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
168 #define KERNEL_READ (READ|REQ_KERNEL)
169 #define KERNEL_WRITE (WRITE|REQ_KERNEL)
171 #define READ_SYNC (READ | REQ_SYNC)
172 #define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE)
173 #define WRITE_ODIRECT (WRITE | REQ_SYNC)
174 #define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
175 #define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
176 #define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
180 * Flag for rw_copy_check_uvector and compat_rw_copy_check_uvector
181 * that indicates that they should check the contents of the iovec are
182 * valid, but not check the memory that the iovec elements
185 #define CHECK_IOVEC_ONLY -1
191 /* public flags for file_system_type */
192 #define FS_REQUIRES_DEV 1
193 #define FS_BINARY_MOUNTDATA 2
194 #define FS_HAS_SUBTYPE 4
195 #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
196 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move()
197 * during rename() internally.
201 * These are the fs-independent mount-flags: up to 32 flags are supported
203 #define MS_RDONLY 1 /* Mount read-only */
204 #define MS_NOSUID 2 /* Ignore suid and sgid bits */
205 #define MS_NODEV 4 /* Disallow access to device special files */
206 #define MS_NOEXEC 8 /* Disallow program execution */
207 #define MS_SYNCHRONOUS 16 /* Writes are synced at once */
208 #define MS_REMOUNT 32 /* Alter flags of a mounted FS */
209 #define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
210 #define MS_DIRSYNC 128 /* Directory modifications are synchronous */
211 #define MS_NOATIME 1024 /* Do not update access times. */
212 #define MS_NODIRATIME 2048 /* Do not update directory access times */
216 #define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
217 MS_VERBOSE is deprecated. */
218 #define MS_SILENT 32768
219 #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
220 #define MS_UNBINDABLE (1<<17) /* change to unbindable */
221 #define MS_PRIVATE (1<<18) /* change to private */
222 #define MS_SLAVE (1<<19) /* change to slave */
223 #define MS_SHARED (1<<20) /* change to shared */
224 #define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */
225 #define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
226 #define MS_I_VERSION (1<<23) /* Update inode I_version field */
227 #define MS_STRICTATIME (1<<24) /* Always perform atime updates */
228 #define MS_NOSEC (1<<28)
229 #define MS_BORN (1<<29)
230 #define MS_ACTIVE (1<<30)
231 #define MS_NOUSER (1<<31)
234 * Superblock flags that can be altered by MS_REMOUNT
236 #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK|MS_I_VERSION)
239 * Old magic mount flag and mask
241 #define MS_MGC_VAL 0xC0ED0000
242 #define MS_MGC_MSK 0xffff0000
244 /* Inode flags - they have nothing to superblock flags now */
246 #define S_SYNC 1 /* Writes are synced at once */
247 #define S_NOATIME 2 /* Do not update access times */
248 #define S_APPEND 4 /* Append-only file */
249 #define S_IMMUTABLE 8 /* Immutable file */
250 #define S_DEAD 16 /* removed, but still open directory */
251 #define S_NOQUOTA 32 /* Inode is not counted to quota */
252 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
253 #define S_NOCMTIME 128 /* Do not update file c/mtime */
254 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
255 #define S_PRIVATE 512 /* Inode is fs-internal */
256 #define S_IMA 1024 /* Inode has an associated IMA struct */
257 #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
258 #define S_NOSEC 4096 /* no suid or xattr security attributes */
261 * Note that nosuid etc flags are inode-specific: setting some file-system
262 * flags just means all the inodes inherit those flags by default. It might be
263 * possible to override it selectively if you really wanted to with some
264 * ioctl() that is not currently implemented.
266 * Exception: MS_RDONLY is always applied to the entire file system.
268 * Unfortunately, it is possible to change a filesystems flags with it mounted
269 * with files in use. This means that all of the inodes will not have their
270 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
271 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
273 #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
275 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
276 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
277 ((inode)->i_flags & S_SYNC))
278 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
279 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
280 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
281 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
282 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
284 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
285 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
286 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
287 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
289 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
290 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
291 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
292 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
293 #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
294 #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
295 #define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
297 /* the read-only stuff doesn't really belong here, but any other place is
298 probably as bad and I don't want to create yet another include file. */
300 #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
301 #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
302 #define BLKRRPART _IO(0x12,95) /* re-read partition table */
303 #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
304 #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
305 #define BLKRASET _IO(0x12,98) /* set read ahead for block device */
306 #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
307 #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
308 #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
309 #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
310 #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
311 #define BLKSSZGET _IO(0x12,104)/* get block device sector size */
313 #define BLKPG _IO(0x12,105)/* See blkpg.h */
315 /* Some people are morons. Do not use sizeof! */
317 #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
318 #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
319 /* This was here just to show that the number is taken -
320 probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
322 /* A jump here: 108-111 have been used for various private purposes. */
323 #define BLKBSZGET _IOR(0x12,112,size_t)
324 #define BLKBSZSET _IOW(0x12,113,size_t)
325 #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
326 #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
327 #define BLKTRACESTART _IO(0x12,116)
328 #define BLKTRACESTOP _IO(0x12,117)
329 #define BLKTRACETEARDOWN _IO(0x12,118)
330 #define BLKDISCARD _IO(0x12,119)
331 #define BLKIOMIN _IO(0x12,120)
332 #define BLKIOOPT _IO(0x12,121)
333 #define BLKALIGNOFF _IO(0x12,122)
334 #define BLKPBSZGET _IO(0x12,123)
335 #define BLKDISCARDZEROES _IO(0x12,124)
336 #define BLKSECDISCARD _IO(0x12,125)
337 #define BLKROTATIONAL _IO(0x12,126)
338 #define BLKZEROOUT _IO(0x12,127)
340 #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
341 #define FIBMAP _IO(0x00,1) /* bmap access */
342 #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
343 #define FIFREEZE _IOWR('X', 119, int) /* Freeze */
344 #define FITHAW _IOWR('X', 120, int) /* Thaw */
345 #define FITRIM _IOWR('X', 121, struct fstrim_range) /* Trim */
347 #define FS_IOC_GETFLAGS _IOR('f', 1, long)
348 #define FS_IOC_SETFLAGS _IOW('f', 2, long)
349 #define FS_IOC_GETVERSION _IOR('v', 1, long)
350 #define FS_IOC_SETVERSION _IOW('v', 2, long)
351 #define FS_IOC_FIEMAP _IOWR('f', 11, struct fiemap)
352 #define FS_IOC32_GETFLAGS _IOR('f', 1, int)
353 #define FS_IOC32_SETFLAGS _IOW('f', 2, int)
354 #define FS_IOC32_GETVERSION _IOR('v', 1, int)
355 #define FS_IOC32_SETVERSION _IOW('v', 2, int)
358 * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS)
360 #define FS_SECRM_FL 0x00000001 /* Secure deletion */
361 #define FS_UNRM_FL 0x00000002 /* Undelete */
362 #define FS_COMPR_FL 0x00000004 /* Compress file */
363 #define FS_SYNC_FL 0x00000008 /* Synchronous updates */
364 #define FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
365 #define FS_APPEND_FL 0x00000020 /* writes to file may only append */
366 #define FS_NODUMP_FL 0x00000040 /* do not dump file */
367 #define FS_NOATIME_FL 0x00000080 /* do not update atime */
368 /* Reserved for compression usage... */
369 #define FS_DIRTY_FL 0x00000100
370 #define FS_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
371 #define FS_NOCOMP_FL 0x00000400 /* Don't compress */
372 #define FS_ECOMPR_FL 0x00000800 /* Compression error */
373 /* End compression flags --- maybe not all used */
374 #define FS_BTREE_FL 0x00001000 /* btree format dir */
375 #define FS_INDEX_FL 0x00001000 /* hash-indexed directory */
376 #define FS_IMAGIC_FL 0x00002000 /* AFS directory */
377 #define FS_JOURNAL_DATA_FL 0x00004000 /* Reserved for ext3 */
378 #define FS_NOTAIL_FL 0x00008000 /* file tail should not be merged */
379 #define FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
380 #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
381 #define FS_EXTENT_FL 0x00080000 /* Extents */
382 #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */
383 #define FS_NOCOW_FL 0x00800000 /* Do not cow file */
384 #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */
386 #define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
387 #define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
390 #define SYNC_FILE_RANGE_WAIT_BEFORE 1
391 #define SYNC_FILE_RANGE_WRITE 2
392 #define SYNC_FILE_RANGE_WAIT_AFTER 4
396 #include <linux/linkage.h>
397 #include <linux/wait.h>
398 #include <linux/kdev_t.h>
399 #include <linux/dcache.h>
400 #include <linux/path.h>
401 #include <linux/stat.h>
402 #include <linux/cache.h>
403 #include <linux/list.h>
404 #include <linux/radix-tree.h>
405 #include <linux/rbtree.h>
406 #include <linux/init.h>
407 #include <linux/pid.h>
408 #include <linux/bug.h>
409 #include <linux/mutex.h>
410 #include <linux/capability.h>
411 #include <linux/semaphore.h>
412 #include <linux/fiemap.h>
413 #include <linux/rculist_bl.h>
414 #include <linux/atomic.h>
415 #include <linux/shrinker.h>
416 #include <linux/migrate_mode.h>
417 #include <linux/uidgid.h>
418 #include <linux/lockdep.h>
419 #include <linux/percpu-rwsem.h>
421 #include <asm/byteorder.h>
423 struct export_operations
;
429 struct pipe_inode_info
;
430 struct poll_table_struct
;
432 struct vm_area_struct
;
435 struct swap_info_struct
;
437 extern void __init
inode_init(void);
438 extern void __init
inode_init_early(void);
439 extern void __init
files_init(unsigned long);
441 extern struct files_stat_struct files_stat
;
442 extern unsigned long get_max_files(void);
443 extern int sysctl_nr_open
;
444 extern struct inodes_stat_t inodes_stat
;
445 extern int leases_enable
, lease_break_time
;
446 extern int sysctl_protected_symlinks
;
447 extern int sysctl_protected_hardlinks
;
450 typedef int (get_block_t
)(struct inode
*inode
, sector_t iblock
,
451 struct buffer_head
*bh_result
, int create
);
452 typedef void (dio_iodone_t
)(struct kiocb
*iocb
, loff_t offset
,
453 ssize_t bytes
, void *private, int ret
,
457 * Attribute flags. These should be or-ed together to figure out what
460 #define ATTR_MODE (1 << 0)
461 #define ATTR_UID (1 << 1)
462 #define ATTR_GID (1 << 2)
463 #define ATTR_SIZE (1 << 3)
464 #define ATTR_ATIME (1 << 4)
465 #define ATTR_MTIME (1 << 5)
466 #define ATTR_CTIME (1 << 6)
467 #define ATTR_ATIME_SET (1 << 7)
468 #define ATTR_MTIME_SET (1 << 8)
469 #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
470 #define ATTR_ATTR_FLAG (1 << 10)
471 #define ATTR_KILL_SUID (1 << 11)
472 #define ATTR_KILL_SGID (1 << 12)
473 #define ATTR_FILE (1 << 13)
474 #define ATTR_KILL_PRIV (1 << 14)
475 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
476 #define ATTR_TIMES_SET (1 << 16)
479 * This is the Inode Attributes structure, used for notify_change(). It
480 * uses the above definitions as flags, to know which values have changed.
481 * Also, in this manner, a Filesystem can look at only the values it cares
482 * about. Basically, these are the attributes that the VFS layer can
483 * request to change from the FS layer.
485 * Derek Atkins <warlord@MIT.EDU> 94-10-20
488 unsigned int ia_valid
;
493 struct timespec ia_atime
;
494 struct timespec ia_mtime
;
495 struct timespec ia_ctime
;
498 * Not an attribute, but an auxiliary info for filesystems wanting to
499 * implement an ftruncate() like method. NOTE: filesystem should
500 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
502 struct file
*ia_file
;
506 * Includes for diskquotas.
508 #include <linux/quota.h>
511 * enum positive_aop_returns - aop return codes with specific semantics
513 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
514 * completed, that the page is still locked, and
515 * should be considered active. The VM uses this hint
516 * to return the page to the active list -- it won't
517 * be a candidate for writeback again in the near
518 * future. Other callers must be careful to unlock
519 * the page if they get this return. Returned by
522 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
523 * unlocked it and the page might have been truncated.
