5 * This file has definitions for some important file table
9 #include <linux/limits.h>
10 #include <linux/ioctl.h>
11 #include <linux/blk_types.h>
12 #include <linux/types.h>
15 * It's silly to have NR_OPEN bigger than NR_FILE, but you can change
16 * the file limit at runtime and only root can increase the per-process
17 * nr_file rlimit, so it's safe to set up a ridiculously high absolute
18 * upper limit on files-per-process.
20 * Some programs (notably those using select()) may have to be
21 * recompiled to take full advantage of the new limits..
24 /* Fixed constants first: */
26 #define INR_OPEN_CUR 1024 /* Initial setting for nfile rlimits */
27 #define INR_OPEN_MAX 4096 /* Hard limit for nfile rlimits */
29 #define BLOCK_SIZE_BITS 10
30 #define BLOCK_SIZE (1<<BLOCK_SIZE_BITS)
32 #define SEEK_SET 0 /* seek relative to beginning of file */
33 #define SEEK_CUR 1 /* seek relative to current file position */
34 #define SEEK_END 2 /* seek relative to end of file */
35 #define SEEK_DATA 3 /* seek to the next data */
36 #define SEEK_HOLE 4 /* seek to the next hole */
37 #define SEEK_MAX SEEK_HOLE
45 /* And dynamically-tunable limits and defaults: */
46 struct files_stat_struct {
47 unsigned long nr_files; /* read only */
48 unsigned long nr_free_files; /* read only */
49 unsigned long max_files; /* tunable */
52 struct inodes_stat_t {
55 int dummy[5]; /* padding for sysctl ABI compatibility */
59 #define NR_FILE 8192 /* this can well be larger on a larger system */
61 #define MAY_EXEC 0x00000001
62 #define MAY_WRITE 0x00000002
63 #define MAY_READ 0x00000004
64 #define MAY_APPEND 0x00000008
65 #define MAY_ACCESS 0x00000010
66 #define MAY_OPEN 0x00000020
67 #define MAY_CHDIR 0x00000040
68 /* called from RCU mode, don't block */
69 #define MAY_NOT_BLOCK 0x00000080
72 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
73 * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
76 /* file is open for reading */
77 #define FMODE_READ ((__force fmode_t)0x1)
78 /* file is open for writing */
79 #define FMODE_WRITE ((__force fmode_t)0x2)
80 /* file is seekable */
81 #define FMODE_LSEEK ((__force fmode_t)0x4)
82 /* file can be accessed using pread */
83 #define FMODE_PREAD ((__force fmode_t)0x8)
84 /* file can be accessed using pwrite */
85 #define FMODE_PWRITE ((__force fmode_t)0x10)
86 /* File is opened for execution with sys_execve / sys_uselib */
87 #define FMODE_EXEC ((__force fmode_t)0x20)
88 /* File is opened with O_NDELAY (only set for block devices) */
89 #define FMODE_NDELAY ((__force fmode_t)0x40)
90 /* File is opened with O_EXCL (only set for block devices) */
91 #define FMODE_EXCL ((__force fmode_t)0x80)
92 /* File is opened using open(.., 3, ..) and is writeable only for ioctls
93 (specialy hack for floppy.c) */
94 #define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
95 /* 32bit hashes as llseek() offset (for directories) */
96 #define FMODE_32BITHASH ((__force fmode_t)0x200)
97 /* 64bit hashes as llseek() offset (for directories) */
98 #define FMODE_64BITHASH ((__force fmode_t)0x400)
101 * Don't update ctime and mtime.
103 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
104 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
106 #define FMODE_NOCMTIME ((__force fmode_t)0x800)
108 /* Expect random access pattern */
109 #define FMODE_RANDOM ((__force fmode_t)0x1000)
111 /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
112 #define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
114 /* File is opened with O_PATH; almost nothing can be done with it */
115 #define FMODE_PATH ((__force fmode_t)0x4000)
117 /* File was opened by fanotify and shouldn't generate fanotify events */
118 #define FMODE_NONOTIFY ((__force fmode_t)0x1000000)
121 * The below are the various read and write types that we support. Some of
122 * them include behavioral modifiers that send information down to the
123 * block layer and IO scheduler. Terminology:
125 * The block layer uses device plugging to defer IO a little bit, in
126 * the hope that we will see more IO very shortly. This increases
127 * coalescing of adjacent IO and thus reduces the number of IOs we
128 * have to send to the device. It also allows for better queuing,
129 * if the IO isn't mergeable. If the caller is going to be waiting
130 * for the IO, then he must ensure that the device is unplugged so
131 * that the IO is dispatched to the driver.
133 * All IO is handled async in Linux. This is fine for background
134 * writes, but for reads or writes that someone waits for completion
135 * on, we want to notify the block layer and IO scheduler so that they
136 * know about it. That allows them to make better scheduling
137 * decisions. So when the below references 'sync' and 'async', it
138 * is referencing this priority hint.
140 * With that in mind, the available types are:
142 * READ A normal read operation. Device will be plugged.
143 * READ_SYNC A synchronous read. Device is not plugged, caller can
144 * immediately wait on this read without caring about
146 * READA Used for read-ahead operations. Lower priority, and the
147 * block layer could (in theory) choose to ignore this
148 * request if it runs into resource problems.
149 * WRITE A normal async write. Device will be plugged.
150 * WRITE_SYNC Synchronous write. Identical to WRITE, but passes down
151 * the hint that someone will be waiting on this IO
152 * shortly. The write equivalent of READ_SYNC.
153 * WRITE_ODIRECT Special case write for O_DIRECT only.
154 * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
155 * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
156 * non-volatile media on completion.
157 * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
158 * by a cache flush and data is guaranteed to be on
159 * non-volatile media on completion.
162 #define RW_MASK REQ_WRITE
163 #define RWA_MASK REQ_RAHEAD
166 #define WRITE RW_MASK
167 #define READA RWA_MASK
169 #define READ_SYNC (READ | REQ_SYNC)
170 #define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE)
171 #define WRITE_ODIRECT (WRITE | REQ_SYNC)
172 #define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
173 #define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
174 #define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
178 * Flag for rw_copy_check_uvector and compat_rw_copy_check_uvector
179 * that indicates that they should check the contents of the iovec are
180 * valid, but not check the memory that the iovec elements
183 #define CHECK_IOVEC_ONLY -1
189 /* public flags for file_system_type */
190 #define FS_REQUIRES_DEV 1
191 #define FS_BINARY_MOUNTDATA 2
192 #define FS_HAS_SUBTYPE 4
193 #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
194 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move()
195 * during rename() internally.
