FIEMAP helper definitions.
1601 struct fiemap_extent_info {
1602 unsigned int fi_flags; /* Flags as passed from user */
1603 unsigned int fi_extents_mapped; /* Number of mapped extents */
1604 unsigned int fi_extents_max; /* Size of fiemap_extent array */
1605 struct fiemap_extent __user *fi_extents_start; /* Start of
1606 fiemap_extent array */
1608 int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
1609 u64 phys, u64 len, u32 flags);
1610 int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
1615 * NOTE! These match bits 12..15 of stat.st_mode
1616 * (ie "(i_mode >> 12) & 15").
1618 #define DT_UNKNOWN 0
1629 * This is the "filldir" function type, used by readdir() to let
1630 * the kernel specify what kind of dirent layout it wants to have.
1631 * This allows the kernel to read directories into kernel space or
1632 * to have different dirent layouts depending on the binary type.
1634 typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned);
1635 struct block_device_operations;
1637 /* These macros are for out of kernel modules to test that
1638 * the kernel supports the unlocked_ioctl and compat_ioctl
1639 * fields in struct file_operations. */
1640 #define HAVE_COMPAT_IOCTL 1
1641 #define HAVE_UNLOCKED_IOCTL 1
1643 struct file_operations {
1644 struct module *owner;
1645 loff_t (*llseek) (struct file *, loff_t, int);
1646 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1647 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
1648 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1649 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1650 int (*readdir) (struct file *, void *, filldir_t);
1651 unsigned int (*poll) (struct file *, struct poll_table_struct *);
1652 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1653 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1654 int (*mmap) (struct file *, struct vm_area_struct *);
1655 int (*open) (struct inode *, struct file *);
1656 int (*flush) (struct file *, fl_owner_t id);
1657 int (*release) (struct inode *, struct file *);
1658 int (*fsync) (struct file *, loff_t, loff_t, int datasync);
1659 int (*aio_fsync) (struct kiocb *, int datasync);
1660 int (*fasync) (int, struct file *, int);
1661 int (*lock) (struct file *, int, struct file_lock *);
1662 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1663 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1664 int (*check_flags)(int);
1665 int (*flock) (struct file *, int, struct file_lock *);
1666 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1667 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
1668 int (*setlease)(struct file *, long, struct file_lock **);
1669 long (*fallocate)(struct file *file, int mode, loff_t offset,
1673 struct inode_operations {
1674 struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
1675 void * (*follow_link) (struct dentry *, struct nameidata *);
1676 int (*permission) (struct inode *, int);
1677 struct posix_acl * (*get_acl)(struct inode *, int);
1679 int (*readlink) (struct dentry *, char __user *,int);
1680 void (*put_link) (struct dentry *, struct nameidata *, void *);
1682 int (*create) (struct inode *,struct dentry *,int, struct nameidata *);
1683 int (*link) (struct dentry *,struct inode *,struct dentry *);
1684 int (*unlink) (struct inode *,struct dentry *);
1685 int (*symlink) (struct inode *,struct dentry *,const char *);
1686 int (*mkdir) (struct inode *,struct dentry *,int);
1687 int (*rmdir) (struct inode *,struct dentry *);
1688 int (*mknod) (struct inode *,struct dentry *,int,dev_t);
1689 int (*rename) (struct inode *, struct dentry *,
1690 struct inode *, struct dentry *);
1691 void (*truncate) (struct inode *);
1692 int (*setattr) (struct dentry *, struct iattr *);
1693 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1694 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1695 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1696 ssize_t (*listxattr) (struct dentry *, char *, size_t);
1697 int (*removexattr) (struct dentry *, const char *);
1698 void (*truncate_range)(struct inode *, loff_t, loff_t);
1699 int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
1701 } ____cacheline_aligned;
1705 ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
1706 unsigned long nr_segs, unsigned long fast_segs,
1707 struct iovec *fast_pointer,
1708 struct iovec **ret_pointer,
1711 extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1712 extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1713 extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1714 unsigned long, loff_t *);
1715 extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1716 unsigned long, loff_t *);
1718 struct super_operations {
1719 struct inode *(*alloc_inode)(struct super_block *sb);
1720 void (*destroy_inode)(struct inode *);
1722 void (*dirty_inode) (struct inode *, int flags);
1723 int (*write_inode) (struct inode *, struct writeback_control *wbc);
1724 int (*drop_inode) (struct inode *);
1725 void (*evict_inode) (struct inode *);
1726 void (*put_super) (struct super_block *);
1727 void (*write_super) (struct super_block *);
1728 int (*sync_fs)(struct super_block *sb, int wait);
1729 int (*freeze_fs) (struct super_block *);
1730 int (*unfreeze_fs) (struct super_block *);
1731 int (*statfs) (struct dentry *, struct kstatfs *);
1732 int (*remount_fs) (struct super_block *, int *, char *);
1733 void (*umount_begin) (struct super_block *);
1735 int (*show_options)(struct seq_file *, struct vfsmount *);
1736 int (*show_devname)(struct seq_file *, struct vfsmount *);
1737 int (*show_path)(struct seq_file *, struct vfsmount *);
1738 int (*show_stats)(struct seq_file *, struct vfsmount *);
1740 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1741 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
1743 int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
1744 int (*nr_cached_objects)(struct super_block *);
1745 void (*free_cached_objects)(struct super_block *, int);
1749 * Inode state bits. Protected by inode->i_lock
1751 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1752 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1754 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1755 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1756 * various stages of removing an inode.