524 * The caller should back up to acquiring a new page and
525 * trying again. The aop will be taking reasonable
526 * precautions not to livelock. If the caller held a page
527 * reference, it should drop it before retrying. Returned
530 * address_space_operation functions return these large constants to indicate
531 * special semantics to the caller. These are much larger than the bytes in a
532 * page to allow for functions that return the number of bytes operated on in a
536 enum positive_aop_returns
{
537 AOP_WRITEPAGE_ACTIVATE
= 0x80000,
538 AOP_TRUNCATED_PAGE
= 0x80001,
541 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
542 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
543 #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
544 * helper code (eg buffer layer)
545 * to clear GFP_FS from alloc */
548 * oh the beauties of C type declarations.
551 struct address_space
;
552 struct writeback_control
;
555 const struct iovec
*iov
;
556 unsigned long nr_segs
;
561 size_t iov_iter_copy_from_user_atomic(struct page
*page
,
562 struct iov_iter
*i
, unsigned long offset
, size_t bytes
);
563 size_t iov_iter_copy_from_user(struct page
*page
,
564 struct iov_iter
*i
, unsigned long offset
, size_t bytes
);
565 void iov_iter_advance(struct iov_iter
*i
, size_t bytes
);
566 int iov_iter_fault_in_readable(struct iov_iter
*i
, size_t bytes
);
567 size_t iov_iter_single_seg_count(struct iov_iter
*i
);
569 static inline void iov_iter_init(struct iov_iter
*i
,
570 const struct iovec
*iov
, unsigned long nr_segs
,
571 size_t count
, size_t written
)
574 i
->nr_segs
= nr_segs
;
576 i
->count
= count
+ written
;
578 iov_iter_advance(i
, written
);
581 static inline size_t iov_iter_count(struct iov_iter
*i
)
587 * "descriptor" for what we're up to with a read.
588 * This allows us to use the same read code yet
589 * have multiple different users of the data that
590 * we read from a file.
592 * The simplest case just copies the data to user
605 typedef int (*read_actor_t
)(read_descriptor_t
*, struct page
*,
606 unsigned long, unsigned long);
608 struct address_space_operations
{
609 int (*writepage
)(struct page
*page
, struct writeback_control
*wbc
);
610 int (*readpage
)(struct file
*, struct page
*);
612 /* Write back some dirty pages from this mapping. */
613 int (*writepages
)(struct address_space
*, struct writeback_control
*);
615 /* Set a page dirty. Return true if this dirtied it */
616 int (*set_page_dirty
)(struct page
*page
);
618 int (*readpages
)(struct file
*filp
, struct address_space
*mapping
,
619 struct list_head
*pages
, unsigned nr_pages
);
621 int (*write_begin
)(struct file
*, struct address_space
*mapping
,
622 loff_t pos
, unsigned len
, unsigned flags
,
623 struct page
**pagep
, void **fsdata
);
624 int (*write_end
)(struct file
*, struct address_space
*mapping
,
625 loff_t pos
, unsigned len
, unsigned copied
,
626 struct page
*page
, void *fsdata
);
628 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
629 sector_t (*bmap
)(struct address_space
*, sector_t
);
630 void (*invalidatepage
) (struct page
*, unsigned long);
631 int (*releasepage
) (struct page
*, gfp_t
);
632 void (*freepage
)(struct page
*);
633 ssize_t (*direct_IO
)(int, struct kiocb
*, const struct iovec
*iov
,
634 loff_t offset
, unsigned long nr_segs
);
635 int (*get_xip_mem
)(struct address_space
*, pgoff_t
, int,
636 void **, unsigned long *);
638 * migrate the contents of a page to the specified target. If sync
639 * is false, it must not block.
641 int (*migratepage
) (struct address_space
*,
642 struct page
*, struct page
*, enum migrate_mode
);
643 int (*launder_page
) (struct page
*);
644 int (*is_partially_uptodate
) (struct page
*, read_descriptor_t
*,
646 int (*error_remove_page
)(struct address_space
*, struct page
*);
648 /* swapfile support */
649 int (*swap_activate
)(struct swap_info_struct
*sis
, struct file
*file
,
651 void (*swap_deactivate
)(struct file
*file
);
654 extern const struct address_space_operations empty_aops
;
657 * pagecache_write_begin/pagecache_write_end must be used by general code
658 * to write into the pagecache.
660 int pagecache_write_begin(struct file
*, struct address_space
*mapping
,
661 loff_t pos
, unsigned len
, unsigned flags
,
662 struct page
**pagep
, void **fsdata
);
664 int pagecache_write_end(struct file
*, struct address_space
*mapping
,
665 loff_t pos
, unsigned len
, unsigned copied
,
666 struct page
*page
, void *fsdata
);
668 struct backing_dev_info
;
669 struct address_space
{
670 struct inode
*host
; /* owner: inode, block_device */
671 struct radix_tree_root page_tree
; /* radix tree of all pages */
672 spinlock_t tree_lock
; /* and lock protecting it */
673 unsigned int i_mmap_writable
;/* count VM_SHARED mappings */
674 struct rb_root i_mmap
; /* tree of private and shared mappings */
675 struct list_head i_mmap_nonlinear
;/*list VM_NONLINEAR mappings */
676 struct mutex i_mmap_mutex
; /* protect tree, count, list */
677 /* Protected by tree_lock together with the radix tree */
678 unsigned long nrpages
; /* number of total pages */
679 pgoff_t writeback_index
;/* writeback starts here */
680 const struct address_space_operations
*a_ops
; /* methods */
681 unsigned long flags
; /* error bits/gfp mask */
682 struct backing_dev_info
*backing_dev_info
; /* device readahead, etc */
683 spinlock_t private_lock
; /* for use by the address_space */
684 struct list_head private_list
; /* ditto */
685 struct address_space
*assoc_mapping
; /* ditto */
686 } __attribute__((aligned(sizeof(long))));
688 * On most architectures that alignment is already the case; but
689 * must be enforced here for CRIS, to let the least significant bit
690 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
692 struct request_queue
;
694 struct block_device
{
695 dev_t bd_dev
; /* not a kdev_t - it's a search key */
697 struct inode
* bd_inode
; /* will die */
698 struct super_block
* bd_super
;
699 struct mutex bd_mutex
; /* open/close mutex */
700 struct list_head bd_inodes
;
704 bool bd_write_holder
;
706 struct list_head bd_holder_disks
;
708 struct block_device
* bd_contains
;
709 unsigned bd_block_size
;
710 struct hd_struct
* bd_part
;
711 /* number of times partitions within this device have been opened. */
712 unsigned bd_part_count
;
714 struct gendisk
* bd_disk
;
715 struct request_queue
* bd_queue
;
716 struct list_head bd_list
;
718 * Private data. You must have bd_claim'ed the block_device
719 * to use this. NOTE: bd_claim allows an owner to claim
720 * the same device multiple times, the owner must take special
721 * care to not mess up bd_private for that case.
723 unsigned long bd_private
;
725 /* The counter of freeze processes */
726 int bd_fsfreeze_count
;
727 /* Mutex for freeze */
728 struct mutex bd_fsfreeze_mutex
;
729 /* A semaphore that prevents I/O while block size is being changed */
730 struct percpu_rw_semaphore bd_block_size_semaphore
;
734 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
737 #define PAGECACHE_TAG_DIRTY 0
738 #define PAGECACHE_TAG_WRITEBACK 1
739 #define PAGECACHE_TAG_TOWRITE 2
741 int mapping_tagged(struct address_space
*mapping
, int tag
);
744 * Might pages of this file be mapped into userspace?
746 static inline int mapping_mapped(struct address_space
*mapping
)
748 return !RB_EMPTY_ROOT(&mapping
->i_mmap
) ||
749 !list_empty(&mapping
->i_mmap_nonlinear
);
753 * Might pages of this file have been modified in userspace?
754 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
755 * marks vma as VM_SHARED if it is shared, and the file was opened for
756 * writing i.e. vma may be mprotected writable even if now readonly.
758 static inline int mapping_writably_mapped(struct address_space
*mapping
)
760 return mapping
->i_mmap_writable
!= 0;
764 * Use sequence counter to get consistent i_size on 32-bit processors.
766 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
767 #include <linux/seqlock.h>
768 #define __NEED_I_SIZE_ORDERED
769 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
771 #define i_size_ordered_init(inode) do { } while (0)
775 #define ACL_NOT_CACHED ((void *)(-1))
777 #define IOP_FASTPERM 0x0001
778 #define IOP_LOOKUP 0x0002
779 #define IOP_NOFOLLOW 0x0004
782 * Keep mostly read-only and often accessed (especially for
783 * the RCU path lookup and 'stat' data) fields at the beginning
784 * of the 'struct inode'
788 unsigned short i_opflags
;
791 unsigned int i_flags
;
793 #ifdef CONFIG_FS_POSIX_ACL
794 struct posix_acl
*i_acl
;
795 struct posix_acl
*i_default_acl
;
798 const struct inode_operations
*i_op
;
799 struct super_block
*i_sb
;
800 struct address_space
*i_mapping
;
802 #ifdef CONFIG_SECURITY
806 /* Stat data, not accessed from path walking */
809 * Filesystems may only read i_nlink directly. They shall use the
810 * following functions for modification:
812 * (set|clear|inc|drop)_nlink
813 * inode_(inc|dec)_link_count
816 const unsigned int i_nlink
;
817 unsigned int __i_nlink
;
821 struct timespec i_atime
;
822 struct timespec i_mtime
;
823 struct timespec i_ctime
;
824 spinlock_t i_lock
; /* i_blocks, i_bytes, maybe i_size */
825 unsigned short i_bytes
;
826 unsigned int i_blkbits
;
829 #ifdef __NEED_I_SIZE_ORDERED
830 seqcount_t i_size_seqcount
;
834 unsigned long i_state
;
835 struct mutex i_mutex
;
837 unsigned long dirtied_when
; /* jiffies of first dirtying */
839 struct hlist_node i_hash
;
840 struct list_head i_wb_list
; /* backing dev IO list */
841 struct list_head i_lru
; /* inode LRU list */
842 struct list_head i_sb_list
;
844 struct hlist_head i_dentry
;
845 struct rcu_head i_rcu
;
849 atomic_t i_dio_count
;
850 atomic_t i_writecount
;
851 const struct file_operations
*i_fop
; /* former ->i_op->default_file_ops */
852 struct file_lock
*i_flock
;
853 struct address_space i_data
;
855 struct dquot
*i_dquot
[MAXQUOTAS
];
857 struct list_head i_devices
;
859 struct pipe_inode_info
*i_pipe
;
860 struct block_device
*i_bdev
;
866 #ifdef CONFIG_FSNOTIFY
867 __u32 i_fsnotify_mask
; /* all events this inode cares about */
868 struct hlist_head i_fsnotify_marks
;
872 atomic_t i_readcount
; /* struct files open RO */
874 void *i_private
; /* fs or device private pointer */
877 static inline int inode_unhashed(struct inode
*inode
)
879 return hlist_unhashed(&inode
->i_hash
);
883 * inode->i_mutex nesting subclasses for the lock validator:
885 * 0: the object of the current VFS operation
890 * The locking order between these classes is
891 * parent -> child -> normal -> xattr -> quota
893 enum inode_i_mutex_lock_class
903 * NOTE: in a 32bit arch with a preemptable kernel and
904 * an UP compile the i_size_read/write must be atomic
905 * with respect to the local cpu (unlike with preempt disabled),
906 * but they don't need to be atomic with respect to other cpus like in
907 * true SMP (so they need either to either locally disable irq around
908 * the read or for example on x86 they can be still implemented as a
909 * cmpxchg8b without the need of the lock prefix). For SMP compiles
910 * and 64bit archs it makes no difference if preempt is enabled or not.