199 * These are the fs-independent mount-flags: up to 32 flags are supported
201 #define MS_RDONLY 1 /* Mount read-only */
202 #define MS_NOSUID 2 /* Ignore suid and sgid bits */
203 #define MS_NODEV 4 /* Disallow access to device special files */
204 #define MS_NOEXEC 8 /* Disallow program execution */
205 #define MS_SYNCHRONOUS 16 /* Writes are synced at once */
206 #define MS_REMOUNT 32 /* Alter flags of a mounted FS */
207 #define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
208 #define MS_DIRSYNC 128 /* Directory modifications are synchronous */
209 #define MS_NOATIME 1024 /* Do not update access times. */
210 #define MS_NODIRATIME 2048 /* Do not update directory access times */
214 #define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
215 MS_VERBOSE is deprecated. */
216 #define MS_SILENT 32768
217 #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
218 #define MS_UNBINDABLE (1<<17) /* change to unbindable */
219 #define MS_PRIVATE (1<<18) /* change to private */
220 #define MS_SLAVE (1<<19) /* change to slave */
221 #define MS_SHARED (1<<20) /* change to shared */
222 #define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */
223 #define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
224 #define MS_I_VERSION (1<<23) /* Update inode I_version field */
225 #define MS_STRICTATIME (1<<24) /* Always perform atime updates */
226 #define MS_NOSEC (1<<28)
227 #define MS_BORN (1<<29)
228 #define MS_ACTIVE (1<<30)
229 #define MS_NOUSER (1<<31)
232 * Superblock flags that can be altered by MS_REMOUNT
234 #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK|MS_I_VERSION)
237 * Old magic mount flag and mask
239 #define MS_MGC_VAL 0xC0ED0000
240 #define MS_MGC_MSK 0xffff0000
242 /* Inode flags - they have nothing to superblock flags now */
244 #define S_SYNC 1 /* Writes are synced at once */
245 #define S_NOATIME 2 /* Do not update access times */
246 #define S_APPEND 4 /* Append-only file */
247 #define S_IMMUTABLE 8 /* Immutable file */
248 #define S_DEAD 16 /* removed, but still open directory */
249 #define S_NOQUOTA 32 /* Inode is not counted to quota */
250 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
251 #define S_NOCMTIME 128 /* Do not update file c/mtime */
252 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
253 #define S_PRIVATE 512 /* Inode is fs-internal */
254 #define S_IMA 1024 /* Inode has an associated IMA struct */
255 #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
256 #define S_NOSEC 4096 /* no suid or xattr security attributes */
259 * Note that nosuid etc flags are inode-specific: setting some file-system
260 * flags just means all the inodes inherit those flags by default. It might be
261 * possible to override it selectively if you really wanted to with some
262 * ioctl() that is not currently implemented.
264 * Exception: MS_RDONLY is always applied to the entire file system.
266 * Unfortunately, it is possible to change a filesystems flags with it mounted
267 * with files in use. This means that all of the inodes will not have their
268 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
269 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
271 #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
273 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
274 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
275 ((inode)->i_flags & S_SYNC))
276 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
277 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
278 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
279 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
280 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
282 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
283 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
284 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
285 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
287 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
288 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
289 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
290 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
291 #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
292 #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
293 #define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
295 /* the read-only stuff doesn't really belong here, but any other place is
296 probably as bad and I don't want to create yet another include file. */
298 #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
299 #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
300 #define BLKRRPART _IO(0x12,95) /* re-read partition table */
301 #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
302 #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
303 #define BLKRASET _IO(0x12,98) /* set read ahead for block device */
304 #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
305 #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
306 #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
307 #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
308 #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
309 #define BLKSSZGET _IO(0x12,104)/* get block device sector size */
311 #define BLKPG _IO(0x12,105)/* See blkpg.h */
313 /* Some people are morons. Do not use sizeof! */
315 #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
316 #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
317 /* This was here just to show that the number is taken -
318 probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
320 /* A jump here: 108-111 have been used for various private purposes. */
321 #define BLKBSZGET _IOR(0x12,112,size_t)
322 #define BLKBSZSET _IOW(0x12,113,size_t)
323 #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
324 #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
325 #define BLKTRACESTART _IO(0x12,116)
326 #define BLKTRACESTOP _IO(0x12,117)
327 #define BLKTRACETEARDOWN _IO(0x12,118)
328 #define BLKDISCARD _IO(0x12,119)
329 #define BLKIOMIN _IO(0x12,120)
330 #define BLKIOOPT _IO(0x12,121)
331 #define BLKALIGNOFF _IO(0x12,122)
332 #define BLKPBSZGET _IO(0x12,123)
333 #define BLKDISCARDZEROES _IO(0x12,124)
334 #define BLKSECDISCARD _IO(0x12,125)
335 #define BLKROTATIONAL _IO(0x12,126)
337 #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
338 #define FIBMAP _IO(0x00,1) /* bmap access */
339 #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
340 #define FIFREEZE _IOWR('X', 119, int) /* Freeze */
341 #define FITHAW _IOWR('X', 120, int) /* Thaw */
342 #define FITRIM _IOWR('X', 121, struct fstrim_range) /* Trim */
344 #define FS_IOC_GETFLAGS _IOR('f', 1, long)
345 #define FS_IOC_SETFLAGS _IOW('f', 2, long)
346 #define FS_IOC_GETVERSION _IOR('v', 1, long)
347 #define FS_IOC_SETVERSION _IOW('v', 2, long)
348 #define FS_IOC_FIEMAP _IOWR('f', 11, struct fiemap)
349 #define FS_IOC32_GETFLAGS _IOR('f', 1, int)
350 #define FS_IOC32_SETFLAGS _IOW('f', 2, int)
351 #define FS_IOC32_GETVERSION _IOR('v', 1, int)
352 #define FS_IOC32_SETVERSION _IOW('v', 2, int)
355 * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS)
357 #define FS_SECRM_FL 0x00000001 /* Secure deletion */
358 #define FS_UNRM_FL 0x00000002 /* Undelete */
359 #define FS_COMPR_FL 0x00000004 /* Compress file */
360 #define FS_SYNC_FL 0x00000008 /* Synchronous updates */
361 #define FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
362 #define FS_APPEND_FL 0x00000020 /* writes to file may only append */
363 #define FS_NODUMP_FL 0x00000040 /* do not dump file */
364 #define FS_NOATIME_FL 0x00000080 /* do not update atime */
365 /* Reserved for compression usage... */
366 #define FS_DIRTY_FL 0x00000100
367 #define FS_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
368 #define FS_NOCOMP_FL 0x00000400 /* Don't compress */
369 #define FS_ECOMPR_FL 0x00000800 /* Compression error */