1758 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1760 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1761 * fdatasync(). i_atime is the usual cause.
1762 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1763 * these changes separately from I_DIRTY_SYNC so that we
1764 * don't have to write inode on fdatasync() when only
1765 * mtime has changed in it.
1766 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1767 * I_NEW Serves as both a mutex and completion notification.
1768 * New inodes set I_NEW. If two processes both create
1769 * the same inode, one of them will release its inode and
1770 * wait for I_NEW to be released before returning.
1771 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1772 * also cause waiting on I_NEW, without I_NEW actually
1773 * being set. find_inode() uses this to prevent returning
1774 * nearly-dead inodes.
1775 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1776 * is zero. I_FREEING must be set when I_WILL_FREE is
1778 * I_FREEING Set when inode is about to be freed but still has dirty
1779 * pages or buffers attached or the inode itself is still
1781 * I_CLEAR Added by end_writeback(). In this state the inode is clean
1782 * and can be destroyed. Inode keeps I_FREEING.
1784 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1785 * prohibited for many purposes. iget() must wait for
1786 * the inode to be completely released, then create it
1787 * anew. Other functions will just ignore such inodes,
1788 * if appropriate. I_NEW is used for waiting.
1790 * I_SYNC Synchonized write of dirty inode data. The bits is
1791 * set during data writeback, and cleared with a wakeup
1792 * on the bit address once it is done.
1794 * I_REFERENCED Marks the inode as recently references on the LRU list.
1796 * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit().
1798 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1800 #define I_DIRTY_SYNC (1 << 0)
1801 #define I_DIRTY_DATASYNC (1 << 1)
1802 #define I_DIRTY_PAGES (1 << 2)
1804 #define I_NEW (1 << __I_NEW)
1805 #define I_WILL_FREE (1 << 4)
1806 #define I_FREEING (1 << 5)
1807 #define I_CLEAR (1 << 6)
1809 #define I_SYNC (1 << __I_SYNC)
1810 #define I_REFERENCED (1 << 8)
1811 #define __I_DIO_WAKEUP 9
1812 #define I_DIO_WAKEUP (1 << I_DIO_WAKEUP)
1814 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1816 extern void __mark_inode_dirty(struct inode *, int);
1817 static inline void mark_inode_dirty(struct inode *inode)
1819 __mark_inode_dirty(inode, I_DIRTY);
1822 static inline void mark_inode_dirty_sync(struct inode *inode)
1824 __mark_inode_dirty(inode, I_DIRTY_SYNC);
1828 * set_nlink - directly set an inode's link count
1830 * @nlink: new nlink (should be non-zero)
1832 * This is a low-level filesystem helper to replace any
1833 * direct filesystem manipulation of i_nlink.
1835 static inline void set_nlink(struct inode *inode, unsigned int nlink)
1837 inode->__i_nlink = nlink;
1841 * inc_nlink - directly increment an inode's link count
1844 * This is a low-level filesystem helper to replace any
1845 * direct filesystem manipulation of i_nlink. Currently,
1846 * it is only here for parity with dec_nlink().
1848 static inline void inc_nlink(struct inode *inode)
1853 static inline void inode_inc_link_count(struct inode *inode)
1856 mark_inode_dirty(inode);
1860 * drop_nlink - directly drop an inode's link count
1863 * This is a low-level filesystem helper to replace any
1864 * direct filesystem manipulation of i_nlink. In cases
1865 * where we are attempting to track writes to the
1866 * filesystem, a decrement to zero means an imminent
1867 * write when the file is truncated and actually unlinked
1868 * on the filesystem.
1870 static inline void drop_nlink(struct inode *inode)
1876 * clear_nlink - directly zero an inode's link count
1879 * This is a low-level filesystem helper to replace any
1880 * direct filesystem manipulation of i_nlink. See
1881 * drop_nlink() for why we care about i_nlink hitting zero.
1883 static inline void clear_nlink(struct inode *inode)
1885 inode->__i_nlink = 0;
1888 static inline void inode_dec_link_count(struct inode *inode)
1891 mark_inode_dirty(inode);
1895 * inode_inc_iversion - increments i_version
1896 * @inode: inode that need to be updated
1898 * Every time the inode is modified, the i_version field will be incremented.