912 static inline loff_t
i_size_read(const struct inode
*inode
)
914 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
919 seq
= read_seqcount_begin(&inode
->i_size_seqcount
);
920 i_size
= inode
->i_size
;
921 } while (read_seqcount_retry(&inode
->i_size_seqcount
, seq
));
923 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
927 i_size
= inode
->i_size
;
931 return inode
->i_size
;
936 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
937 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
938 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
940 static inline void i_size_write(struct inode
*inode
, loff_t i_size
)
942 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
943 write_seqcount_begin(&inode
->i_size_seqcount
);
944 inode
->i_size
= i_size
;
945 write_seqcount_end(&inode
->i_size_seqcount
);
946 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
948 inode
->i_size
= i_size
;
951 inode
->i_size
= i_size
;
955 /* Helper functions so that in most cases filesystems will
956 * not need to deal directly with kuid_t and kgid_t and can
957 * instead deal with the raw numeric values that are stored
960 static inline uid_t
i_uid_read(const struct inode
*inode
)
962 return from_kuid(&init_user_ns
, inode
->i_uid
);
965 static inline gid_t
i_gid_read(const struct inode
*inode
)
967 return from_kgid(&init_user_ns
, inode
->i_gid
);
970 static inline void i_uid_write(struct inode
*inode
, uid_t uid
)
972 inode
->i_uid
= make_kuid(&init_user_ns
, uid
);
975 static inline void i_gid_write(struct inode
*inode
, gid_t gid
)
977 inode
->i_gid
= make_kgid(&init_user_ns
, gid
);
980 static inline unsigned iminor(const struct inode
*inode
)
982 return MINOR(inode
->i_rdev
);
985 static inline unsigned imajor(const struct inode
*inode
)
987 return MAJOR(inode
->i_rdev
);
990 extern struct block_device
*I_BDEV(struct inode
*inode
);
993 rwlock_t lock
; /* protects pid, uid, euid fields */
994 struct pid
*pid
; /* pid or -pgrp where SIGIO should be sent */
995 enum pid_type pid_type
; /* Kind of process group SIGIO should be sent to */
996 kuid_t uid
, euid
; /* uid/euid of process setting the owner */
997 int signum
; /* posix.1b rt signal to be delivered on IO */
1001 * Track a single file's readahead state
1003 struct file_ra_state
{
1004 pgoff_t start
; /* where readahead started */
1005 unsigned int size
; /* # of readahead pages */
1006 unsigned int async_size
; /* do asynchronous readahead when
1007 there are only # of pages ahead */
1009 unsigned int ra_pages
; /* Maximum readahead window */
1010 unsigned int mmap_miss
; /* Cache miss stat for mmap accesses */
1011 loff_t prev_pos
; /* Cache last read() position */
1015 * Check if @index falls in the readahead windows.
1017 static inline int ra_has_index(struct file_ra_state
*ra
, pgoff_t index
)
1019 return (index
>= ra
->start
&&
1020 index
< ra
->start
+ ra
->size
);
1023 #define FILE_MNT_WRITE_TAKEN 1
1024 #define FILE_MNT_WRITE_RELEASED 2
1028 * fu_list becomes invalid after file_free is called and queued via
1029 * fu_rcuhead for RCU freeing
1032 struct list_head fu_list
;
1033 struct rcu_head fu_rcuhead
;
1036 #define f_dentry f_path.dentry
1037 #define f_vfsmnt f_path.mnt
1038 const struct file_operations
*f_op
;
1041 * Protects f_ep_links, f_flags, f_pos vs i_size in lseek SEEK_CUR.
1042 * Must not be taken from IRQ context.
1048 atomic_long_t f_count
;
1049 unsigned int f_flags
;
1052 struct fown_struct f_owner
;
1053 const struct cred
*f_cred
;
1054 struct file_ra_state f_ra
;
1057 #ifdef CONFIG_SECURITY
1060 /* needed for tty driver, and maybe others */
1064 /* Used by fs/eventpoll.c to link all the hooks to this file */
1065 struct list_head f_ep_links
;
1066 struct list_head f_tfile_llink
;
1067 #endif /* #ifdef CONFIG_EPOLL */
1068 struct address_space
*f_mapping
;
1069 #ifdef CONFIG_DEBUG_WRITECOUNT
1070 unsigned long f_mnt_write_state
;
1074 struct file_handle
{
1077 /* file identifier */
1078 unsigned char f_handle
[0];
1081 static inline struct file
*get_file(struct file
*f
)
1083 atomic_long_inc(&f
->f_count
);
1086 #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
1087 #define file_count(x) atomic_long_read(&(x)->f_count)
1089 #ifdef CONFIG_DEBUG_WRITECOUNT
1090 static inline void file_take_write(struct file
*f
)
1092 WARN_ON(f
->f_mnt_write_state
!= 0);
1093 f
->f_mnt_write_state
= FILE_MNT_WRITE_TAKEN
;
1095 static inline void file_release_write(struct file
*f
)
1097 f
->f_mnt_write_state
|= FILE_MNT_WRITE_RELEASED
;
1099 static inline void file_reset_write(struct file
*f
)
1101 f
->f_mnt_write_state
= 0;
1103 static inline void file_check_state(struct file
*f
)
1106 * At this point, either both or neither of these bits
1109 WARN_ON(f
->f_mnt_write_state
== FILE_MNT_WRITE_TAKEN
);
1110 WARN_ON(f
->f_mnt_write_state
== FILE_MNT_WRITE_RELEASED
);
1112 static inline int file_check_writeable(struct file
*f
)
1114 if (f
->f_mnt_write_state
== FILE_MNT_WRITE_TAKEN
)
1116 printk(KERN_WARNING
"writeable file with no "
1117 "mnt_want_write()\n");
1121 #else /* !CONFIG_DEBUG_WRITECOUNT */
1122 static inline void file_take_write(struct file
*filp
) {}
1123 static inline void file_release_write(struct file
*filp
) {}
1124 static inline void file_reset_write(struct file
*filp
) {}
1125 static inline void file_check_state(struct file
*filp
) {}
1126 static inline int file_check_writeable(struct file
*filp
)
1130 #endif /* CONFIG_DEBUG_WRITECOUNT */
1132 #define MAX_NON_LFS ((1UL<<31) - 1)
1134 /* Page cache limit. The filesystems should put that into their s_maxbytes
1135 limits, otherwise bad things can happen in VM. */
1136 #if BITS_PER_LONG==32
1137 #define MAX_LFS_FILESIZE (((loff_t)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
1138 #elif BITS_PER_LONG==64
1139 #define MAX_LFS_FILESIZE ((loff_t)0x7fffffffffffffff)
1144 #define FL_ACCESS 8 /* not trying to lock, just looking */
1145 #define FL_EXISTS 16 /* when unlocking, test for existence */
1146 #define FL_LEASE 32 /* lease held on this file */
1147 #define FL_CLOSE 64 /* unlock on close */
1148 #define FL_SLEEP 128 /* A blocking lock */
1149 #define FL_DOWNGRADE_PENDING 256 /* Lease is being downgraded */
1150 #define FL_UNLOCK_PENDING 512 /* Lease is being broken */
1153 * Special return value from posix_lock_file() and vfs_lock_file() for
1154 * asynchronous locking.
1156 #define FILE_LOCK_DEFERRED 1
1159 * The POSIX file lock owner is determined by
1160 * the "struct files_struct" in the thread group
1161 * (or NULL for no owner - BSD locks).
1163 * Lockd stuffs a "host" pointer into this.
1165 typedef struct files_struct
*fl_owner_t
;
1167 struct file_lock_operations
{
1168 void (*fl_copy_lock
)(struct file_lock
*, struct file_lock
*);
1169 void (*fl_release_private
)(struct file_lock
*);
1172 struct lock_manager_operations
{
1173 int (*lm_compare_owner
)(struct file_lock
*, struct file_lock
*);
1174 void (*lm_notify
)(struct file_lock
*); /* unblock callback */
1175 int (*lm_grant
)(struct file_lock
*, struct file_lock
*, int);
1176 void (*lm_break
)(struct file_lock
*);
1177 int (*lm_change
)(struct file_lock
**, int);
1180 struct lock_manager
{
1181 struct list_head list
;
1185 void locks_start_grace(struct net
*, struct lock_manager
*);
1186 void locks_end_grace(struct lock_manager
*);
1187 int locks_in_grace(struct net
*);
1189 /* that will die - we need it for nfs_lock_info */
1190 #include <linux/nfs_fs_i.h>
1193 struct file_lock
*fl_next
; /* singly linked list for this inode */
1194 struct list_head fl_link
; /* doubly linked list of all locks */
1195 struct list_head fl_block
; /* circular list of blocked processes */
1196 fl_owner_t fl_owner
;
1197 unsigned int fl_flags
;
1198 unsigned char fl_type
;
1199 unsigned int fl_pid
;
1200 struct pid
*fl_nspid
;
1201 wait_queue_head_t fl_wait
;
1202 struct file
*fl_file
;
1206 struct fasync_struct
* fl_fasync
; /* for lease break notifications */
1207 /* for lease breaks: */
1208 unsigned long fl_break_time
;
1209 unsigned long fl_downgrade_time
;
1211 const struct file_lock_operations
*fl_ops
; /* Callbacks for filesystems */
1212 const struct lock_manager_operations
*fl_lmops
; /* Callbacks for lockmanagers */
1214 struct nfs_lock_info nfs_fl
;
1215 struct nfs4_lock_info nfs4_fl
;
1217 struct list_head link
; /* link in AFS vnode's pending_locks list */
1218 int state
; /* state of grant or error if -ve */
1223 /* The following constant reflects the upper bound of the file/locking space */
1225 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
1226 #define OFFSET_MAX INT_LIMIT(loff_t)
1227 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
1230 #include <linux/fcntl.h>
1232 extern void send_sigio(struct fown_struct
*fown
, int fd
, int band
);
1234 #ifdef CONFIG_FILE_LOCKING
1235 extern int fcntl_getlk(struct file
*, struct flock __user
*);
1236 extern int fcntl_setlk(unsigned int, struct file
*, unsigned int,
1237 struct flock __user
*);
1239 #if BITS_PER_LONG == 32
1240 extern int fcntl_getlk64(struct file
*, struct flock64 __user
*);
1241 extern int fcntl_setlk64(unsigned int, struct file
*, unsigned int,
1242 struct flock64 __user
*);
1245 extern int fcntl_setlease(unsigned int fd
, struct file
*filp
, long arg
);
1246 extern int fcntl_getlease(struct file
*filp
);
1249 void locks_free_lock(struct file_lock
*fl
);
1250 extern void locks_init_lock(struct file_lock
*);
1251 extern struct file_lock
* locks_alloc_lock(void);
1252 extern void locks_copy_lock(struct file_lock
*, struct file_lock
*);
1253 extern void __locks_copy_lock(struct file_lock
*, const struct file_lock
*);
1254 extern void locks_remove_posix(struct file
*, fl_owner_t
);
1255 extern void locks_remove_flock(struct file
*);
1256 extern void locks_release_private(struct file_lock
*);
1257 extern void posix_test_lock(struct file
*, struct file_lock
*);
1258 extern int posix_lock_file(struct file
*, struct file_lock
*, struct file_lock
*);
1259 extern int posix_lock_file_wait(struct file
*, struct file_lock
*);
1260 extern int posix_unblock_lock(struct file
*, struct file_lock
*);
1261 extern int vfs_test_lock(struct file
*, struct file_lock
*);
1262 extern int vfs_lock_file(struct file
*, unsigned int, struct file_lock
*, struct file_lock
*);
1263 extern int vfs_cancel_lock(struct file
*filp
, struct file_lock
*fl
);
1264 extern int flock_lock_file_wait(struct file
*filp
, struct file_lock
*fl
);
1265 extern int __break_lease(struct inode
*inode
, unsigned int flags
);
1266 extern void lease_get_mtime(struct inode
*, struct timespec
*time
);
1267 extern int generic_setlease(struct file
*, long, struct file_lock
**);
1268 extern int vfs_setlease(struct file
*, long, struct file_lock
**);
1269 extern int lease_modify(struct file_lock
**, int);
1270 extern int lock_may_read(struct inode
*, loff_t start
, unsigned long count
);
1271 extern int lock_may_write(struct inode
*, loff_t start
, unsigned long count
);
1272 extern void locks_delete_block(struct file_lock
*waiter
);
1273 extern void lock_flocks(void);
1274 extern void unlock_flocks(void);
1275 #else /* !CONFIG_FILE_LOCKING */
1276 static inline int fcntl_getlk(struct file
*file
, struct flock __user
*user
)
1281 static inline int fcntl_setlk(unsigned int fd
, struct file
*file
,
1282 unsigned int cmd
, struct flock __user
*user
)
1287 #if BITS_PER_LONG == 32
1288 static inline int fcntl_getlk64(struct file
*file
, struct flock64 __user
*user
)
1293 static inline int fcntl_setlk64(unsigned int fd
, struct file
*file
,
1294 unsigned int cmd