370 /* End compression flags --- maybe not all used */
371 #define FS_BTREE_FL 0x00001000 /* btree format dir */
372 #define FS_INDEX_FL 0x00001000 /* hash-indexed directory */
373 #define FS_IMAGIC_FL 0x00002000 /* AFS directory */
374 #define FS_JOURNAL_DATA_FL 0x00004000 /* Reserved for ext3 */
375 #define FS_NOTAIL_FL 0x00008000 /* file tail should not be merged */
376 #define FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
377 #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
378 #define FS_EXTENT_FL 0x00080000 /* Extents */
379 #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */
380 #define FS_NOCOW_FL 0x00800000 /* Do not cow file */
381 #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */
383 #define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
384 #define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
387 #define SYNC_FILE_RANGE_WAIT_BEFORE 1
388 #define SYNC_FILE_RANGE_WRITE 2
389 #define SYNC_FILE_RANGE_WAIT_AFTER 4
393 #include <linux/linkage.h>
394 #include <linux/wait.h>
395 #include <linux/kdev_t.h>
396 #include <linux/dcache.h>
397 #include <linux/path.h>
398 #include <linux/stat.h>
399 #include <linux/cache.h>
400 #include <linux/list.h>
401 #include <linux/radix-tree.h>
402 #include <linux/prio_tree.h>
403 #include <linux/init.h>
404 #include <linux/pid.h>
405 #include <linux/bug.h>
406 #include <linux/mutex.h>
407 #include <linux/capability.h>
408 #include <linux/semaphore.h>
409 #include <linux/fiemap.h>
410 #include <linux/rculist_bl.h>
411 #include <linux/atomic.h>
412 #include <linux/shrinker.h>
413 #include <linux/migrate_mode.h>
414 #include <linux/uidgid.h>
416 #include <asm/byteorder.h>
418 struct export_operations;
424 struct pipe_inode_info;
425 struct poll_table_struct;
427 struct vm_area_struct;
431 extern void __init inode_init(void);
432 extern void __init inode_init_early(void);
433 extern void __init files_init(unsigned long);
435 extern struct files_stat_struct files_stat;
436 extern unsigned long get_max_files(void);
437 extern int sysctl_nr_open;
438 extern struct inodes_stat_t inodes_stat;
439 extern int leases_enable, lease_break_time;
442 typedef int (get_block_t)(struct inode *inode, sector_t iblock,
443 struct buffer_head *bh_result, int create);
444 typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
445 ssize_t bytes, void *private, int ret,
449 * Attribute flags. These should be or-ed together to figure out what
452 #define ATTR_MODE (1 << 0)
453 #define ATTR_UID (1 << 1)
454 #define ATTR_GID (1 << 2)
455 #define ATTR_SIZE (1 << 3)
456 #define ATTR_ATIME (1 << 4)
457 #define ATTR_MTIME (1 << 5)
458 #define ATTR_CTIME (1 << 6)
459 #define ATTR_ATIME_SET (1 << 7)
460 #define ATTR_MTIME_SET (1 << 8)
461 #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
462 #define ATTR_ATTR_FLAG (1 << 10)
463 #define ATTR_KILL_SUID (1 << 11)
464 #define ATTR_KILL_SGID (1 << 12)
465 #define ATTR_FILE (1 << 13)
466 #define ATTR_KILL_PRIV (1 << 14)
467 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
468 #define ATTR_TIMES_SET (1 << 16)
471 * This is the Inode Attributes structure, used for notify_change(). It
472 * uses the above definitions as flags, to know which values have changed.
473 * Also, in this manner, a Filesystem can look at only the values it cares
474 * about. Basically, these are the attributes that the VFS layer can
475 * request to change from the FS layer.
477 * Derek Atkins <warlord@MIT.EDU> 94-10-20
480 unsigned int ia_valid;
485 struct timespec ia_atime;
486 struct timespec ia_mtime;
487 struct timespec ia_ctime;
490 * Not an attribute, but an auxiliary info for filesystems wanting to
491 * implement an ftruncate() like method. NOTE: filesystem should
492 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
494 struct file *ia_file;
498 * Includes for diskquotas.
500 #include <linux/quota.h>
503 * enum positive_aop_returns - aop return codes with specific semantics
505 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
506 * completed, that the page is still locked, and
507 * should be considered active. The VM uses this hint
508 * to return the page to the active list -- it won't
509 * be a candidate for writeback again in the near
510 * future. Other callers must be careful to unlock
511 * the page if they get this return. Returned by
514 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
515 * unlocked it and the page might have been truncated.
516 * The caller should back up to acquiring a new page and
517 * trying again. The aop will be taking reasonable
518 * precautions not to livelock. If the caller held a page
519 * reference, it should drop it before retrying. Returned
522 * address_space_operation functions return these large constants to indicate
523 * special semantics to the caller. These are much larger than the bytes in a
524 * page to allow for functions that return the number of bytes operated on in a
528 enum positive_aop_returns {
529 AOP_WRITEPAGE_ACTIVATE = 0x80000,
530 AOP_TRUNCATED_PAGE = 0x80001,
533 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
534 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
535 #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
536 * helper code (eg buffer layer)
537 * to clear GFP_FS from alloc */
540 * oh the beauties of C type declarations.
543 struct address_space;
544 struct writeback_control;
547 const struct iovec *iov;
548 unsigned long nr_segs;
553 size_t iov_iter_copy_from_user_atomic(struct page *page,
554 struct iov_iter *i, unsigned long offset, size_t bytes);
555 size_t iov_iter_copy_from_user(struct page *page,
556 struct iov_iter *i, unsigned long offset, size_t bytes);
557 void iov_iter_advance(struct iov_iter *i, size_t bytes);
558 int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes);
559 size_t iov_iter_single_seg_count(struct iov_iter *i);
561 static inline void iov_iter_init(struct iov_iter *i,
562 const struct iovec *iov, unsigned long nr_segs,
563 size_t count, size_t written)
566 i->nr_segs = nr_segs;
568 i->count = count + written;
570 iov_iter_advance(i, written);
573 static inline size_t iov_iter_count(struct iov_iter *i)
579 * "descriptor" for what we're up to with a read.
580 * This allows us to use the same read code yet
581 * have multiple different users of the data that
582 * we read from a file.
584 * The simplest case just copies the data to user
597 typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
598 unsigned long, unsigned long);
600 struct address_space_operations {
601 int (*writepage)(struct page *page, struct writeback_control *wbc);
602 int (*readpage)(struct file *, struct page *);
604 /* Write back some dirty pages from this mapping. */
605 int (*writepages)(struct address_space *, struct writeback_control *);
607 /* Set a page dirty. Return true if this dirtied it */
608 int (*set_page_dirty)(struct page *page);
610 int (*readpages)(struct file *filp, struct address_space *mapping,
611 struct list_head *pages, unsigned nr_pages);
613 int (*write_begin)(struct file *, struct address_space *mapping,
614 loff_t pos, unsigned len, unsigned flags,
615 struct page **pagep, void **fsdata);
616 int (*write_end)(struct file *, struct address_space *mapping,
617 loff_t pos, unsigned len, unsigned copied,
618 struct page *page, void *fsdata);
620 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
621 sector_t (*bmap)(struct address_space *, sector_t);
622 void (*invalidatepage) (struct page *, unsigned long);
623 int (*releasepage) (struct page *, gfp_t);
624 void (*freepage)(struct page *);
625 ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
626 loff_t offset, unsigned long nr_segs);
627 int (*get_xip_mem)(struct address_space *, pgoff_t, int,
628 void **, unsigned long *);
630 * migrate the contents of a page to the specified target. If sync
631 * is false, it must not block.