1899 * The filesystem has to be mounted with i_version flag
1902 static inline void inode_inc_iversion(struct inode *inode)
1904 spin_lock(&inode->i_lock);
1906 spin_unlock(&inode->i_lock);
1909 extern void touch_atime(struct vfsmount *mnt, struct dentry *dentry);
1910 static inline void file_accessed(struct file *file)
1912 if (!(file->f_flags & O_NOATIME))
1913 touch_atime(file->f_path.mnt, file->f_path.dentry);
1916 int sync_inode(struct inode *inode, struct writeback_control *wbc);
1917 int sync_inode_metadata(struct inode *inode, int wait);
1919 struct file_system_type {
1922 struct dentry *(*mount) (struct file_system_type *, int,
1923 const char *, void *);
1924 void (*kill_sb) (struct super_block *);
1925 struct module *owner;
1926 struct file_system_type * next;
1927 struct list_head fs_supers;
1929 struct lock_class_key s_lock_key;
1930 struct lock_class_key s_umount_key;
1931 struct lock_class_key s_vfs_rename_key;
1933 struct lock_class_key i_lock_key;
1934 struct lock_class_key i_mutex_key;
1935 struct lock_class_key i_mutex_dir_key;
1938 extern struct dentry *mount_ns(struct file_system_type *fs_type, int flags,
1939 void *data, int (*fill_super)(struct super_block *, void *, int));
1940 extern struct dentry *mount_bdev(struct file_system_type *fs_type,
1941 int flags, const char *dev_name, void *data,
1942 int (*fill_super)(struct super_block *, void *, int));
1943 extern struct dentry *mount_single(struct file_system_type *fs_type,
1944 int flags, void *data,
1945 int (*fill_super)(struct super_block *, void *, int));
1946 extern struct dentry *mount_nodev(struct file_system_type *fs_type,
1947 int flags, void *data,
1948 int (*fill_super)(struct super_block *, void *, int));
1949 extern struct dentry *mount_subtree(struct vfsmount *mnt, const char *path);
1950 void generic_shutdown_super(struct super_block *sb);
1951 void kill_block_super(struct super_block *sb);
1952 void kill_anon_super(struct super_block *sb);
1953 void kill_litter_super(struct super_block *sb);
1954 void deactivate_super(struct super_block *sb);
1955 void deactivate_locked_super(struct super_block *sb);
1956 int set_anon_super(struct super_block *s, void *data);
1957 int get_anon_bdev(dev_t *);
1958 void free_anon_bdev(dev_t);
1959 struct super_block *sget(struct file_system_type *type,
1960 int (*test)(struct super_block *,void *),
1961 int (*set)(struct super_block *,void *),
1963 extern struct dentry *mount_pseudo(struct file_system_type *, char *,
1964 const struct super_operations *ops,
1965 const struct dentry_operations *dops,
1968 static inline void sb_mark_dirty(struct super_block *sb)
1972 static inline void sb_mark_clean(struct super_block *sb)
1976 static inline int sb_is_dirty(struct super_block *sb)
1981 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1982 #define fops_get(fops) \
1983 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1984 #define fops_put(fops) \
1985 do { if (fops) module_put((fops)->owner); } while(0)
1987 extern int register_filesystem(struct file_system_type *);
1988 extern int unregister_filesystem(struct file_system_type *);
1989 extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
1990 #define kern_mount(type) kern_mount_data(type, NULL)
1991 extern void kern_unmount(struct vfsmount *mnt);
1992 extern int may_umount_tree(struct vfsmount *);
1993 extern int may_umount(struct vfsmount *);
1994 extern long do_mount(char *, char *, char *, unsigned long, void *);
1995 extern struct vfsmount *collect_mounts(struct path *);
1996 extern void drop_collected_mounts(struct vfsmount *);
1997 extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
1999 extern int vfs_statfs(struct path *, struct kstatfs *);
2000 extern int user_statfs(const char __user *, struct kstatfs *);
2001 extern int fd_statfs(int, struct kstatfs *);
2002 extern int statfs_by_dentry(struct dentry *, struct kstatfs *);
2003 extern int freeze_super(struct super_block *super);
2004 extern int thaw_super(struct super_block *super);
2005 extern bool our_mnt(struct vfsmount *mnt);
2007 extern int current_umask(void);
2009 static inline struct inode *file_inode(struct file *f)
2011 return f->f_dentry->d_inode;
2015 extern struct kobject *fs_kobj;
2017 #define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
2018 extern int rw_verify_area(int, struct file *, loff_t *, size_t);
2020 #define FLOCK_VERIFY_READ 1
2021 #define FLOCK_VERIFY_WRITE 2
2023 #ifdef CONFIG_FILE_LOCKING
2024 extern int locks_mandatory_locked(struct inode *);
2025 extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
2028 * Candidates for mandatory locking have the setgid bit set
2029 * but no group execute bit - an otherwise meaningless combination.