, struct flock64 __user
*user
)
1299 static inline int fcntl_setlease(unsigned int fd
, struct file
*filp
, long arg
)
1304 static inline int fcntl_getlease(struct file
*filp
)
1309 static inline void locks_init_lock(struct file_lock
*fl
)
1314 static inline void __locks_copy_lock(struct file_lock
*new, struct file_lock
*fl
)
1319 static inline void locks_copy_lock(struct file_lock
*new, struct file_lock
*fl
)
1324 static inline void locks_remove_posix(struct file
*filp
, fl_owner_t owner
)
1329 static inline void locks_remove_flock(struct file
*filp
)
1334 static inline void posix_test_lock(struct file
*filp
, struct file_lock
*fl
)
1339 static inline int posix_lock_file(struct file
*filp
, struct file_lock
*fl
,
1340 struct file_lock
*conflock
)
1345 static inline int posix_lock_file_wait(struct file
*filp
, struct file_lock
*fl
)
1350 static inline int posix_unblock_lock(struct file
*filp
,
1351 struct file_lock
*waiter
)
1356 static inline int vfs_test_lock(struct file
*filp
, struct file_lock
*fl
)
1361 static inline int vfs_lock_file(struct file
*filp
, unsigned int cmd
,
1362 struct file_lock
*fl
, struct file_lock
*conf
)
1367 static inline int vfs_cancel_lock(struct file
*filp
, struct file_lock
*fl
)
1372 static inline int flock_lock_file_wait(struct file
*filp
,
1373 struct file_lock
*request
)
1378 static inline int __break_lease(struct inode
*inode
, unsigned int mode
)
1383 static inline void lease_get_mtime(struct inode
*inode
, struct timespec
*time
)
1388 static inline int generic_setlease(struct file
*filp
, long arg
,
1389 struct file_lock
**flp
)
1394 static inline int vfs_setlease(struct file
*filp
, long arg
,
1395 struct file_lock
**lease
)
1400 static inline int lease_modify(struct file_lock
**before
, int arg
)
1405 static inline int lock_may_read(struct inode
*inode
, loff_t start
,
1411 static inline int lock_may_write(struct inode
*inode
, loff_t start
,
1417 static inline void locks_delete_block(struct file_lock
*waiter
)
1421 static inline void lock_flocks(void)
1425 static inline void unlock_flocks(void)
1429 #endif /* !CONFIG_FILE_LOCKING */
1432 struct fasync_struct
{
1436 struct fasync_struct
*fa_next
; /* singly linked list */
1437 struct file
*fa_file
;
1438 struct rcu_head fa_rcu
;
1441 #define FASYNC_MAGIC 0x4601
1443 /* SMP safe fasync helpers: */
1444 extern int fasync_helper(int, struct file
*, int, struct fasync_struct
**);
1445 extern struct fasync_struct
*fasync_insert_entry(int, struct file
*, struct fasync_struct
**, struct fasync_struct
*);
1446 extern int fasync_remove_entry(struct file
*, struct fasync_struct
**);
1447 extern struct fasync_struct
*fasync_alloc(void);
1448 extern void fasync_free(struct fasync_struct
*);
1450 /* can be called from interrupts */
1451 extern void kill_fasync(struct fasync_struct
**, int, int);
1453 extern int __f_setown(struct file
*filp
, struct pid
*, enum pid_type
, int force
);
1454 extern int f_setown(struct file
*filp
, unsigned long arg
, int force
);
1455 extern void f_delown(struct file
*filp
);
1456 extern pid_t
f_getown(struct file
*filp
);
1457 extern int send_sigurg(struct fown_struct
*fown
);
1465 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1466 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1467 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1468 #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1469 #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
1471 extern struct list_head super_blocks
;
1472 extern spinlock_t sb_lock
;
1474 /* Possible states of 'frozen' field */
1476 SB_UNFROZEN
= 0, /* FS is unfrozen */
1477 SB_FREEZE_WRITE
= 1, /* Writes, dir ops, ioctls frozen */
1478 SB_FREEZE_PAGEFAULT
= 2, /* Page faults stopped as well */
1479 SB_FREEZE_FS
= 3, /* For internal FS use (e.g. to stop
1480 * internal threads if needed) */
1481 SB_FREEZE_COMPLETE
= 4, /* ->freeze_fs finished successfully */
1484 #define SB_FREEZE_LEVELS (SB_FREEZE_COMPLETE - 1)
1487 /* Counters for counting writers at each level */
1488 struct percpu_counter counter
[SB_FREEZE_LEVELS
];
1489 wait_queue_head_t wait
; /* queue for waiting for
1490 writers / faults to finish */
1491 int frozen
; /* Is sb frozen? */
1492 wait_queue_head_t wait_unfrozen
; /* queue for waiting for
1494 #ifdef CONFIG_DEBUG_LOCK_ALLOC
1495 struct lockdep_map lock_map
[SB_FREEZE_LEVELS
];
1499 struct super_block
{
1500 struct list_head s_list
; /* Keep this first */
1501 dev_t s_dev
; /* search index; _not_ kdev_t */
1502 unsigned char s_blocksize_bits
;
1503 unsigned long s_blocksize
;
1504 loff_t s_maxbytes
; /* Max file size */
1505 struct file_system_type
*s_type
;
1506 const struct super_operations
*s_op
;
1507 const struct dquot_operations
*dq_op
;
1508 const struct quotactl_ops
*s_qcop
;
1509 const struct export_operations
*s_export_op
;
1510 unsigned long s_flags
;
1511 unsigned long s_magic
;
1512 struct dentry
*s_root
;
1513 struct rw_semaphore s_umount
;
1514 struct mutex s_lock
;
1517 #ifdef CONFIG_SECURITY
1520 const struct xattr_handler
**s_xattr
;
1522 struct list_head s_inodes
; /* all inodes */
1523 struct hlist_bl_head s_anon
; /* anonymous dentries for (nfs) exporting */
1525 struct list_head __percpu
*s_files
;
1527 struct list_head s_files
;
1529 struct list_head s_mounts
; /* list of mounts; _not_ for fs use */
1530 /* s_dentry_lru, s_nr_dentry_unused protected by dcache.c lru locks */
1531 struct list_head s_dentry_lru
; /* unused dentry lru */
1532 int s_nr_dentry_unused
; /* # of dentry on lru */
1534 /* s_inode_lru_lock protects s_inode_lru and s_nr_inodes_unused */
1535 spinlock_t s_inode_lru_lock ____cacheline_aligned_in_smp
;
1536 struct list_head s_inode_lru
; /* unused inode lru */
1537 int s_nr_inodes_unused
; /* # of inodes on lru */
1539 struct block_device
*s_bdev
;
1540 struct backing_dev_info
*s_bdi
;
1541 struct mtd_info
*s_mtd
;
1542 struct hlist_node s_instances
;
1543 struct quota_info s_dquot
; /* Diskquota specific options */
1545 struct sb_writers s_writers
;
1547 char s_id
[32]; /* Informational name */
1548 u8 s_uuid
[16]; /* UUID */
1550 void *s_fs_info
; /* Filesystem private info */
1551 unsigned int s_max_links
;
1554 /* Granularity of c/m/atime in ns.
1555 Cannot be worse than a second */
1559 * The next field is for VFS *only*. No filesystems have any business
1560 * even looking at it. You had been warned.
1562 struct mutex s_vfs_rename_mutex
; /* Kludge */
1565 * Filesystem subtype. If non-empty the filesystem type field
1566 * in /proc/mounts will be "type.subtype"
1571 * Saved mount options for lazy filesystems using
1572 * generic_show_options()
1574 char __rcu
*s_options
;
1575 const struct dentry_operations
*s_d_op
; /* default d_op for dentries */
1578 * Saved pool identifier for cleancache (-1 means none)
1580 int cleancache_poolid
;
1582 struct shrinker s_shrink
; /* per-sb shrinker handle */
1584 /* Number of inodes with nlink == 0 but still referenced */
1585 atomic_long_t s_remove_count
;
1587 /* Being remounted read-only */
1588 int s_readonly_remount
;
1591 /* superblock cache pruning functions */
1592 extern void prune_icache_sb(struct super_block
*sb
, int nr_to_scan
);
1593 extern void prune_dcache_sb(struct super_block
*sb
, int nr_to_scan
);
1595 extern struct timespec
current_fs_time(struct super_block
*sb
);
1598 * Snapshotting support.
1601 void __sb_end_write(struct super_block
*sb
, int level
);
1602 int __sb_start_write(struct super_block
*sb
, int level
, bool wait
);
1605 * sb_end_write - drop write access to a superblock
1606 * @sb: the super we wrote to
1608 * Decrement number of writers to the filesystem. Wake up possible waiters
1609 * wanting to freeze the filesystem.
1611 static inline void sb_end_write(struct super_block
*sb
)
1613 __sb_end_write(sb
, SB_FREEZE_WRITE
);
1617 * sb_end_pagefault - drop write access to a superblock from a page fault
1618 * @sb: the super we wrote to
1620 * Decrement number of processes handling write page fault to the filesystem.
1621 * Wake up possible waiters wanting to freeze the filesystem.
1623 static inline void sb_end_pagefault(struct super_block
*sb
)
1625 __sb_end_write(sb
, SB_FREEZE_PAGEFAULT
);
1629 * sb_end_intwrite - drop write access to a superblock for internal fs purposes
1630 * @sb: the super we wrote to
1632 * Decrement fs-internal number of writers to the filesystem. Wake up possible
1633 * waiters wanting to freeze the filesystem.
1635 static inline void sb_end_intwrite(struct super_block
*sb
)
1637 __sb_end_write(sb
, SB_FREEZE_FS
);
1641 * sb_start_write - get write access to a superblock
1642 * @sb: the super we write to
1644 * When a process wants to write data or metadata to a file system (i.e. dirty
1645 * a page or an inode), it should embed the operation in a sb_start_write() -
1646 * sb_end_write() pair to get exclusion against file system freezing. This
1647 * function increments number of writers preventing freezing. If the file
1648 * system is already frozen, the function waits until the file system is
1651 * Since freeze protection behaves as a lock, users have to preserve
1652 * ordering of freeze protection and other filesystem locks. Generally,
1653 * freeze protection should be the outermost lock. In particular, we have:
1656 * -> i_mutex (write path, truncate, directory ops, ...)
1657 * -> s_umount (freeze_super, thaw_super)
1659 static inline void sb_start_write(struct super_block
*sb
)
1661 __sb_start_write(sb
, SB_FREEZE_WRITE
, true);
1664 static inline int sb_start_write_trylock(struct super_block
*sb
)
1666 return __sb_start_write(sb
, SB_FREEZE_WRITE
, false);
1670 * sb_start_pagefault - get write access to a superblock from a page fault
1671 * @sb: the super we write to
1673 * When a process starts handling write page fault, it should embed the
1674 * operation into sb_start_pagefault() - sb_end_pagefault() pair to get
1675 * exclusion against file system freezing. This is needed since the page fault
1676 * is going to dirty a page. This function increments number of running page
1677 * faults preventing freezing. If the file system is already frozen, the
1678 * function waits until the file system is thawed.
1680 * Since page fault freeze protection behaves as a lock, users have to preserve
1681 * ordering of freeze protection and other filesystem locks. It is advised to
1682 * put sb_start_pagefault() close to mmap_sem in lock ordering. Page fault
1683 * handling code implies lock dependency:
1686 * -> sb_start_pagefault
1688 static inline void sb_start_pagefault(struct super_block
*sb
)
1690 __sb_start_write(sb
, SB_FREEZE_PAGEFAULT
, true);
1694 * sb_start_intwrite - get write access to a superblock for internal fs purposes
1695 * @sb: the super we write to
1697 * This is the third level of protection against filesystem freezing. It is
1698 * free for use by a filesystem. The only requirement is that it must rank
1699 * below sb_start_pagefault.
1701 * For example filesystem can call sb_start_intwrite() when starting a
1702 * transaction which somewhat eases handling of freezing for internal sources
1703 * of filesystem changes (internal fs threads, discarding preallocation on file
1706 static inline void sb_start_intwrite(struct super_block
*sb
)
1708 __sb_start_write(sb
, SB_FREEZE_FS
, true);
1712 extern bool inode_owner_or_capable(const struct inode
*inode
);
1714 /* not quite ready to be deprecated, but... */
1715 extern void lock_super(struct super_block
*);
1716 extern void unlock_super(struct super_block
*);
1719 * VFS helper functions..