633 int (*migratepage) (struct address_space *,
634 struct page *, struct page *, enum migrate_mode);
635 int (*launder_page) (struct page *);
636 int (*is_partially_uptodate) (struct page *, read_descriptor_t *,
638 int (*error_remove_page)(struct address_space *, struct page *);
641 extern const struct address_space_operations empty_aops;
644 * pagecache_write_begin/pagecache_write_end must be used by general code
645 * to write into the pagecache.
647 int pagecache_write_begin(struct file *, struct address_space *mapping,
648 loff_t pos, unsigned len, unsigned flags,
649 struct page **pagep, void **fsdata);
651 int pagecache_write_end(struct file *, struct address_space *mapping,
652 loff_t pos, unsigned len, unsigned copied,
653 struct page *page, void *fsdata);
655 struct backing_dev_info;
656 struct address_space {
657 struct inode *host; /* owner: inode, block_device */
658 struct radix_tree_root page_tree; /* radix tree of all pages */
659 spinlock_t tree_lock; /* and lock protecting it */
660 unsigned int i_mmap_writable;/* count VM_SHARED mappings */
661 struct prio_tree_root i_mmap; /* tree of private and shared mappings */
662 struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
663 struct mutex i_mmap_mutex; /* protect tree, count, list */
664 /* Protected by tree_lock together with the radix tree */
665 unsigned long nrpages; /* number of total pages */
666 pgoff_t writeback_index;/* writeback starts here */
667 const struct address_space_operations *a_ops; /* methods */
668 unsigned long flags; /* error bits/gfp mask */
669 struct backing_dev_info *backing_dev_info; /* device readahead, etc */
670 spinlock_t private_lock; /* for use by the address_space */
671 struct list_head private_list; /* ditto */
672 struct address_space *assoc_mapping; /* ditto */
673 } __attribute__((aligned(sizeof(long))));
675 * On most architectures that alignment is already the case; but
676 * must be enforced here for CRIS, to let the least significant bit
677 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
679 struct request_queue;
681 struct block_device {
682 dev_t bd_dev; /* not a kdev_t - it's a search key */
684 struct inode * bd_inode; /* will die */
685 struct super_block * bd_super;
686 struct mutex bd_mutex; /* open/close mutex */
687 struct list_head bd_inodes;
691 bool bd_write_holder;
693 struct list_head bd_holder_disks;
695 struct block_device * bd_contains;
696 unsigned bd_block_size;
697 struct hd_struct * bd_part;
698 /* number of times partitions within this device have been opened. */
699 unsigned bd_part_count;
701 struct gendisk * bd_disk;
702 struct request_queue * bd_queue;
703 struct list_head bd_list;
705 * Private data. You must have bd_claim'ed the block_device
706 * to use this. NOTE: bd_claim allows an owner to claim
707 * the same device multiple times, the owner must take special
708 * care to not mess up bd_private for that case.
710 unsigned long bd_private;
712 /* The counter of freeze processes */
713 int bd_fsfreeze_count;
714 /* Mutex for freeze */
715 struct mutex bd_fsfreeze_mutex;
719 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
722 #define PAGECACHE_TAG_DIRTY 0
723 #define PAGECACHE_TAG_WRITEBACK 1
724 #define PAGECACHE_TAG_TOWRITE 2
726 int mapping_tagged(struct address_space *mapping, int tag);
729 * Might pages of this file be mapped into userspace?
731 static inline int mapping_mapped(struct address_space *mapping)
733 return !prio_tree_empty(&mapping->i_mmap) ||
734 !list_empty(&mapping->i_mmap_nonlinear);
738 * Might pages of this file have been modified in userspace?
739 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
740 * marks vma as VM_SHARED if it is shared, and the file was opened for
741 * writing i.e. vma may be mprotected writable even if now readonly.
743 static inline int mapping_writably_mapped(struct address_space *mapping)
745 return mapping->i_mmap_writable != 0;
749 * Use sequence counter to get consistent i_size on 32-bit processors.
751 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
752 #include <linux/seqlock.h>
753 #define __NEED_I_SIZE_ORDERED
754 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
756 #define i_size_ordered_init(inode) do { } while (0)
760 #define ACL_NOT_CACHED ((void *)(-1))
762 #define IOP_FASTPERM 0x0001
763 #define IOP_LOOKUP 0x0002
764 #define IOP_NOFOLLOW 0x0004
767 * Keep mostly read-only and often accessed (especially for
768 * the RCU path lookup and 'stat' data) fields at the beginning
769 * of the 'struct inode'
773 unsigned short i_opflags;
776 unsigned int i_flags;
778 #ifdef CONFIG_FS_POSIX_ACL
779 struct posix_acl *i_acl;
780 struct posix_acl *i_default_acl;
783 const struct inode_operations *i_op;
784 struct super_block *i_sb;
785 struct address_space *i_mapping;
787 #ifdef CONFIG_SECURITY
791 /* Stat data, not accessed from path walking */
794 * Filesystems may only read i_nlink directly. They shall use the
795 * following functions for modification:
797 * (set|clear|inc|drop)_nlink
798 * inode_(inc|dec)_link_count
801 const unsigned int i_nlink;
802 unsigned int __i_nlink;
806 struct timespec i_atime;
807 struct timespec i_mtime;
808 struct timespec i_ctime;
809 spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
810 unsigned short i_bytes;
811 unsigned int i_blkbits;
814 #ifdef __NEED_I_SIZE_ORDERED
815 seqcount_t i_size_seqcount;
819 unsigned long i_state;
820 struct mutex i_mutex;
822 unsigned long dirtied_when; /* jiffies of first dirtying */
824 struct hlist_node i_hash;
825 struct list_head i_wb_list; /* backing dev IO list */
826 struct list_head i_lru; /* inode LRU list */
827 struct list_head i_sb_list;
829 struct list_head i_dentry;
830 struct rcu_head i_rcu;
834 atomic_t i_dio_count;
835 atomic_t i_writecount;
836 const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
837 struct file_lock *i_flock;
838 struct address_space i_data;
840 struct dquot *i_dquot[MAXQUOTAS];
842 struct list_head i_devices;
844 struct pipe_inode_info *i_pipe;
845 struct block_device *i_bdev;
851 #ifdef CONFIG_FSNOTIFY
852 __u32 i_fsnotify_mask; /* all events this inode cares about */
853 struct hlist_head i_fsnotify_marks;
857 atomic_t i_readcount; /* struct files open RO */
859 void *i_private; /* fs or device private pointer */
862 static inline int inode_unhashed(struct inode *inode)
864 return hlist_unhashed(&inode->i_hash);
868 * inode->i_mutex nesting subclasses for the lock validator:
870 * 0: the object of the current VFS operation
875 * The locking order between these classes is
876 * parent -> child -> normal -> xattr -> quota
878 enum inode_i_mutex_lock_class
888 * NOTE: in a 32bit arch with a preemptable kernel and
889 * an UP compile the i_size_read/write must be atomic
890 * with respect to the local cpu (unlike with preempt disabled),
891 * but they don't need to be atomic with respect to other cpus like in
892 * true SMP (so they need either to either locally disable irq around
893 * the read or for example on x86 they can be still implemented as a
894 * cmpxchg8b without the need of the lock prefix). For SMP compiles
895 * and 64bit archs it makes no difference if preempt is enabled or not.