2032 static inline int __mandatory_lock(struct inode *ino)
2034 return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
2038 * ... and these candidates should be on MS_MANDLOCK mounted fs,
2039 * otherwise these will be advisory locks
2042 static inline int mandatory_lock(struct inode *ino)
2044 return IS_MANDLOCK(ino) && __mandatory_lock(ino);
2047 static inline int locks_verify_locked(struct inode *inode)
2049 if (mandatory_lock(inode))
2050 return locks_mandatory_locked(inode);
2054 static inline int locks_verify_truncate(struct inode *inode,
2058 if (inode->i_flock && mandatory_lock(inode))
2059 return locks_mandatory_area(
2060 FLOCK_VERIFY_WRITE, inode, filp,
2061 size < inode->i_size ? size : inode->i_size,
2062 (size < inode->i_size ? inode->i_size - size
2063 : size - inode->i_size)
2068 static inline int break_lease(struct inode *inode, unsigned int mode)
2071 return __break_lease(inode, mode);
2074 #else /* !CONFIG_FILE_LOCKING */
2075 static inline int locks_mandatory_locked(struct inode *inode)
2080 static inline int locks_mandatory_area(int rw, struct inode *inode,
2081 struct file *filp, loff_t offset,
2087 static inline int __mandatory_lock(struct inode *inode)
2092 static inline int mandatory_lock(struct inode *inode)
2097 static inline int locks_verify_locked(struct inode *inode)
2102 static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
2108 static inline int break_lease(struct inode *inode, unsigned int mode)
2113 #endif /* CONFIG_FILE_LOCKING */
2117 extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
2119 extern int do_fallocate(struct file *file, int mode, loff_t offset,
2121 extern long do_sys_open(int dfd, const char __user *filename, int flags,
2123 extern struct file *filp_open(const char *, int, int);
2124 extern struct file *file_open_root(struct dentry *, struct vfsmount *,
2125 const char *, int, umode_t);
2126 extern struct file * dentry_open(struct dentry *, struct vfsmount *, int,
2127 const struct cred *);
2128 extern int filp_close(struct file *, fl_owner_t id);
2129 extern char * getname(const char __user *);
2133 extern int ioctl_preallocate(struct file *filp, void __user *argp);
2136 extern void __init vfs_caches_init_early(void);
2137 extern void __init vfs_caches_init(unsigned long);
2139 extern struct kmem_cache *names_cachep;
2141 #define __getname_gfp(gfp) kmem_cache_alloc(names_cachep, (gfp))
2142 #define __getname() __getname_gfp(GFP_KERNEL)
2143 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
2144 #ifndef CONFIG_AUDITSYSCALL
2145 #define putname(name) __putname(name)
2147 extern void putname(const char *name);
2151 extern int register_blkdev(unsigned int, const char *);
2152 extern void unregister_blkdev(unsigned int, const char *);
2153 extern struct block_device *bdget(dev_t);
2154 extern struct block_device *bdgrab(struct block_device *bdev);
2155 extern void bd_set_size(struct block_device *, loff_t size);
2156 extern sector_t blkdev_max_block(struct block_device *bdev);
2157 extern void bd_forget(struct inode *inode);
2158 extern void bdput(struct block_device *);
2159 extern void invalidate_bdev(struct block_device *);
2160 extern int sync_blockdev(struct block_device *bdev);
2161 extern void kill_bdev(struct block_device *);
2162 extern struct super_block *freeze_bdev(struct block_device *);
2163 extern void emergency_thaw_all(void);
2164 extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
2165 extern int fsync_bdev(struct block_device *);
2167 static inline void bd_forget(struct inode *inode) {}
2168 static inline int sync_blockdev(struct block_device *bdev) { return 0; }
2169 static inline void kill_bdev(struct block_device *bdev) {}
2170 static inline void invalidate_bdev(struct block_device *bdev) {}
2172 static inline struct super_block *freeze_bdev(struct block_device *sb)
2177 static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
2182 extern int sync_filesystem(struct super_block *);
2183 extern const struct file_operations def_blk_fops;
2184 extern const struct file_operations def_chr_fops;
2185 extern const struct file_operations bad_sock_fops;
2186 extern const struct file_operations def_fifo_fops;
2188 extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
2189 extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
2190 extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
2191 extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
2192 extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
2194 extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
2196 extern int blkdev_put(struct block_device *bdev, fmode_t mode);
2198 extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
2199 extern void bd_unlink_disk_holder(struct block_device *bdev,
2200 struct gendisk *disk);
2202 static inline int bd_link_disk_holder(struct block_device *bdev,
2203 struct gendisk *disk)
2207 static inline void bd_unlink_disk_holder(struct block_device *bdev,
2208 struct gendisk *disk)