1721 extern int vfs_create(struct inode
*, struct dentry
*, umode_t
, bool);
1722 extern int vfs_mkdir(struct inode
*, struct dentry
*, umode_t
);
1723 extern int vfs_mknod(struct inode
*, struct dentry
*, umode_t
, dev_t
);
1724 extern int vfs_symlink(struct inode
*, struct dentry
*, const char *);
1725 extern int vfs_link(struct dentry
*, struct inode
*, struct dentry
*);
1726 extern int vfs_rmdir(struct inode
*, struct dentry
*);
1727 extern int vfs_unlink(struct inode
*, struct dentry
*);
1728 extern int vfs_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
1731 * VFS dentry helper functions.
1733 extern void dentry_unhash(struct dentry
*dentry
);
1736 * VFS file helper functions.
1738 extern void inode_init_owner(struct inode
*inode
, const struct inode
*dir
,
1741 * VFS FS_IOC_FIEMAP helper definitions.
1743 struct fiemap_extent_info
{
1744 unsigned int fi_flags
; /* Flags as passed from user */
1745 unsigned int fi_extents_mapped
; /* Number of mapped extents */
1746 unsigned int fi_extents_max
; /* Size of fiemap_extent array */
1747 struct fiemap_extent __user
*fi_extents_start
; /* Start of
1748 fiemap_extent array */
1750 int fiemap_fill_next_extent(struct fiemap_extent_info
*info
, u64 logical
,
1751 u64 phys
, u64 len
, u32 flags
);
1752 int fiemap_check_flags(struct fiemap_extent_info
*fieinfo
, u32 fs_flags
);
1757 * NOTE! These match bits 12..15 of stat.st_mode
1758 * (ie "(i_mode >> 12) & 15").
1760 #define DT_UNKNOWN 0
1771 * This is the "filldir" function type, used by readdir() to let
1772 * the kernel specify what kind of dirent layout it wants to have.
1773 * This allows the kernel to read directories into kernel space or
1774 * to have different dirent layouts depending on the binary type.
1776 typedef int (*filldir_t
)(void *, const char *, int, loff_t
, u64
, unsigned);
1777 struct block_device_operations
;
1779 /* These macros are for out of kernel modules to test that
1780 * the kernel supports the unlocked_ioctl and compat_ioctl
1781 * fields in struct file_operations. */
1782 #define HAVE_COMPAT_IOCTL 1
1783 #define HAVE_UNLOCKED_IOCTL 1
1785 struct file_operations
{
1786 struct module
*owner
;
1787 loff_t (*llseek
) (struct file
*, loff_t
, int);
1788 ssize_t (*read
) (struct file
*, char __user
*, size_t, loff_t
*);
1789 ssize_t (*write
) (struct file
*, const char __user
*, size_t, loff_t
*);
1790 ssize_t (*aio_read
) (struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1791 ssize_t (*aio_write
) (struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1792 int (*readdir
) (struct file
*, void *, filldir_t
);
1793 unsigned int (*poll
) (struct file
*, struct poll_table_struct
*);
1794 long (*unlocked_ioctl
) (struct file
*, unsigned int, unsigned long);
1795 long (*compat_ioctl
) (struct file
*, unsigned int, unsigned long);
1796 int (*mmap
) (struct file
*, struct vm_area_struct
*);
1797 int (*open
) (struct inode
*, struct file
*);
1798 int (*flush
) (struct file
*, fl_owner_t id
);
1799 int (*release
) (struct inode
*, struct file
*);
1800 int (*fsync
) (struct file
*, loff_t
, loff_t
, int datasync
);
1801 int (*aio_fsync
) (struct kiocb
*, int datasync
);
1802 int (*fasync
) (int, struct file
*, int);
1803 int (*lock
) (struct file
*, int, struct file_lock
*);
1804 ssize_t (*sendpage
) (struct file
*, struct page
*, int, size_t, loff_t
*, int);
1805 unsigned long (*get_unmapped_area
)(struct file
*, unsigned long, unsigned long, unsigned long, unsigned long);
1806 int (*check_flags
)(int);
1807 int (*flock
) (struct file
*, int, struct file_lock
*);
1808 ssize_t (*splice_write
)(struct pipe_inode_info
*, struct file
*, loff_t
*, size_t, unsigned int);
1809 ssize_t (*splice_read
)(struct file
*, loff_t
*, struct pipe_inode_info
*, size_t, unsigned int);
1810 int (*setlease
)(struct file
*, long, struct file_lock
**);
1811 long (*fallocate
)(struct file
*file
, int mode
, loff_t offset
,
1815 struct inode_operations
{
1816 struct dentry
* (*lookup
) (struct inode
*,struct dentry
*, unsigned int);
1817 void * (*follow_link
) (struct dentry
*, struct nameidata
*);
1818 int (*permission
) (struct inode
*, int);
1819 struct posix_acl
* (*get_acl
)(struct inode
*, int);
1821 int (*readlink
) (struct dentry
*, char __user
*,int);
1822 void (*put_link
) (struct dentry
*, struct nameidata
*, void *);
1824 int (*create
) (struct inode
*,struct dentry
*, umode_t
, bool);
1825 int (*link
) (struct dentry
*,struct inode
*,struct dentry
*);
1826 int (*unlink
) (struct inode
*,struct dentry
*);
1827 int (*symlink
) (struct inode
*,struct dentry
*,const char *);
1828 int (*mkdir
) (struct inode
*,struct dentry
*,umode_t
);
1829 int (*rmdir
) (struct inode
*,struct dentry
*);
1830 int (*mknod
) (struct inode
*,struct dentry
*,umode_t
,dev_t
);
1831 int (*rename
) (struct inode
*, struct dentry
*,
1832 struct inode
*, struct dentry
*);
1833 void (*truncate
) (struct inode
*);
1834 int (*setattr
) (struct dentry
*, struct iattr
*);
1835 int (*getattr
) (struct vfsmount
*mnt
, struct dentry
*, struct kstat
*);
1836 int (*setxattr
) (struct dentry
*, const char *,const void *,size_t,int);
1837 ssize_t (*getxattr
) (struct dentry
*, const char *, void *, size_t);
1838 ssize_t (*listxattr
) (struct dentry
*, char *, size_t);
1839 int (*removexattr
) (struct dentry
*, const char *);
1840 int (*fiemap
)(struct inode
*, struct fiemap_extent_info
*, u64 start
,
1842 int (*update_time
)(struct inode
*, struct timespec
*, int);
1843 int (*atomic_open
)(struct inode
*, struct dentry
*,
1844 struct file
*, unsigned open_flag
,
1845 umode_t create_mode
, int *opened
);
1846 } ____cacheline_aligned
;
1850 ssize_t
rw_copy_check_uvector(int type
, const struct iovec __user
* uvector
,
1851 unsigned long nr_segs
, unsigned long fast_segs
,
1852 struct iovec
*fast_pointer
,
1853 struct iovec
**ret_pointer
);
1855 extern ssize_t
vfs_read(struct file
*, char __user
*, size_t, loff_t
*);
1856 extern ssize_t
vfs_write(struct file
*, const char __user
*, size_t, loff_t
*);
1857 extern ssize_t
vfs_readv(struct file
*, const struct iovec __user
*,
1858 unsigned long, loff_t
*);
1859 extern ssize_t
vfs_writev(struct file
*, const struct iovec __user
*,
1860 unsigned long, loff_t
*);
1862 struct super_operations
{
1863 struct inode
*(*alloc_inode
)(struct super_block
*sb
);
1864 void (*destroy_inode
)(struct inode
*);
1866 void (*dirty_inode
) (struct inode
*, int flags
);
1867 int (*write_inode
) (struct inode
*, struct writeback_control
*wbc
);
1868 int (*drop_inode
) (struct inode
*);
1869 void (*evict_inode
) (struct inode
*);
1870 void (*put_super
) (struct super_block
*);
1871 int (*sync_fs
)(struct super_block
*sb
, int wait
);
1872 int (*freeze_fs
) (struct super_block
*);
1873 int (*unfreeze_fs
) (struct super_block
*);
1874 int (*statfs
) (struct dentry
*, struct kstatfs
*);
1875 int (*remount_fs
) (struct super_block
*, int *, char *);
1876 void (*umount_begin
) (struct super_block
*);
1878 int (*show_options
)(struct seq_file
*, struct dentry
*);
1879 int (*show_devname
)(struct seq_file
*, struct dentry
*);
1880 int (*show_path
)(struct seq_file
*, struct dentry
*);
1881 int (*show_stats
)(struct seq_file
*, struct dentry
*);
1883 ssize_t (*quota_read
)(struct super_block
*, int, char *, size_t, loff_t
);
1884 ssize_t (*quota_write
)(struct super_block
*, int, const char *, size_t, loff_t
);
1886 int (*bdev_try_to_free_page
)(struct super_block
*, struct page
*, gfp_t
);
1887 int (*nr_cached_objects
)(struct super_block
*);
1888 void (*free_cached_objects
)(struct super_block
*, int);
1892 * Inode state bits. Protected by inode->i_lock
1894 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1895 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1897 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1898 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1899 * various stages of removing an inode.
1901 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1903 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1904 * fdatasync(). i_atime is the usual cause.
1905 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1906 * these changes separately from I_DIRTY_SYNC so that we
1907 * don't have to write inode on fdatasync() when only
1908 * mtime has changed in it.
1909 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1910 * I_NEW Serves as both a mutex and completion notification.
1911 * New inodes set I_NEW. If two processes both create
1912 * the same inode, one of them will release its inode and
1913 * wait for I_NEW to be released before returning.
1914 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1915 * also cause waiting on I_NEW, without I_NEW actually
1916 * being set. find_inode() uses this to prevent returning
1917 * nearly-dead inodes.
1918 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1919 * is zero. I_FREEING must be set when I_WILL_FREE is
1921 * I_FREEING Set when inode is about to be freed but still has dirty
1922 * pages or buffers attached or the inode itself is still
1924 * I_CLEAR Added by clear_inode(). In this state the inode is
1925 * clean and can be destroyed. Inode keeps I_FREEING.
1927 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1928 * prohibited for many purposes. iget() must wait for
1929 * the inode to be completely released, then create it
1930 * anew. Other functions will just ignore such inodes,
1931 * if appropriate. I_NEW is used for waiting.
1933 * I_SYNC Writeback of inode is running. The bit is set during
1934 * data writeback, and cleared with a wakeup on the bit
1935 * address once it is done. The bit is also used to pin
1936 * the inode in memory for flusher thread.
1938 * I_REFERENCED Marks the inode as recently references on the LRU list.
1940 * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit().
1942 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1944 #define I_DIRTY_SYNC (1 << 0)
1945 #define I_DIRTY_DATASYNC (1 << 1)
1946 #define I_DIRTY_PAGES (1 << 2)
1948 #define I_NEW (1 << __I_NEW)
1949 #define I_WILL_FREE (1 << 4)
1950 #define I_FREEING (1 << 5)
1951 #define I_CLEAR (1 << 6)
1953 #define I_SYNC (1 << __I_SYNC)
1954 #define I_REFERENCED (1 << 8)
1955 #define __I_DIO_WAKEUP 9
1956 #define I_DIO_WAKEUP (1 << I_DIO_WAKEUP)
1958 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1960 extern void __mark_inode_dirty(struct inode
*, int);
1961 static inline void mark_inode_dirty(struct inode
*inode
)
1963 __mark_inode_dirty(inode
, I_DIRTY
);
1966 static inline void mark_inode_dirty_sync(struct inode
*inode
)
1968 __mark_inode_dirty(inode
, I_DIRTY_SYNC
);
1971 extern void inc_nlink(struct inode
*inode
);
1972 extern void drop_nlink(struct inode
*inode
);
1973 extern void clear_nlink(struct inode
*inode
);
1974 extern void set_nlink(struct inode
*inode
, unsigned int nlink
);
1976 static inline void inode_inc_link_count(struct inode
*inode
)
1979 mark_inode_dirty(inode
);
1982 static inline void inode_dec_link_count(struct inode
*inode
)
1985 mark_inode_dirty(inode
);
1989 * inode_inc_iversion - increments i_version
1990 * @inode: inode that need to be updated
1992 * Every time the inode is modified, the i_version field will be incremented.