897 static inline loff_t i_size_read(const struct inode *inode)
899 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
904 seq = read_seqcount_begin(&inode->i_size_seqcount);
905 i_size = inode->i_size;
906 } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
908 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
912 i_size = inode->i_size;
916 return inode->i_size;
921 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
922 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
923 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
925 static inline void i_size_write(struct inode *inode, loff_t i_size)
927 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
928 write_seqcount_begin(&inode->i_size_seqcount);
929 inode->i_size = i_size;
930 write_seqcount_end(&inode->i_size_seqcount);
931 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
933 inode->i_size = i_size;
936 inode->i_size = i_size;
940 /* Helper functions so that in most cases filesystems will
941 * not need to deal directly with kuid_t and kgid_t and can
942 * instead deal with the raw numeric values that are stored
945 static inline uid_t i_uid_read(const struct inode *inode)
947 return from_kuid(&init_user_ns, inode->i_uid);
950 static inline gid_t i_gid_read(const struct inode *inode)
952 return from_kgid(&init_user_ns, inode->i_gid);
955 static inline void i_uid_write(struct inode *inode, uid_t uid)
957 inode->i_uid = make_kuid(&init_user_ns, uid);
960 static inline void i_gid_write(struct inode *inode, gid_t gid)
962 inode->i_gid = make_kgid(&init_user_ns, gid);
965 static inline unsigned iminor(const struct inode *inode)
967 return MINOR(inode->i_rdev);
970 static inline unsigned imajor(const struct inode *inode)
972 return MAJOR(inode->i_rdev);
975 extern struct block_device *I_BDEV(struct inode *inode);
978 rwlock_t lock; /* protects pid, uid, euid fields */
979 struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
980 enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
981 kuid_t uid, euid; /* uid/euid of process setting the owner */
982 int signum; /* posix.1b rt signal to be delivered on IO */
986 * Track a single file's readahead state
988 struct file_ra_state {
989 pgoff_t start; /* where readahead started */
990 unsigned int size; /* # of readahead pages */
991 unsigned int async_size; /* do asynchronous readahead when
992 there are only # of pages ahead */
994 unsigned int ra_pages; /* Maximum readahead window */
995 unsigned int mmap_miss; /* Cache miss stat for mmap accesses */
996 loff_t prev_pos; /* Cache last read() position */
1000 * Check if @index falls in the readahead windows.
1002 static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
1004 return (index >= ra->start &&
1005 index < ra->start + ra->size);
1008 #define FILE_MNT_WRITE_TAKEN 1
1009 #define FILE_MNT_WRITE_RELEASED 2
1013 * fu_list becomes invalid after file_free is called and queued via
1014 * fu_rcuhead for RCU freeing
1017 struct list_head fu_list;
1018 struct rcu_head fu_rcuhead;
1021 #define f_dentry f_path.dentry
1022 #define f_vfsmnt f_path.mnt
1023 const struct file_operations *f_op;
1026 * Protects f_ep_links, f_flags, f_pos vs i_size in lseek SEEK_CUR.
1027 * Must not be taken from IRQ context.
1033 atomic_long_t f_count;
1034 unsigned int f_flags;
1037 struct fown_struct f_owner;
1038 const struct cred *f_cred;
1039 struct file_ra_state f_ra;
1042 #ifdef CONFIG_SECURITY
1045 /* needed for tty driver, and maybe others */
1049 /* Used by fs/eventpoll.c to link all the hooks to this file */
1050 struct list_head f_ep_links;
1051 struct list_head f_tfile_llink;
1052 #endif /* #ifdef CONFIG_EPOLL */
1053 struct address_space *f_mapping;
1054 #ifdef CONFIG_DEBUG_WRITECOUNT
1055 unsigned long f_mnt_write_state;
1059 struct file_handle {
1062 /* file identifier */
1063 unsigned char f_handle[0];
1066 #define get_file(x) atomic_long_inc(&(x)->f_count)
1067 #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
1068 #define file_count(x) atomic_long_read(&(x)->f_count)
1070 #ifdef CONFIG_DEBUG_WRITECOUNT
1071 static inline void file_take_write(struct file *f)
1073 WARN_ON(f->f_mnt_write_state != 0);
1074 f->f_mnt_write_state = FILE_MNT_WRITE_TAKEN;
1076 static inline void file_release_write(struct file *f)
1078 f->f_mnt_write_state |= FILE_MNT_WRITE_RELEASED;
1080 static inline void file_reset_write(struct file *f)
1082 f->f_mnt_write_state = 0;
1084 static inline void file_check_state(struct file *f)
1087 * At this point, either both or neither of these bits
1090 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN);
1091 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_RELEASED);
1093 static inline int file_check_writeable(struct file *f)
1095 if (f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN)
1097 printk(KERN_WARNING "writeable file with no "
1098 "mnt_want_write()\n");
1102 #else /* !CONFIG_DEBUG_WRITECOUNT */
1103 static inline void file_take_write(struct file *filp) {}
1104 static inline void file_release_write(struct file *filp) {}
1105 static inline void file_reset_write(struct file *filp) {}
1106 static inline void file_check_state(struct file *filp) {}
1107 static inline int file_check_writeable(struct file *filp)
1111 #endif /* CONFIG_DEBUG_WRITECOUNT */
1113 #define MAX_NON_LFS ((1UL<<31) - 1)
1115 /* Page cache limit. The filesystems should put that into their s_maxbytes
1116 limits, otherwise bad things can happen in VM. */
1117 #if BITS_PER_LONG==32
1118 #define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
1119 #elif BITS_PER_LONG==64
1120 #define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
1125 #define FL_ACCESS 8 /* not trying to lock, just looking */
1126 #define FL_EXISTS 16 /* when unlocking, test for existence */
1127 #define FL_LEASE 32 /* lease held on this file */
1128 #define FL_CLOSE 64 /* unlock on close */
1129 #define FL_SLEEP 128 /* A blocking lock */
1130 #define FL_DOWNGRADE_PENDING 256 /* Lease is being downgraded */
1131 #define FL_UNLOCK_PENDING 512 /* Lease is being broken */
1134 * Special return value from posix_lock_file() and vfs_lock_file() for
1135 * asynchronous locking.