2215 #define CHRDEV_MAJOR_HASH_SIZE 255
2216 extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
2217 extern int register_chrdev_region(dev_t, unsigned, const char *);
2218 extern int __register_chrdev(unsigned int major, unsigned int baseminor,
2219 unsigned int count, const char *name,
2220 const struct file_operations *fops);
2221 extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
2222 unsigned int count, const char *name);
2223 extern void unregister_chrdev_region(dev_t, unsigned);
2224 extern void chrdev_show(struct seq_file *,off_t);
2226 static inline int register_chrdev(unsigned int major, const char *name,
2227 const struct file_operations *fops)
2229 return __register_chrdev(major, 0, 256, name, fops);
2232 static inline void unregister_chrdev(unsigned int major, const char *name)
2234 __unregister_chrdev(major, 0, 256, name);
2237 /* fs/block_dev.c */
2238 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
2239 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
2242 #define BLKDEV_MAJOR_HASH_SIZE 255
2243 extern const char *__bdevname(dev_t, char *buffer);
2244 extern const char *bdevname(struct block_device *bdev, char *buffer);
2245 extern struct block_device *lookup_bdev(const char *);
2246 extern void blkdev_show(struct seq_file *,off_t);
2249 #define BLKDEV_MAJOR_HASH_SIZE 0
2252 extern void init_special_inode(struct inode *, umode_t, dev_t);
2254 /* Invalid inode operations -- fs/bad_inode.c */
2255 extern void make_bad_inode(struct inode *);
2256 extern int is_bad_inode(struct inode *);
2258 extern const struct file_operations read_pipefifo_fops;
2259 extern const struct file_operations write_pipefifo_fops;
2260 extern const struct file_operations rdwr_pipefifo_fops;
2262 extern int fs_may_remount_ro(struct super_block *);
2266 * return READ, READA, or WRITE
2268 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2271 * return data direction, READ or WRITE
2273 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
2275 extern void check_disk_size_change(struct gendisk *disk,
2276 struct block_device *bdev);
2277 extern int revalidate_disk(struct gendisk *);
2278 extern int check_disk_change(struct block_device *);
2279 extern int __invalidate_device(struct block_device *, bool);
2280 extern int invalidate_partition(struct gendisk *, int);
2282 unsigned long invalidate_mapping_pages(struct address_space *mapping,
2283 pgoff_t start, pgoff_t end);
2285 static inline void invalidate_remote_inode(struct inode *inode)
2287 if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
2288 S_ISLNK(inode->i_mode))
2289 invalidate_mapping_pages(inode->i_mapping, 0, -1);
2291 extern int invalidate_inode_pages2(struct address_space *mapping);
2292 extern int invalidate_inode_pages2_range(struct address_space *mapping,
2293 pgoff_t start, pgoff_t end);
2294 extern int write_inode_now(struct inode *, int);
2295 extern int filemap_fdatawrite(struct address_space *);
2296 extern int filemap_flush(struct address_space *);
2297 extern int filemap_fdatawait(struct address_space *);
2298 extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
2300 extern int filemap_write_and_wait(struct address_space *mapping);
2301 extern int filemap_write_and_wait_range(struct address_space *mapping,
2302 loff_t lstart, loff_t lend);
2303 extern int __filemap_fdatawrite_range(struct address_space *mapping,
2304 loff_t start, loff_t end, int sync_mode);
2305 extern int filemap_fdatawrite_range(struct address_space *mapping,
2306 loff_t start, loff_t end);
2308 extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
2310 extern int vfs_fsync(struct file *file, int datasync);
2311 extern int generic_write_sync(struct file *file, loff_t pos, loff_t count);
2312 extern void sync_supers(void);
2313 extern void emergency_sync(void);
2314 extern void emergency_remount(void);
2316 extern sector_t bmap(struct inode *, sector_t);
2318 extern int notify_change(struct dentry *, struct iattr *);
2319 extern int inode_permission(struct inode *, int);
2320 extern int generic_permission(struct inode *, int);
2322 static inline bool execute_ok(struct inode *inode)
2324 return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
2328 * get_write_access() gets write permission for a file.
2329 * put_write_access() releases this write permission.
2330 * This is used for regular files.
2331 * We cannot support write (and maybe mmap read-write shared) accesses and
2332 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2333 * can have the following values:
2334 * 0: no writers, no VM_DENYWRITE mappings
2335 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2336 * > 0: (i_writecount) users are writing to the file.
2338 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2339 * except for the cases where we don't hold i_writecount yet. Then we need to
2340 * use {get,deny}_write_access() - these functions check the sign and refuse
2341 * to do the change if sign is wrong.