1993 * The filesystem has to be mounted with i_version flag
1996 static inline void inode_inc_iversion(struct inode
*inode
)
1998 spin_lock(&inode
->i_lock
);
2000 spin_unlock(&inode
->i_lock
);
2003 enum file_time_flags
{
2010 extern void touch_atime(struct path
*);
2011 static inline void file_accessed(struct file
*file
)
2013 if (!(file
->f_flags
& O_NOATIME
))
2014 touch_atime(&file
->f_path
);
2017 int sync_inode(struct inode
*inode
, struct writeback_control
*wbc
);
2018 int sync_inode_metadata(struct inode
*inode
, int wait
);
2020 struct file_system_type
{
2023 struct dentry
*(*mount
) (struct file_system_type
*, int,
2024 const char *, void *);
2025 void (*kill_sb
) (struct super_block
*);
2026 struct module
*owner
;
2027 struct file_system_type
* next
;
2028 struct hlist_head fs_supers
;
2030 struct lock_class_key s_lock_key
;
2031 struct lock_class_key s_umount_key
;
2032 struct lock_class_key s_vfs_rename_key
;
2033 struct lock_class_key s_writers_key
[SB_FREEZE_LEVELS
];
2035 struct lock_class_key i_lock_key
;
2036 struct lock_class_key i_mutex_key
;
2037 struct lock_class_key i_mutex_dir_key
;
2040 extern struct dentry
*mount_ns(struct file_system_type
*fs_type
, int flags
,
2041 void *data
, int (*fill_super
)(struct super_block
*, void *, int));
2042 extern struct dentry
*mount_bdev(struct file_system_type
*fs_type
,
2043 int flags
, const char *dev_name
, void *data
,
2044 int (*fill_super
)(struct super_block
*, void *, int));
2045 extern struct dentry
*mount_single(struct file_system_type
*fs_type
,
2046 int flags
, void *data
,
2047 int (*fill_super
)(struct super_block
*, void *, int));
2048 extern struct dentry
*mount_nodev(struct file_system_type
*fs_type
,
2049 int flags
, void *data
,
2050 int (*fill_super
)(struct super_block
*, void *, int));
2051 extern struct dentry
*mount_subtree(struct vfsmount
*mnt
, const char *path
);
2052 void generic_shutdown_super(struct super_block
*sb
);
2053 void kill_block_super(struct super_block
*sb
);
2054 void kill_anon_super(struct super_block
*sb
);
2055 void kill_litter_super(struct super_block
*sb
);
2056 void deactivate_super(struct super_block
*sb
);
2057 void deactivate_locked_super(struct super_block
*sb
);
2058 int set_anon_super(struct super_block
*s
, void *data
);
2059 int get_anon_bdev(dev_t
*);
2060 void free_anon_bdev(dev_t
);
2061 struct super_block
*sget(struct file_system_type
*type
,
2062 int (*test
)(struct super_block
*,void *),
2063 int (*set
)(struct super_block
*,void *),
2064 int flags
, void *data
);
2065 extern struct dentry
*mount_pseudo(struct file_system_type
*, char *,
2066 const struct super_operations
*ops
,
2067 const struct dentry_operations
*dops
,
2070 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
2071 #define fops_get(fops) \
2072 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
2073 #define fops_put(fops) \
2074 do { if (fops) module_put((fops)->owner); } while(0)
2076 extern int register_filesystem(struct file_system_type
*);
2077 extern int unregister_filesystem(struct file_system_type
*);
2078 extern struct vfsmount
*kern_mount_data(struct file_system_type
*, void *data
);
2079 #define kern_mount(type) kern_mount_data(type, NULL)
2080 extern void kern_unmount(struct vfsmount
*mnt
);
2081 extern int may_umount_tree(struct vfsmount
*);
2082 extern int may_umount(struct vfsmount
*);
2083 extern long do_mount(char *, char *, char *, unsigned long, void *);
2084 extern struct vfsmount
*collect_mounts(struct path
*);
2085 extern void drop_collected_mounts(struct vfsmount
*);
2086 extern int iterate_mounts(int (*)(struct vfsmount
*, void *), void *,
2088 extern int vfs_statfs(struct path
*, struct kstatfs
*);
2089 extern int user_statfs(const char __user
*, struct kstatfs
*);
2090 extern int fd_statfs(int, struct kstatfs
*);
2091 extern int vfs_ustat(dev_t
, struct kstatfs
*);
2092 extern int freeze_super(struct super_block
*super
);
2093 extern int thaw_super(struct super_block
*super
);
2094 extern bool our_mnt(struct vfsmount
*mnt
);
2096 extern int current_umask(void);
2099 extern struct kobject
*fs_kobj
;
2101 #define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
2102 extern int rw_verify_area(int, struct file
*, loff_t
*, size_t);
2104 #define FLOCK_VERIFY_READ 1
2105 #define FLOCK_VERIFY_WRITE 2
2107 #ifdef CONFIG_FILE_LOCKING
2108 extern int locks_mandatory_locked(struct inode
*);
2109 extern int locks_mandatory_area(int, struct inode
*, struct file
*, loff_t
, size_t);
2112 * Candidates for mandatory locking have the setgid bit set
2113 * but no group execute bit - an otherwise meaningless combination.
2116 static inline int __mandatory_lock(struct inode
*ino
)
2118 return (ino
->i_mode
& (S_ISGID
| S_IXGRP
)) == S_ISGID
;
2122 * ... and these candidates should be on MS_MANDLOCK mounted fs,
2123 * otherwise these will be advisory locks
2126 static inline int mandatory_lock(struct inode
*ino
)
2128 return IS_MANDLOCK(ino
) && __mandatory_lock(ino
);
2131 static inline int locks_verify_locked(struct inode
*inode
)
2133 if (mandatory_lock(inode
))
2134 return locks_mandatory_locked(inode
);
2138 static inline int locks_verify_truncate(struct inode
*inode
,
2142 if (inode
->i_flock
&& mandatory_lock(inode
))
2143 return locks_mandatory_area(
2144 FLOCK_VERIFY_WRITE
, inode
, filp
,
2145 size
< inode
->i_size
? size
: inode
->i_size
,
2146 (size
< inode
->i_size
? inode
->i_size
- size
2147 : size
- inode
->i_size
)
2152 static inline int break_lease(struct inode
*inode
, unsigned int mode
)
2155 return __break_lease(inode
, mode
);
2158 #else /* !CONFIG_FILE_LOCKING */
2159 static inline int locks_mandatory_locked(struct inode
*inode
)
2164 static inline int locks_mandatory_area(int rw
, struct inode
*inode
,
2165 struct file
*filp
, loff_t offset
,
2171 static inline int __mandatory_lock(struct inode
*inode
)
2176 static inline int mandatory_lock(struct inode
*inode
)
2181 static inline int locks_verify_locked(struct inode
*inode
)
2186 static inline int locks_verify_truncate(struct inode
*inode
, struct file
*filp
,
2192 static inline int break_lease(struct inode
*inode
, unsigned int mode
)
2197 #endif /* CONFIG_FILE_LOCKING */
2201 extern int do_truncate(struct dentry
*, loff_t start
, unsigned int time_attrs
,
2203 extern int do_fallocate(struct file
*file
, int mode
, loff_t offset
,
2205 extern long do_sys_open(int dfd
, const char __user
*filename
, int flags
,
2207 extern struct file
*filp_open(const char *, int, umode_t
);
2208 extern struct file
*file_open_root(struct dentry
*, struct vfsmount
*,
2210 extern struct file
* dentry_open(const struct path
*, int, const struct cred
*);
2211 extern int filp_close(struct file
*, fl_owner_t id
);
2212 extern char * getname(const char __user
*);
2217 extern int finish_open(struct file
*file
, struct dentry
*dentry
,
2218 int (*open
)(struct inode
*, struct file
*),
2220 extern int finish_no_open(struct file
*file
, struct dentry
*dentry
);
2224 extern int ioctl_preallocate(struct file
*filp
, void __user
*argp
);
2227 extern void __init
vfs_caches_init_early(void);
2228 extern void __init
vfs_caches_init(unsigned long);
2230 extern struct kmem_cache
*names_cachep
;
2232 #define __getname_gfp(gfp) kmem_cache_alloc(names_cachep, (gfp))
2233 #define __getname() __getname_gfp(GFP_KERNEL)
2234 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
2235 #ifndef CONFIG_AUDITSYSCALL
2236 #define putname(name) __putname(name)
2238 extern void putname(const char *name
);
2242 extern int register_blkdev(unsigned int, const char *);
2243 extern void unregister_blkdev(unsigned int, const char *);
2244 extern struct block_device
*bdget(dev_t
);
2245 extern struct block_device
*bdgrab(struct block_device
*bdev
);
2246 extern void bd_set_size(struct block_device
*, loff_t size
);
2247 extern sector_t
blkdev_max_block(struct block_device
*bdev
);
2248 extern void bd_forget(struct inode
*inode
);
2249 extern void bdput(struct block_device
*);
2250 extern void invalidate_bdev(struct block_device
*);
2251 extern void iterate_bdevs(void (*)(struct block_device
*, void *), void *);
2252 extern int sync_blockdev(struct block_device
*bdev
);
2253 extern void kill_bdev(struct block_device
*);
2254 extern struct super_block
*freeze_bdev(struct block_device
*);
2255 extern void emergency_thaw_all(void);
2256 extern int thaw_bdev(struct block_device
*bdev
, struct super_block
*sb
);
2257 extern int fsync_bdev(struct block_device
*);
2259 static inline void bd_forget(struct inode
*inode
) {}
2260 static inline int sync_blockdev(struct block_device
*bdev
) { return 0; }
2261 static inline void kill_bdev(struct block_device
*bdev
) {}
2262 static inline void invalidate_bdev(struct block_device
*bdev
) {}
2264 static inline struct super_block
*freeze_bdev(struct block_device
*sb
)
2269 static inline int thaw_bdev(struct block_device
*bdev
, struct super_block
*sb
)
2274 static inline void iterate_bdevs(void (*f
)(struct block_device
*, void *), void *arg
)
2278 extern int sync_filesystem(struct super_block
*);
2279 extern const struct file_operations def_blk_fops
;
2280 extern const struct file_operations def_chr_fops
;
2281 extern const struct file_operations bad_sock_fops
;
2282 extern const struct file_operations def_fifo_fops
;
2284 extern int ioctl_by_bdev(struct block_device
*, unsigned, unsigned long);
2285 extern int blkdev_ioctl(struct block_device
*, fmode_t
, unsigned, unsigned long);
2286 extern long compat_blkdev_ioctl(struct file
*, unsigned, unsigned long);
2287 extern int blkdev_get(struct block_device
*bdev
, fmode_t mode
, void *holder
);
2288 extern struct block_device
*blkdev_get_by_path(const char *path
, fmode_t mode
,
2290 extern struct block_device
*blkdev_get_by_dev(dev_t dev
, fmode_t mode
,
2292 extern int blkdev_put(struct block_device
*bdev
, fmode_t mode
);
2294 extern int bd_link_disk_holder(struct block_device
*bdev
, struct gendisk
*disk
);
2295 extern void bd_unlink_disk_holder(struct block_device
*bdev
,
2296 struct gendisk
*disk
);
2298 static inline int bd_link_disk_holder(struct block_device
*bdev
,
2299 struct gendisk
*disk
)
2303 static inline void bd_unlink_disk_holder(struct block_device
*bdev
,
2304 struct gendisk
*disk
)
2311 #define CHRDEV_MAJOR_HASH_SIZE 255
2312 extern int alloc_chrdev_region(dev_t
*, unsigned, unsigned, const char *);
2313 extern int register_chrdev_region(dev_t
, unsigned, const char *);
2314 extern int __register_chrdev(unsigned int major
, unsigned int baseminor
,
2315 unsigned int count
, const char *name
,
2316 const struct file_operations
*fops
);
2317 extern void __unregister_chrdev(unsigned int major
, unsigned int baseminor
,
2318 unsigned int count
, const char *name
);
2319 extern void unregister_chrdev_region(dev_t
, unsigned);
2320 extern void chrdev_show(struct seq_file
*,off_t
);
2322 static inline int register_chrdev(unsigned int major
, const char *name
,
2323 const struct file_operations
*fops
)
2325 return __register_chrdev(major
, 0, 256, name
, fops
);
2328 static inline void unregister_chrdev(unsigned int major
, const char *name
)
2330 __unregister_chrdev(major
, 0, 256, name
);
2333 /* fs/block_dev.c */
2334 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
2335 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
2338 #define BLKDEV_MAJOR_HASH_SIZE 255
2339 extern const char *__bdevname(dev_t
, char *buffer
);
2340 extern const char *bdevname(struct block_device
*bdev
, char *buffer
);
2341 extern struct block_device
*lookup_bdev(const char *);
2342 extern void blkdev_show(struct seq_file
*,off_t
);
2345 #define BLKDEV_MAJOR_HASH_SIZE 0
2348 extern void init_special_inode(struct inode
*, umode_t
, dev_t
);
2350 /* Invalid inode operations -- fs/bad_inode.c */
2351 extern void make_bad_inode(struct inode
*);
2352 extern int is_bad_inode(struct inode
*);
2354 extern const struct file_operations read_pipefifo_fops
;
2355 extern const struct file_operations write_pipefifo_fops
;
2356 extern const struct file_operations rdwr_pipefifo_fops
;
2360 * return READ, READA, or WRITE
2362 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2365 * return data direction, READ or WRITE
2367 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
2369 extern void check_disk_size_change(struct gendisk
*disk
,
2370 struct block_device
*bdev
);
2371 extern int revalidate_disk(struct gendisk
*);
2372 extern int check_disk_change(struct block_device
*);
2373 extern int __invalidate_device(struct block_device
*, bool);
2374 extern int invalidate_partition(struct gendisk
*, int);
2376 unsigned long invalidate_mapping_pages(struct address_space
*mapping
,
2377 pgoff_t start
, pgoff_t end
);
2379 static inline void invalidate_remote_inode(struct inode
*inode
)
2381 if (S_ISREG(inode
->i_mode
) || S_ISDIR(inode
->i_mode
) ||
2382 S_ISLNK(inode
->i_mode
))
2383 invalidate_mapping_pages(inode
->i_mapping
, 0, -1);
2385 extern int invalidate_inode_pages2(struct address_space
*mapping
);
2386 extern int invalidate_inode_pages2_range(struct address_space
*mapping
,
2387 pgoff_t start
, pgoff_t end
);
2388 extern int write_inode_now(struct inode
*, int);
2389 extern int filemap_fdatawrite(struct address_space
*);
2390 extern int filemap_flush(struct address_space
*);
2391 extern int filemap_fdatawait(struct address_space
*);
2392 extern int filemap_fdatawait_range(struct address_space
*, loff_t lstart
,
2394 extern int filemap_write_and_wait(struct address_space
*mapping
);
2395 extern int filemap_write_and_wait_range(struct address_space
*mapping
,
2396 loff_t lstart
, loff_t lend
);
2397 extern int __filemap_fdatawrite_range(struct address_space
*mapping
,
2398 loff_t start
, loff_t end
, int sync_mode
);
2399 extern int filemap_fdatawrite_range(struct address_space
*mapping
,
2400 loff_t start
, loff_t end
);
2402 extern int vfs_fsync_range(struct file
*file
, loff_t start
, loff_t end
,
2404 extern int vfs_fsync(struct file
*file
, int datasync
);
2405 extern int generic_write_sync(struct file
*file
, loff_t pos
, loff_t count
);
2406 extern void emergency_sync(void);
2407 extern void emergency_remount(void);
2409 extern sector_t
bmap(struct inode
*, sector_t
);
2411 extern int notify_change(struct dentry
*, struct iattr
*);
2412 extern int inode_permission(struct inode
*, int);
2413 extern int generic_permission(struct inode
*, int);
2415 static inline bool execute_ok(struct inode
*inode
)
2417 return (inode
->i_mode
& S_IXUGO
) || S_ISDIR(inode
->i_mode
);
2421 * get_write_access() gets write permission for a file.