1137 #define FILE_LOCK_DEFERRED 1
1140 * The POSIX file lock owner is determined by
1141 * the "struct files_struct" in the thread group
1142 * (or NULL for no owner - BSD locks).
1144 * Lockd stuffs a "host" pointer into this.
1146 typedef struct files_struct *fl_owner_t;
1148 struct file_lock_operations {
1149 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
1150 void (*fl_release_private)(struct file_lock *);
1153 struct lock_manager_operations {
1154 int (*lm_compare_owner)(struct file_lock *, struct file_lock *);
1155 void (*lm_notify)(struct file_lock *); /* unblock callback */
1156 int (*lm_grant)(struct file_lock *, struct file_lock *, int);
1157 void (*lm_release_private)(struct file_lock *);
1158 void (*lm_break)(struct file_lock *);
1159 int (*lm_change)(struct file_lock **, int);
1162 struct lock_manager {
1163 struct list_head list;
1167 void locks_start_grace(struct net *, struct lock_manager *);
1168 void locks_end_grace(struct lock_manager *);
1169 int locks_in_grace(struct net *);
1171 /* that will die - we need it for nfs_lock_info */
1172 #include <linux/nfs_fs_i.h>
1175 struct file_lock *fl_next; /* singly linked list for this inode */
1176 struct list_head fl_link; /* doubly linked list of all locks */
1177 struct list_head fl_block; /* circular list of blocked processes */
1178 fl_owner_t fl_owner;
1179 unsigned int fl_flags;
1180 unsigned char fl_type;
1181 unsigned int fl_pid;
1182 struct pid *fl_nspid;
1183 wait_queue_head_t fl_wait;
1184 struct file *fl_file;
1188 struct fasync_struct * fl_fasync; /* for lease break notifications */
1189 /* for lease breaks: */
1190 unsigned long fl_break_time;
1191 unsigned long fl_downgrade_time;
1193 const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
1194 const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
1196 struct nfs_lock_info nfs_fl;
1197 struct nfs4_lock_info nfs4_fl;
1199 struct list_head link; /* link in AFS vnode's pending_locks list */
1200 int state; /* state of grant or error if -ve */
1205 /* The following constant reflects the upper bound of the file/locking space */
1207 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
1208 #define OFFSET_MAX INT_LIMIT(loff_t)
1209 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
1212 #include <linux/fcntl.h>
1214 extern void send_sigio(struct fown_struct *fown, int fd, int band);
1216 #ifdef CONFIG_FILE_LOCKING
1217 extern int fcntl_getlk(struct file *, struct flock __user *);
1218 extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
1219 struct flock __user *);
1221 #if BITS_PER_LONG == 32
1222 extern int fcntl_getlk64(struct file *, struct flock64 __user *);
1223 extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
1224 struct flock64 __user *);
1227 extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
1228 extern int fcntl_getlease(struct file *filp);
1231 void locks_free_lock(struct file_lock *fl);
1232 extern void locks_init_lock(struct file_lock *);
1233 extern struct file_lock * locks_alloc_lock(void);
1234 extern void locks_copy_lock(struct file_lock *, struct file_lock *);
1235 extern void __locks_copy_lock(struct file_lock *, const struct file_lock *);
1236 extern void locks_remove_posix(struct file *, fl_owner_t);
1237 extern void locks_remove_flock(struct file *);
1238 extern void locks_release_private(struct file_lock *);
1239 extern void posix_test_lock(struct file *, struct file_lock *);
1240 extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
1241 extern int posix_lock_file_wait(struct file *, struct file_lock *);
1242 extern int posix_unblock_lock(struct file *, struct file_lock *);
1243 extern int vfs_test_lock(struct file *, struct file_lock *);
1244 extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
1245 extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
1246 extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
1247 extern int __break_lease(struct inode *inode, unsigned int flags);
1248 extern void lease_get_mtime(struct inode *, struct timespec *time);
1249 extern int generic_setlease(struct file *, long, struct file_lock **);
1250 extern int vfs_setlease(struct file *, long, struct file_lock **);
1251 extern int lease_modify(struct file_lock **, int);
1252 extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
1253 extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
1254 extern void locks_delete_block(struct file_lock *waiter);
1255 extern void lock_flocks(void);
1256 extern void unlock_flocks(void);
1257 #else /* !CONFIG_FILE_LOCKING */
1258 static inline int fcntl_getlk(struct file *file, struct flock __user *user)
1263 static inline int fcntl_setlk(unsigned int fd, struct file *file,
1264 unsigned int cmd, struct flock __user *user)
1269 #if BITS_PER_LONG == 32
1270 static inline int fcntl_getlk64(struct file *file, struct flock64 __user *user)
1275 static inline int fcntl_setlk64(unsigned int fd, struct file *file,
1276 unsigned int cmd, struct flock64 __user *user)
1281 static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
1286 static inline int fcntl_getlease(struct file *filp)
1291 static inline void locks_init_lock(struct file_lock *fl)
1296 static inline void __locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1301 static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1306 static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
1311 static inline void locks_remove_flock(struct file *filp)
1316 static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
1321 static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
1322 struct file_lock *conflock)
1327 static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
1332 static inline int posix_unblock_lock(struct file *filp,
1333 struct file_lock *waiter)
1338 static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
1343 static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
1344 struct file_lock *fl, struct file_lock *conf)
1349 static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
1354 static inline int flock_lock_file_wait(struct file *filp,
1355 struct file_lock *request)
1360 static inline int __break_lease(struct inode *inode, unsigned int mode)
1365 static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
1370 static inline int generic_setlease(struct file *filp, long arg,
1371 struct file_lock **flp)
1376 static inline int vfs_setlease(struct file *filp, long arg,
1377 struct file_lock **lease)
1382 static inline int lease_modify(struct file_lock **before, int arg)
1387 static inline int lock_may_read(struct inode *inode, loff_t start,