2343 static inline int get_write_access(struct inode *inode)
2345 return atomic_inc_unless_negative(&inode->i_writecount) ? 0 : -ETXTBSY;
2347 static inline int deny_write_access(struct file *file)
2349 struct inode *inode = file->f_path.dentry->d_inode;
2350 return atomic_dec_unless_positive(&inode->i_writecount) ? 0 : -ETXTBSY;
2352 static inline void put_write_access(struct inode * inode)
2354 atomic_dec(&inode->i_writecount);
2356 static inline void allow_write_access(struct file *file)
2359 atomic_inc(&file->f_path.dentry->d_inode->i_writecount);
2362 static inline void i_readcount_dec(struct inode *inode)
2364 BUG_ON(!atomic_read(&inode->i_readcount));
2365 atomic_dec(&inode->i_readcount);
2367 static inline void i_readcount_inc(struct inode *inode)
2369 atomic_inc(&inode->i_readcount);
2372 static inline void i_readcount_dec(struct inode *inode)
2376 static inline void i_readcount_inc(struct inode *inode)
2381 extern int do_pipe_flags(int *, int);
2382 extern struct file *create_read_pipe(struct file *f, int flags);
2383 extern struct file *create_write_pipe(int flags);
2384 extern void free_write_pipe(struct file *);
2386 extern int kernel_read(struct file *, loff_t, char *, unsigned long);
2387 extern struct file * open_exec(const char *);
2389 /* fs/dcache.c -- generic fs support functions */
2390 extern int is_subdir(struct dentry *, struct dentry *);
2391 extern int path_is_under(struct path *, struct path *);
2392 extern ino_t find_inode_number(struct dentry *, struct qstr *);
2394 #include <linux/err.h>
2396 /* needed for stackable file system support */
2397 extern loff_t default_llseek(struct file *file, loff_t offset, int origin);
2399 extern loff_t vfs_llseek(struct file *file, loff_t offset, int origin);
2401 extern int inode_init_always(struct super_block *, struct inode *);
2402 extern void inode_init_once(struct inode *);
2403 extern void address_space_init_once(struct address_space *mapping);
2404 extern void ihold(struct inode * inode);
2405 extern void iput(struct inode *);
2406 extern struct inode * igrab(struct inode *);
2407 extern ino_t iunique(struct super_block *, ino_t);
2408 extern int inode_needs_sync(struct inode *inode);
2409 extern int generic_delete_inode(struct inode *inode);
2410 extern int generic_drop_inode(struct inode *inode);
2412 extern struct inode *ilookup5_nowait(struct super_block *sb,
2413 unsigned long hashval, int (*test)(struct inode *, void *),
2415 extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
2416 int (*test)(struct inode *, void *), void *data);
2417 extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
2419 extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
2420 extern struct inode * iget_locked(struct super_block *, unsigned long);
2421 extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
2422 extern int insert_inode_locked(struct inode *);
2423 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2424 extern void lockdep_annotate_inode_mutex_key(struct inode *inode);
2426 static inline void lockdep_annotate_inode_mutex_key(struct inode *inode) { };
2428 extern void unlock_new_inode(struct inode *);
2429 extern unsigned int get_next_ino(void);
2431 extern void __iget(struct inode * inode);
2432 extern void iget_failed(struct inode *);
2433 extern void end_writeback(struct inode *);
2434 extern void __destroy_inode(struct inode *);
2435 extern struct inode *new_inode_pseudo(struct super_block *sb);
2436 extern struct inode *new_inode(struct super_block *sb);
2437 extern void free_inode_nonrcu(struct inode *inode);
2438 extern int should_remove_suid(struct dentry *);
2439 extern int file_remove_suid(struct file *);
2441 extern void __insert_inode_hash(struct inode *, unsigned long hashval);
2442 static inline void insert_inode_hash(struct inode *inode)
2444 __insert_inode_hash(inode, inode->i_ino);
2447 extern void __remove_inode_hash(struct inode *);
2448 static inline void remove_inode_hash(struct inode *inode)
2450 if (!inode_unhashed(inode))
2451 __remove_inode_hash(inode);
2454 extern void inode_sb_list_add(struct inode *inode);
2457 extern void submit_bio(int, struct bio *);
2458 extern int bdev_read_only(struct block_device *);
2460 extern int set_blocksize(struct block_device *, int);
2461 extern int sb_set_blocksize(struct super_block *, int);
2462 extern int sb_min_blocksize(struct super_block *, int);
2464 extern int generic_file_mmap(struct file *, struct vm_area_struct *);
2465 extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
2466 extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
2467 int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
2468 extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2469 extern ssize_t __generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long,
2471 extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2472 extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
2473 unsigned long *, loff_t, loff_t *, size_t, size_t);
2474 extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
2475 unsigned long, loff_t, loff_t *, size_t, ssize_t);
2476 extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
2477 extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
2478 extern int generic_segment_checks(const struct iovec *iov,
2479 unsigned long *nr_segs, size_t *count, int access_flags);
2481 /* fs/block_dev.c */
2482 extern ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
2483 unsigned long nr_segs, loff_t pos);
2484 extern int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
2488 extern ssize_t generic_file_splice_read(struct file *, loff_t *,
2489 struct pipe_inode_info *, size_t, unsigned int);
2490 extern ssize_t default_file_splice_read(struct file *, loff_t *,