2422 * put_write_access() releases this write permission.
2423 * This is used for regular files.
2424 * We cannot support write (and maybe mmap read-write shared) accesses and
2425 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2426 * can have the following values:
2427 * 0: no writers, no VM_DENYWRITE mappings
2428 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2429 * > 0: (i_writecount) users are writing to the file.
2431 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2432 * except for the cases where we don't hold i_writecount yet. Then we need to
2433 * use {get,deny}_write_access() - these functions check the sign and refuse
2434 * to do the change if sign is wrong.
2436 static inline int get_write_access(struct inode
*inode
)
2438 return atomic_inc_unless_negative(&inode
->i_writecount
) ? 0 : -ETXTBSY
;
2440 static inline int deny_write_access(struct file
*file
)
2442 struct inode
*inode
= file
->f_path
.dentry
->d_inode
;
2443 return atomic_dec_unless_positive(&inode
->i_writecount
) ? 0 : -ETXTBSY
;
2445 static inline void put_write_access(struct inode
* inode
)
2447 atomic_dec(&inode
->i_writecount
);
2449 static inline void allow_write_access(struct file
*file
)
2452 atomic_inc(&file
->f_path
.dentry
->d_inode
->i_writecount
);
2455 static inline void i_readcount_dec(struct inode
*inode
)
2457 BUG_ON(!atomic_read(&inode
->i_readcount
));
2458 atomic_dec(&inode
->i_readcount
);
2460 static inline void i_readcount_inc(struct inode
*inode
)
2462 atomic_inc(&inode
->i_readcount
);
2465 static inline void i_readcount_dec(struct inode
*inode
)
2469 static inline void i_readcount_inc(struct inode
*inode
)
2474 extern int do_pipe_flags(int *, int);
2476 extern int kernel_read(struct file
*, loff_t
, char *, unsigned long);
2477 extern struct file
* open_exec(const char *);
2479 /* fs/dcache.c -- generic fs support functions */
2480 extern int is_subdir(struct dentry
*, struct dentry
*);
2481 extern int path_is_under(struct path
*, struct path
*);
2482 extern ino_t
find_inode_number(struct dentry
*, struct qstr
*);
2484 #include <linux/err.h>
2486 /* needed for stackable file system support */
2487 extern loff_t
default_llseek(struct file
*file
, loff_t offset
, int origin
);
2489 extern loff_t
vfs_llseek(struct file
*file
, loff_t offset
, int origin
);
2491 extern int inode_init_always(struct super_block
*, struct inode
*);
2492 extern void inode_init_once(struct inode
*);
2493 extern void address_space_init_once(struct address_space
*mapping
);
2494 extern void ihold(struct inode
* inode
);
2495 extern void iput(struct inode
*);
2496 extern struct inode
* igrab(struct inode
*);
2497 extern ino_t
iunique(struct super_block
*, ino_t
);
2498 extern int inode_needs_sync(struct inode
*inode
);
2499 extern int generic_delete_inode(struct inode
*inode
);
2500 static inline int generic_drop_inode(struct inode
*inode
)
2502 return !inode
->i_nlink
|| inode_unhashed(inode
);
2505 extern struct inode
*ilookup5_nowait(struct super_block
*sb
,
2506 unsigned long hashval
, int (*test
)(struct inode
*, void *),
2508 extern struct inode
*ilookup5(struct super_block
*sb
, unsigned long hashval
,
2509 int (*test
)(struct inode
*, void *), void *data
);
2510 extern struct inode
*ilookup(struct super_block
*sb
, unsigned long ino
);
2512 extern struct inode
* iget5_locked(struct super_block
*, unsigned long, int (*test
)(struct inode
*, void *), int (*set
)(struct inode
*, void *), void *);
2513 extern struct inode
* iget_locked(struct super_block
*, unsigned long);
2514 extern int insert_inode_locked4(struct inode
*, unsigned long, int (*test
)(struct inode
*, void *), void *);
2515 extern int insert_inode_locked(struct inode
*);
2516 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2517 extern void lockdep_annotate_inode_mutex_key(struct inode
*inode
);
2519 static inline void lockdep_annotate_inode_mutex_key(struct inode
*inode
) { };
2521 extern void unlock_new_inode(struct inode
*);
2522 extern unsigned int get_next_ino(void);
2524 extern void __iget(struct inode
* inode
);
2525 extern void iget_failed(struct inode
*);
2526 extern void clear_inode(struct inode
*);
2527 extern void __destroy_inode(struct inode
*);
2528 extern struct inode
*new_inode_pseudo(struct super_block
*sb
);
2529 extern struct inode
*new_inode(struct super_block
*sb
);
2530 extern void free_inode_nonrcu(struct inode
*inode
);
2531 extern int should_remove_suid(struct dentry
*);
2532 extern int file_remove_suid(struct file
*);
2534 extern void __insert_inode_hash(struct inode
*, unsigned long hashval
);
2535 static inline void insert_inode_hash(struct inode
*inode
)
2537 __insert_inode_hash(inode
, inode
->i_ino
);
2540 extern void __remove_inode_hash(struct inode
*);
2541 static inline void remove_inode_hash(struct inode
*inode
)
2543 if (!inode_unhashed(inode
))
2544 __remove_inode_hash(inode
);
2547 extern void inode_sb_list_add(struct inode
*inode
);
2550 extern void submit_bio(int, struct bio
*);
2551 extern int bdev_read_only(struct block_device
*);
2553 extern int set_blocksize(struct block_device
*, int);
2554 extern int sb_set_blocksize(struct super_block
*, int);
2555 extern int sb_min_blocksize(struct super_block
*, int);
2557 extern int generic_file_mmap(struct file
*, struct vm_area_struct
*);
2558 extern int generic_file_readonly_mmap(struct file
*, struct vm_area_struct
*);
2559 extern int generic_file_remap_pages(struct vm_area_struct
*, unsigned long addr
,
2560 unsigned long size
, pgoff_t pgoff
);
2561 extern int file_read_actor(read_descriptor_t
* desc
, struct page
*page
, unsigned long offset
, unsigned long size
);
2562 int generic_write_checks(struct file
*file
, loff_t
*pos
, size_t *count
, int isblk
);
2563 extern ssize_t
generic_file_aio_read(struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
2564 extern ssize_t
__generic_file_aio_write(struct kiocb
*, const struct iovec
*, unsigned long,
2566 extern ssize_t
generic_file_aio_write(struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
2567 extern ssize_t
generic_file_direct_write(struct kiocb
*, const struct iovec
*,
2568 unsigned long *, loff_t
, loff_t
*, size_t, size_t);
2569 extern ssize_t
generic_file_buffered_write(struct kiocb
*, const struct iovec
*,
2570 unsigned long, loff_t
, loff_t
*, size_t, ssize_t
);
2571 extern ssize_t
do_sync_read(struct file
*filp
, char __user
*buf
, size_t len
, loff_t
*ppos
);
2572 extern ssize_t
do_sync_write(struct file
*filp
, const char __user
*buf
, size_t len
, loff_t
*ppos
);
2573 extern int generic_segment_checks(const struct iovec
*iov
,
2574 unsigned long *nr_segs
, size_t *count
, int access_flags
);
2576 /* fs/block_dev.c */
2577 extern ssize_t
blkdev_aio_read(struct kiocb
*iocb
, const struct iovec
*iov
,
2578 unsigned long nr_segs
, loff_t pos
);
2579 extern ssize_t
blkdev_aio_write(struct kiocb
*iocb
, const struct iovec
*iov
,
2580 unsigned long nr_segs
, loff_t pos
);
2581 extern int blkdev_fsync(struct file
*filp
, loff_t start
, loff_t end
,
2583 extern void block_sync_page(struct page
*page
);
2586 extern ssize_t
generic_file_splice_read(struct file
*, loff_t
*,
2587 struct pipe_inode_info
*, size_t, unsigned int);
2588 extern ssize_t
default_file_splice_read(struct file
*, loff_t
*,
2589 struct pipe_inode_info
*, size_t, unsigned int);
2590 extern ssize_t
generic_file_splice_write(struct pipe_inode_info
*,
2591 struct file
*, loff_t
*, size_t, unsigned int);
2592 extern ssize_t
generic_splice_sendpage(struct pipe_inode_info
*pipe
,
2593 struct file
*out
, loff_t
*, size_t len
, unsigned int flags
);
2594 extern long do_splice_direct(struct file
*in
, loff_t
*ppos
, struct file
*out
,
2595 size_t len
, unsigned int flags
);
2598 file_ra_state_init(struct file_ra_state
*ra
, struct address_space
*mapping
);
2599 extern loff_t
noop_llseek(struct file
*file
, loff_t offset
, int origin
);
2600 extern loff_t
no_llseek(struct file
*file
, loff_t offset
, int origin
);
2601 extern loff_t
generic_file_llseek(struct file
*file
, loff_t offset
, int origin
);
2602 extern loff_t
generic_file_llseek_size(struct file
*file
, loff_t offset
,
2603 int origin
, loff_t maxsize
, loff_t eof
);
2604 extern int generic_file_open(struct inode
* inode
, struct file
* filp
);
2605 extern int nonseekable_open(struct inode
* inode
, struct file
* filp
);
2607 #ifdef CONFIG_FS_XIP
2608 extern ssize_t
xip_file_read(struct file
*filp
, char __user
*buf
, size_t len
,
2610 extern int xip_file_mmap(struct file
* file
, struct vm_area_struct
* vma
);
2611 extern ssize_t
xip_file_write(struct file
*filp
, const char __user
*buf
,
2612 size_t len
, loff_t
*ppos
);
2613 extern int xip_truncate_page(struct address_space
*mapping
, loff_t from
);
2615 static inline int xip_truncate_page(struct address_space
*mapping
, loff_t from
)
2622 typedef void (dio_submit_t
)(int rw
, struct bio
*bio
, struct inode
*inode
,
2623 loff_t file_offset
);
2626 /* need locking between buffered and direct access */
2629 /* filesystem does not support filling holes */
2630 DIO_SKIP_HOLES
= 0x02,
2633 void dio_end_io(struct bio
*bio
, int error
);
2635 ssize_t
__blockdev_direct_IO(int rw
, struct kiocb
*iocb
, struct inode
*inode
,
2636 struct block_device
*bdev
, const struct iovec
*iov
, loff_t offset
,
2637 unsigned long nr_segs
, get_block_t get_block
, dio_iodone_t end_io
,
2638 dio_submit_t submit_io
, int flags
);
2640 static inline ssize_t
blockdev_direct_IO(int rw
, struct kiocb
*iocb
,
2641 struct inode
*inode
, const struct iovec
*iov
, loff_t offset
,
2642 unsigned long nr_segs
, get_block_t get_block
)
2644 return __blockdev_direct_IO(rw
, iocb
, inode
, inode
->i_sb