1393 static inline int lock_may_write(struct inode *inode, loff_t start,
1399 static inline void locks_delete_block(struct file_lock *waiter)
1403 static inline void lock_flocks(void)
1407 static inline void unlock_flocks(void)
1411 #endif /* !CONFIG_FILE_LOCKING */
1414 struct fasync_struct {
1418 struct fasync_struct *fa_next; /* singly linked list */
1419 struct file *fa_file;
1420 struct rcu_head fa_rcu;
1423 #define FASYNC_MAGIC 0x4601
1425 /* SMP safe fasync helpers: */
1426 extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
1427 extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
1428 extern int fasync_remove_entry(struct file *, struct fasync_struct **);
1429 extern struct fasync_struct *fasync_alloc(void);
1430 extern void fasync_free(struct fasync_struct *);
1432 /* can be called from interrupts */
1433 extern void kill_fasync(struct fasync_struct **, int, int);
1435 extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
1436 extern int f_setown(struct file *filp, unsigned long arg, int force);
1437 extern void f_delown(struct file *filp);
1438 extern pid_t f_getown(struct file *filp);
1439 extern int send_sigurg(struct fown_struct *fown);
1445 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1446 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1447 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1448 #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1449 #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
1451 extern struct list_head super_blocks;
1452 extern spinlock_t sb_lock;
1454 struct super_block {
1455 struct list_head s_list; /* Keep this first */
1456 dev_t s_dev; /* search index; _not_ kdev_t */
1457 unsigned char s_dirt;
1458 unsigned char s_blocksize_bits;
1459 unsigned long s_blocksize;
1460 loff_t s_maxbytes; /* Max file size */
1461 struct file_system_type *s_type;
1462 const struct super_operations *s_op;
1463 const struct dquot_operations *dq_op;
1464 const struct quotactl_ops *s_qcop;
1465 const struct export_operations *s_export_op;
1466 unsigned long s_flags;
1467 unsigned long s_magic;
1468 struct dentry *s_root;
1469 struct rw_semaphore s_umount;
1470 struct mutex s_lock;
1473 #ifdef CONFIG_SECURITY
1476 const struct xattr_handler **s_xattr;
1478 struct list_head s_inodes; /* all inodes */
1479 struct hlist_bl_head s_anon; /* anonymous dentries for (nfs) exporting */
1481 struct list_head __percpu *s_files;
1483 struct list_head s_files;
1485 struct list_head s_mounts; /* list of mounts; _not_ for fs use */
1486 /* s_dentry_lru, s_nr_dentry_unused protected by dcache.c lru locks */
1487 struct list_head s_dentry_lru; /* unused dentry lru */
1488 int s_nr_dentry_unused; /* # of dentry on lru */
1490 /* s_inode_lru_lock protects s_inode_lru and s_nr_inodes_unused */
1491 spinlock_t s_inode_lru_lock ____cacheline_aligned_in_smp;
1492 struct list_head s_inode_lru; /* unused inode lru */
1493 int s_nr_inodes_unused; /* # of inodes on lru */
1495 struct block_device *s_bdev;
1496 struct backing_dev_info *s_bdi;
1497 struct mtd_info *s_mtd;
1498 struct hlist_node s_instances;
1499 struct quota_info s_dquot; /* Diskquota specific options */
1502 wait_queue_head_t s_wait_unfrozen;
1504 char s_id[32]; /* Informational name */
1505 u8 s_uuid[16]; /* UUID */
1507 void *s_fs_info; /* Filesystem private info */
1508 unsigned int s_max_links;
1511 /* Granularity of c/m/atime in ns.
1512 Cannot be worse than a second */
1516 * The next field is for VFS *only*. No filesystems have any business
1517 * even looking at it. You had been warned.
1519 struct mutex s_vfs_rename_mutex; /* Kludge */
1522 * Filesystem subtype. If non-empty the filesystem type field
1523 * in /proc/mounts will be "type.subtype"
1528 * Saved mount options for lazy filesystems using
1529 * generic_show_options()
1531 char __rcu *s_options;
1532 const struct dentry_operations *s_d_op; /* default d_op for dentries */
1535 * Saved pool identifier for cleancache (-1 means none)
1537 int cleancache_poolid;
1539 struct shrinker s_shrink; /* per-sb shrinker handle */
1541 /* Number of inodes with nlink == 0 but still referenced */
1542 atomic_long_t s_remove_count;
1544 /* Being remounted read-only */
1545 int s_readonly_remount;
1548 /* superblock cache pruning functions */
1549 extern void prune_icache_sb(struct super_block *sb, int nr_to_scan);
1550 extern void prune_dcache_sb(struct super_block *sb, int nr_to_scan);
1552 extern struct timespec current_fs_time(struct super_block *sb);
1555 * Snapshotting support.
1559 SB_FREEZE_WRITE = 1,
1560 SB_FREEZE_TRANS = 2,
1563 #define vfs_check_frozen(sb, level) \
1564 wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
1566 extern bool inode_owner_or_capable(const struct inode *inode);
1568 /* not quite ready to be deprecated, but... */
1569 extern void lock_super(struct super_block *);
1570 extern void unlock_super(struct super_block *);
1573 * VFS helper functions..
1575 extern int vfs_create(struct inode *, struct dentry *, umode_t, struct nameidata *);
1576 extern int vfs_mkdir(struct inode *, struct dentry *, umode_t);
1577 extern int vfs_mknod(struct inode *, struct dentry *, umode_t, dev_t);
1578 extern int vfs_symlink(struct inode *, struct dentry *, const char *);
1579 extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
1580 extern int vfs_rmdir(struct inode *, struct dentry *);
1581 extern int vfs_unlink(struct inode *, struct dentry *);
1582 extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
1585 * VFS dentry helper functions.
1587 extern void dentry_unhash(struct dentry *dentry);
1590 * VFS file helper functions.
1592 extern void inode_init_owner(struct inode *inode, const struct inode *dir,
1595 * VFS FS_IOC_FIEMAP helper definitions.
1597 struct fiemap_extent_info {
1598 unsigned int fi_flags; /* Flags as passed from user */
1599 unsigned int fi_extents_mapped; /* Number of mapped extents */
1600 unsigned int fi_extents_max; /* Size of fiemap_extent array */
1601 struct fiemap_extent __user *fi_extents_start; /* Start of
1602 fiemap_extent array */
1604 int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
1605 u64 phys, u64 len, u32 flags);
1606 int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
1611 * NOTE! These match bits 12..15 of stat.st_mode
1612 * (ie "(i_mode >> 12) & 15").
1614 #define DT_UNKNOWN 0
1625 * This is the "filldir" function type, used by readdir() to let
1626 * the kernel specify what kind of dirent layout it wants to have.
1627 * This allows the kernel to read directories into kernel space or
1628 * to have different dirent layouts depending on the binary type.