2491 struct pipe_inode_info *, size_t, unsigned int);
2492 extern ssize_t generic_file_splice_write(struct pipe_inode_info *,
2493 struct file *, loff_t *, size_t, unsigned int);
2494 extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
2495 struct file *out, loff_t *, size_t len, unsigned int flags);
2496 extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
2497 size_t len, unsigned int flags);
2500 file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
2501 extern loff_t noop_llseek(struct file *file, loff_t offset, int origin);
2502 extern loff_t no_llseek(struct file *file, loff_t offset, int origin);
2503 extern loff_t generic_file_llseek(struct file *file, loff_t offset, int origin);
2504 extern loff_t generic_file_llseek_size(struct file *file, loff_t offset,
2505 int origin, loff_t maxsize);
2506 extern int generic_file_open(struct inode * inode, struct file * filp);
2507 extern int nonseekable_open(struct inode * inode, struct file * filp);
2509 #ifdef CONFIG_FS_XIP
2510 extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
2512 extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
2513 extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
2514 size_t len, loff_t *ppos);
2515 extern int xip_truncate_page(struct address_space *mapping, loff_t from);
2517 static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
2524 typedef void (dio_submit_t)(int rw, struct bio *bio, struct inode *inode,
2525 loff_t file_offset);
2528 /* need locking between buffered and direct access */
2531 /* filesystem does not support filling holes */
2532 DIO_SKIP_HOLES = 0x02,
2535 void dio_end_io(struct bio *bio, int error);
2536 void inode_dio_wait(struct inode *inode);
2537 void inode_dio_done(struct inode *inode);
2539 ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
2540 struct block_device *bdev, const struct iovec *iov, loff_t offset,
2541 unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
2542 dio_submit_t submit_io, int flags);
2544 static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
2545 struct inode *inode, const struct iovec *iov, loff_t offset,
2546 unsigned long nr_segs, get_block_t get_block)
2548 return __blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
2549 offset, nr_segs, get_block, NULL, NULL,
2550 DIO_LOCKING | DIO_SKIP_HOLES);
2553 static inline void inode_dio_wait(struct inode *inode)
2558 extern const struct file_operations generic_ro_fops;
2560 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2562 extern int vfs_readlink(struct dentry *, char __user *, int, const char *);
2563 extern int vfs_follow_link(struct nameidata *, const char *);
2564 extern int page_readlink(struct dentry *, char __user *, int);
2565 extern void *page_follow_link_light(struct dentry *, struct nameidata *);
2566 extern void page_put_link(struct dentry *, struct nameidata *, void *);
2567 extern int __page_symlink(struct inode *inode, const char *symname, int len,
2569 extern int page_symlink(struct inode *inode, const char *symname, int len);
2570 extern const struct inode_operations page_symlink_inode_operations;
2571 extern int generic_readlink(struct dentry *, char __user *, int);
2572 extern void generic_fillattr(struct inode *, struct kstat *);
2573 extern int vfs_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2574 void __inode_add_bytes(struct inode *inode, loff_t bytes);
2575 void inode_add_bytes(struct inode *inode, loff_t bytes);
2576 void inode_sub_bytes(struct inode *inode, loff_t bytes);
2577 loff_t inode_get_bytes(struct inode *inode);
2578 void inode_set_bytes(struct inode *inode, loff_t bytes);
2580 extern int vfs_readdir(struct file *, filldir_t, void *);
2582 extern int vfs_stat(const char __user *, struct kstat *);
2583 extern int vfs_lstat(const char __user *, struct kstat *);
2584 extern int vfs_fstat(unsigned int, struct kstat *);
2585 extern int vfs_fstatat(int , const char __user *, struct kstat *, int);
2587 extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
2589 extern int __generic_block_fiemap(struct inode *inode,
2590 struct fiemap_extent_info *fieinfo,
2591 loff_t start, loff_t len,
2592 get_block_t *get_block);
2593 extern int generic_block_fiemap(struct inode *inode,
2594 struct fiemap_extent_info *fieinfo, u64 start,
2595 u64 len, get_block_t *get_block);
2597 extern void get_filesystem(struct file_system_type *fs);
2598 extern void put_filesystem(struct file_system_type *fs);
2599 extern struct file_system_type *get_fs_type(const char *name);
2600 extern struct super_block *get_super(struct block_device *);
2601 extern struct super_block *get_active_super(struct block_device *bdev);
2602 extern struct super_block *user_get_super(dev_t);
2603 extern void drop_super(struct super_block *sb);
2604 extern void iterate_supers(void (*)(struct super_block *, void *), void *);
2605 extern void iterate_supers_type(struct file_system_type *,
2606 void (*)(struct super_block *, void *), void *);
2608 extern int dcache_dir_open(struct inode *, struct file *);
2609 extern int dcache_dir_close(struct inode *, struct file *);
2610 extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
2611 extern int dcache_readdir(struct file *, void *, filldir_t);
2612 extern int simple_setattr(struct dentry *, struct iattr *);
2613 extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2614 extern int simple_statfs(struct dentry *, struct kstatfs *);
2615 extern int simple_link(struct dentry *, struct inode *, struct dentry *);
2616 extern int simple_unlink(struct inode *, struct dentry *);
2617 extern int simple_rmdir(struct inode *, struct dentry *);
2618 extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