->s_bdev
, iov
,
2645 offset
, nr_segs
, get_block
, NULL
, NULL
,
2646 DIO_LOCKING
| DIO_SKIP_HOLES
);
2650 void inode_dio_wait(struct inode
*inode
);
2651 void inode_dio_done(struct inode
*inode
);
2653 extern const struct file_operations generic_ro_fops
;
2655 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2657 extern int vfs_readlink(struct dentry
*, char __user
*, int, const char *);
2658 extern int vfs_follow_link(struct nameidata
*, const char *);
2659 extern int page_readlink(struct dentry
*, char __user
*, int);
2660 extern void *page_follow_link_light(struct dentry
*, struct nameidata
*);
2661 extern void page_put_link(struct dentry
*, struct nameidata
*, void *);
2662 extern int __page_symlink(struct inode
*inode
, const char *symname
, int len
,
2664 extern int page_symlink(struct inode
*inode
, const char *symname
, int len
);
2665 extern const struct inode_operations page_symlink_inode_operations
;
2666 extern int generic_readlink(struct dentry
*, char __user
*, int);
2667 extern void generic_fillattr(struct inode
*, struct kstat
*);
2668 extern int vfs_getattr(struct vfsmount
*, struct dentry
*, struct kstat
*);
2669 void __inode_add_bytes(struct inode
*inode
, loff_t bytes
);
2670 void inode_add_bytes(struct inode
*inode
, loff_t bytes
);
2671 void inode_sub_bytes(struct inode
*inode
, loff_t bytes
);
2672 loff_t
inode_get_bytes(struct inode
*inode
);
2673 void inode_set_bytes(struct inode
*inode
, loff_t bytes
);
2675 extern int vfs_readdir(struct file
*, filldir_t
, void *);
2677 extern int vfs_stat(const char __user
*, struct kstat
*);
2678 extern int vfs_lstat(const char __user
*, struct kstat
*);
2679 extern int vfs_fstat(unsigned int, struct kstat
*);
2680 extern int vfs_fstatat(int , const char __user
*, struct kstat
*, int);
2682 extern int do_vfs_ioctl(struct file
*filp
, unsigned int fd
, unsigned int cmd
,
2684 extern int __generic_block_fiemap(struct inode
*inode
,
2685 struct fiemap_extent_info
*fieinfo
,
2686 loff_t start
, loff_t len
,
2687 get_block_t
*get_block
);
2688 extern int generic_block_fiemap(struct inode
*inode
,
2689 struct fiemap_extent_info
*fieinfo
, u64 start
,
2690 u64 len
, get_block_t
*get_block
);
2692 extern void get_filesystem(struct file_system_type
*fs
);
2693 extern void put_filesystem(struct file_system_type
*fs
);
2694 extern struct file_system_type
*get_fs_type(const char *name
);
2695 extern struct super_block
*get_super(struct block_device
*);
2696 extern struct super_block
*get_super_thawed(struct block_device
*);
2697 extern struct super_block
*get_active_super(struct block_device
*bdev
);
2698 extern void drop_super(struct super_block
*sb
);
2699 extern void iterate_supers(void (*)(struct super_block
*, void *), void *);
2700 extern void iterate_supers_type(struct file_system_type
*,
2701 void (*)(struct super_block
*, void *), void *);
2703 extern int dcache_dir_open(struct inode
*, struct file
*);
2704 extern int dcache_dir_close(struct inode
*, struct file
*);
2705 extern loff_t
dcache_dir_lseek(struct file
*, loff_t
, int);
2706 extern int dcache_readdir(struct file
*, void *, filldir_t
);
2707 extern int simple_setattr(struct dentry
*, struct iattr
*);
2708 extern int simple_getattr(struct vfsmount
*, struct dentry
*, struct kstat
*);
2709 extern int simple_statfs(struct dentry
*, struct kstatfs
*);
2710 extern int simple_open(struct inode
*inode
, struct file
*file
);
2711 extern int simple_link(struct dentry
*, struct inode
*, struct dentry
*);
2712 extern int simple_unlink(struct inode
*, struct dentry
*);
2713 extern int simple_rmdir(struct inode
*, struct dentry
*);
2714 extern int simple_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
2715 extern int noop_fsync(struct file
*, loff_t
, loff_t
, int);
2716 extern int simple_empty(struct dentry
*);
2717 extern int simple_readpage(struct file
*file
, struct page
*page
);
2718 extern int simple_write_begin(struct file
*file
, struct address_space
*mapping
,
2719 loff_t pos
, unsigned len
, unsigned flags
,
2720 struct page
**pagep
, void **fsdata
);
2721 extern int simple_write_end(struct file
*file
, struct address_space
*mapping
,
2722 loff_t pos
, unsigned len
, unsigned copied
,
2723 struct page
*page
, void *fsdata
);
2725 extern struct dentry
*simple_lookup(struct inode
*, struct dentry
*, unsigned int flags
);
2726 extern ssize_t
generic_read_dir(struct file
*, char __user
*, size_t, loff_t
*);
2727 extern const struct file_operations simple_dir_operations
;
2728 extern const struct inode_operations simple_dir_inode_operations
;
2729 struct tree_descr
{ char *name
; const struct file_operations
*ops
; int mode
; };
2730 struct dentry
*d_alloc_name(struct dentry
*, const char *);
2731 extern int simple_fill_super(struct super_block
*, unsigned long, struct tree_descr
*);
2732 extern int simple_pin_fs(struct file_system_type
*, struct vfsmount
**mount
, int *count
);
2733 extern void simple_release_fs(struct vfsmount
**mount
, int *count
);
2735 extern ssize_t
simple_read_from_buffer(void __user
*to
, size_t count
,
2736 loff_t
*ppos
, const void *from
, size_t available
);
2737 extern ssize_t
simple_write_to_buffer(void *to
, size_t available
, loff_t
*ppos
,
2738 const void __user
*from
, size_t count
);
2740 extern int generic_file_fsync(struct file
*, loff_t
, loff_t
, int);
2742 extern int generic_check_addressable(unsigned, u64
);
2744 #ifdef CONFIG_MIGRATION
2745 extern int buffer_migrate_page(struct address_space
*,
2746 struct page
*, struct page
*,
2749 #define buffer_migrate_page NULL
2752 extern int inode_change_ok(const struct inode
*, struct iattr
*);
2753 extern int inode_newsize_ok(const struct inode
*, loff_t offset
);
2754 extern void setattr_copy(struct inode
*inode
, const struct iattr
*attr
);
2756 extern int file_update_time(struct file
*file
);
2758 extern int generic_show_options(struct seq_file
*m
, struct dentry
*root
);
2759 extern void save_mount_options(struct super_block
*sb
, char *options
);
2760 extern void replace_mount_options(struct super_block
*sb
, char *options
);
2762 static inline ino_t
parent_ino(struct dentry
*dentry
)
2767 * Don't strictly need d_lock here? If the parent ino could change
2768 * then surely we'd have a deeper race in the caller?
2770 spin_lock(&dentry
->d_lock
);
2771 res
= dentry
->d_parent
->d_inode
->i_ino
;
2772 spin_unlock(&dentry
->d_lock
);
2776 /* Transaction based IO helpers */
2779 * An argresp is stored in an allocated page and holds the
2780 * size of the argument or response, along with its content
2782 struct simple_transaction_argresp
{
2787 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2789 char *simple_transaction_get(struct file
*file
, const char __user
*buf
,
2791 ssize_t
simple_transaction_read(struct file
*file
, char __user
*buf
,
2792 size_t size
, loff_t
*pos
);
2793 int simple_transaction_release(struct inode
*inode
, struct file
*file
);
2795 void simple_transaction_set(struct file
*file
, size_t n
);
2798 * simple attribute files
2800 * These attributes behave similar to those in sysfs:
2802 * Writing to an attribute immediately sets a value, an open file can be
2803 * written to multiple times.
2805 * Reading from an attribute creates a buffer from the value that might get
2806 * read with multiple read calls. When the attribute has been read
2807 * completely, no further read calls are possible until the file is opened
2810 * All attributes contain a text representation of a numeric value
2811 * that are accessed with the get() and set() functions.
2813 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2814 static int __fops ## _open(struct inode *inode, struct file *file) \
2816 __simple_attr_check_format(__fmt, 0ull); \
2817 return simple_attr_open(inode, file, __get, __set, __fmt); \
2819 static const struct file_operations __fops = { \
2820 .owner = THIS_MODULE, \
2821 .open = __fops ## _open, \
2822 .release = simple_attr_release, \
2823 .read = simple_attr_read, \
2824 .write = simple_attr_write, \
2825 .llseek = generic_file_llseek, \
2828 static inline __printf(1, 2)
2829 void __simple_attr_check_format(const char *fmt
, ...)
2831 /* don't do anything, just let the compiler check the arguments; */
2834 int simple_attr_open(struct inode
*inode
, struct file
*file
,
2835 int (*get
)(void *, u64
*), int (*set
)(void *, u64
),
2837 int simple_attr_release(struct inode
*inode
, struct file
*file
);
2838 ssize_t
simple_attr_read(struct file
*file
, char __user
*buf
,
2839 size_t len
, loff_t
*ppos
);
2840 ssize_t
simple_attr_write(struct file
*file
, const char __user
*buf
,
2841 size_t len
, loff_t
*ppos
);
2844 int proc_nr_files(struct ctl_table
*table
, int write
,
2845 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2846 int proc_nr_dentry(struct ctl_table
*table
, int write
,
2847 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2848 int proc_nr_inodes(struct ctl_table
*table
, int write
,
2849 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2850 int __init
get_filesystem_list(char *buf
);
2852 #define __FMODE_EXEC ((__force int) FMODE_EXEC)
2853 #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
2855 #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
2856 #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
2857 (flag & __FMODE_NONOTIFY)))
2859 static inline int is_sxid(umode_t mode
)
2861 return (mode
& S_ISUID
) || ((mode
& S_ISGID
) && (mode
& S_IXGRP
));
2864 static inline void inode_has_no_xattr(struct inode
*inode
)
2866 if (!is_sxid(inode
->i_mode
) && (inode
->i_sb
->s_flags
& MS_NOSEC
))
2867 inode
->i_flags
|= S_NOSEC
;
2870 #endif /* __KERNEL__ */
2871 #endif /* _LINUX_FS_H */