2619 extern int noop_fsync(struct file *, loff_t, loff_t, int);
2620 extern int simple_empty(struct dentry *);
2621 extern int simple_readpage(struct file *file, struct page *page);
2622 extern int simple_write_begin(struct file *file, struct address_space *mapping,
2623 loff_t pos, unsigned len, unsigned flags,
2624 struct page **pagep, void **fsdata);
2625 extern int simple_write_end(struct file *file, struct address_space *mapping,
2626 loff_t pos, unsigned len, unsigned copied,
2627 struct page *page, void *fsdata);
2629 extern struct dentry *simple_lookup(struct inode *, struct dentry *, struct nameidata *);
2630 extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
2631 extern const struct file_operations simple_dir_operations;
2632 extern const struct inode_operations simple_dir_inode_operations;
2633 struct tree_descr { char *name; const struct file_operations *ops; int mode; };
2634 struct dentry *d_alloc_name(struct dentry *, const char *);
2635 extern int simple_fill_super(struct super_block *, unsigned long, struct tree_descr *);
2636 extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
2637 extern void simple_release_fs(struct vfsmount **mount, int *count);
2639 extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
2640 loff_t *ppos, const void *from, size_t available);
2641 extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
2642 const void __user *from, size_t count);
2644 extern int generic_file_fsync(struct file *, loff_t, loff_t, int);
2646 extern int generic_check_addressable(unsigned, u64);
2648 #ifdef CONFIG_MIGRATION
2649 extern int buffer_migrate_page(struct address_space *,
2650 struct page *, struct page *,
2653 #define buffer_migrate_page NULL
2656 extern int setattr_prepare(struct dentry *, struct iattr *);
2657 extern int inode_newsize_ok(const struct inode *, loff_t offset);
2658 extern void setattr_copy(struct inode *inode, const struct iattr *attr);
2660 extern void file_update_time(struct file *file);
2662 extern int generic_show_options(struct seq_file *m, struct vfsmount *mnt);
2663 extern void save_mount_options(struct super_block *sb, char *options);
2664 extern void replace_mount_options(struct super_block *sb, char *options);
2666 static inline ino_t parent_ino(struct dentry *dentry)
2671 * Don't strictly need d_lock here? If the parent ino could change
2672 * then surely we'd have a deeper race in the caller?
2674 spin_lock(&dentry->d_lock);
2675 res = dentry->d_parent->d_inode->i_ino;
2676 spin_unlock(&dentry->d_lock);
2680 /* Transaction based IO helpers */
2683 * An argresp is stored in an allocated page and holds the
2684 * size of the argument or response, along with its content
2686 struct simple_transaction_argresp {
2691 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2693 char *simple_transaction_get(struct file *file, const char __user *buf,
2695 ssize_t simple_transaction_read(struct file *file, char __user *buf,
2696 size_t size, loff_t *pos);
2697 int simple_transaction_release(struct inode *inode, struct file *file);
2699 void simple_transaction_set(struct file *file, size_t n);
2702 * simple attribute files
2704 * These attributes behave similar to those in sysfs:
2706 * Writing to an attribute immediately sets a value, an open file can be
2707 * written to multiple times.
2709 * Reading from an attribute creates a buffer from the value that might get
2710 * read with multiple read calls. When the attribute has been read
2711 * completely, no further read calls are possible until the file is opened
2714 * All attributes contain a text representation of a numeric value
2715 * that are accessed with the get() and set() functions.
2717 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2718 static int __fops ## _open(struct inode *inode, struct file *file) \
2720 __simple_attr_check_format(__fmt, 0ull); \
2721 return simple_attr_open(inode, file, __get, __set, __fmt); \
2723 static const struct file_operations __fops = { \
2724 .owner = THIS_MODULE, \
2725 .open = __fops ## _open, \
2726 .release = simple_attr_release, \
2727 .read = simple_attr_read, \
2728 .write = simple_attr_write, \
2729 .llseek = generic_file_llseek, \
2732 static inline __printf(1, 2)
2733 void __simple_attr_check_format(const char *fmt, ...)
2735 /* don't do anything, just let the compiler check the arguments; */
2738 int simple_attr_open(struct inode *inode, struct file *file,
2739 int (*get)(void *, u64 *), int (*set)(void *, u64),
2741 int simple_attr_release(struct inode *inode, struct file *file);
2742 ssize_t simple_attr_read(struct file *file, char __user *buf,
2743 size_t len, loff_t *ppos);
2744 ssize_t simple_attr_write(struct file *file, const char __user *buf,
2745 size_t len, loff_t *ppos);
2748 int proc_nr_files(struct ctl_table *table, int write,
2749 void __user *buffer, size_t *lenp, loff_t *ppos);
2750 int proc_nr_dentry(struct ctl_table *table, int write,
2751 void __user *buffer, size_t *lenp, loff_t *ppos);
2752 int proc_nr_inodes(struct ctl_table *table, int write,
2753 void __user *buffer, size_t *lenp, loff_t *ppos);
2754 int __init get_filesystem_list(char *buf);
2756 #define __FMODE_EXEC ((__force int) FMODE_EXEC)
2757 #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
2759 #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
2760 #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
2761 (flag & __FMODE_NONOTIFY)))
2763 static inline int is_sxid(mode_t mode)
2765 return (mode & S_ISUID) || ((mode & S_ISGID) && (mode & S_IXGRP));
2768 static inline void inode_has_no_xattr(struct inode *inode)
2770 if (!is_sxid(inode->i_mode) && (inode->i_sb->s_flags & MS_NOSEC))
2771 inode->i_flags |= S_NOSEC;
2774 #endif /* __KERNEL__ */
2775 #endif /* _LINUX_FS_H */