1630 typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned);
1631 struct block_device_operations;
1633 /* These macros are for out of kernel modules to test that
1634 * the kernel supports the unlocked_ioctl and compat_ioctl
1635 * fields in struct file_operations. */
1636 #define HAVE_COMPAT_IOCTL 1
1637 #define HAVE_UNLOCKED_IOCTL 1
1639 struct file_operations {
1640 struct module *owner;
1641 loff_t (*llseek) (struct file *, loff_t, int);
1642 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1643 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
1644 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1645 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1646 int (*readdir) (struct file *, void *, filldir_t);
1647 unsigned int (*poll) (struct file *, struct poll_table_struct *);
1648 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1649 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1650 int (*mmap) (struct file *, struct vm_area_struct *);
1651 int (*open) (struct inode *, struct file *);
1652 int (*flush) (struct file *, fl_owner_t id);
1653 int (*release) (struct inode *, struct file *);
1654 int (*fsync) (struct file *, loff_t, loff_t, int datasync);
1655 int (*aio_fsync) (struct kiocb *, int datasync);
1656 int (*fasync) (int, struct file *, int);
1657 int (*lock) (struct file *, int, struct file_lock *);
1658 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1659 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1660 int (*check_flags)(int);
1661 int (*flock) (struct file *, int, struct file_lock *);
1662 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1663 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
1664 int (*setlease)(struct file *, long, struct file_lock **);
1665 long (*fallocate)(struct file *file, int mode, loff_t offset,
1669 struct inode_operations {
1670 struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
1671 void * (*follow_link) (struct dentry *, struct nameidata *);
1672 int (*permission) (struct inode *, int);
1673 struct posix_acl * (*get_acl)(struct inode *, int);
1675 int (*readlink) (struct dentry *, char __user *,int);
1676 void (*put_link) (struct dentry *, struct nameidata *, void *);
1678 int (*create) (struct inode *,struct dentry *,umode_t,struct nameidata *);
1679 int (*link) (struct dentry *,struct inode *,struct dentry *);
1680 int (*unlink) (struct inode *,struct dentry *);
1681 int (*symlink) (struct inode *,struct dentry *,const char *);
1682 int (*mkdir) (struct inode *,struct dentry *,umode_t);
1683 int (*rmdir) (struct inode *,struct dentry *);
1684 int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t);
1685 int (*rename) (struct inode *, struct dentry *,
1686 struct inode *, struct dentry *);
1687 void (*truncate) (struct inode *);
1688 int (*setattr) (struct dentry *, struct iattr *);
1689 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1690 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1691 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1692 ssize_t (*listxattr) (struct dentry *, char *, size_t);
1693 int (*removexattr) (struct dentry *, const char *);
1694 int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
1696 int (*update_time)(struct inode *, struct timespec *, int);
1697 } ____cacheline_aligned;
1701 ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
1702 unsigned long nr_segs, unsigned long fast_segs,
1703 struct iovec *fast_pointer,
1704 struct iovec **ret_pointer);
1706 extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1707 extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1708 extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1709 unsigned long, loff_t *);
1710 extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1711 unsigned long, loff_t *);
1713 struct super_operations {
1714 struct inode *(*alloc_inode)(struct super_block *sb);
1715 void (*destroy_inode)(struct inode *);
1717 void (*dirty_inode) (struct inode *, int flags);
1718 int (*write_inode) (struct inode *, struct writeback_control *wbc);
1719 int (*drop_inode) (struct inode *);
1720 void (*evict_inode) (struct inode *);
1721 void (*put_super) (struct super_block *);
1722 void (*write_super) (struct super_block *);
1723 int (*sync_fs)(struct super_block *sb, int wait);
1724 int (*freeze_fs) (struct super_block *);
1725 int (*unfreeze_fs) (struct super_block *);
1726 int (*statfs) (struct dentry *, struct kstatfs *);
1727 int (*remount_fs) (struct super_block *, int *, char *);
1728 void (*umount_begin) (struct super_block *);
1730 int (*show_options)(struct seq_file *, struct dentry *);
1731 int (*show_devname)(struct seq_file *, struct dentry *);
1732 int (*show_path)(struct seq_file *, struct dentry *);
1733 int (*show_stats)(struct seq_file *, struct dentry *);
1735 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1736 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
1738 int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
1739 int (*nr_cached_objects)(struct super_block *);
1740 void (*free_cached_objects)(struct super_block *, int);
1744 * Inode state bits. Protected by inode->i_lock
1746 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1747 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1749 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1750 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1751 * various stages of removing an inode.
1753 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1755 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1756 * fdatasync(). i_atime is the usual cause.
1757 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1758 * these changes separately from I_DIRTY_SYNC so that we
1759 * don't have to write inode on fdatasync() when only
1760 * mtime has changed in it.
1761 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1762 * I_NEW Serves as both a mutex and completion notification.
1763 * New inodes set I_NEW. If two processes both create
1764 * the same inode, one of them will release its inode and
1765 * wait for I_NEW to be released before returning.
1766 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1767 * also cause waiting on I_NEW, without I_NEW actually
1768 * being set. find_inode() uses this to prevent returning
1769 * nearly-dead inodes.
1770 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1771 * is zero. I_FREEING must be set when I_WILL_FREE is
1773 * I_FREEING Set when inode is about to be freed but still has dirty
1774 * pages or buffers attached or the inode itself is still
1776 * I_CLEAR Added by clear_inode(). In this state the inode is
1777 * clean and can be destroyed. Inode keeps I_FREEING.
1779 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1780 * prohibited for many purposes. iget() must wait for
1781 * the inode to be completely released, then create it
1782 * anew. Other functions will just ignore such inodes,
1783 * if appropriate. I_NEW is used for waiting.
1785 * I_SYNC Writeback of inode is running. The bit is set during
1786 * data writeback, and cleared with a wakeup on the bit
1787 * address once it is done. The bit is also used to pin
1788 * the inode in memory for flusher thread.
1790 * I_REFERENCED Marks the inode as recently references on the LRU list.
1792 * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit().
1794 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1796 #define I_DIRTY_SYNC (1 << 0)
1797 #define I_DIRTY_DATASYNC (1 << 1)
1798 #define I_DIRTY_PAGES (1 << 2)
1800 #define I_NEW (1 << __I_NEW)
1801 #define I_WILL_FREE (1 << 4)
1802 #define I_FREEING (1 << 5)
1803 #define I_CLEAR (1 << 6)
1805 #define I_SYNC (1 << __I_SYNC)
1806 #define I_REFERENCED (1 << 8)
1807 #define __I_DIO_WAKEUP 9
1808 #define I_DIO_WAKEUP (1 << I_DIO_WAKEUP)
1810 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1812 extern void __mark_inode_dirty(struct inode *, int);
1813 static inline void mark_inode_dirty(struct inode *inode)
1815 __mark_inode_dirty(inode, I_DIRTY);
1818 static inline void mark_inode_dirty_sync(struct inode *inode)
1820 __mark_inode_dirty(inode, I_DIRTY_SYNC);
1823 extern void inc_nlink(struct inode *inode);
1824 extern void drop_nlink(struct inode *inode);
1825 extern void clear_nlink(struct inode *inode);
1826 extern void set_nlink(struct inode *inode, unsigned int nlink);
1828 static inline void inode_inc_link_count(struct inode *inode)
1831 mark_inode_dirty(inode);
1834 static inline void inode_dec_link_count(struct inode *inode)
1837 mark_inode_dirty(inode);
1841 * inode_inc_iversion - increments i_version
1842 * @inode: inode that need to be updated
1844 * Every time the inode is modified, the i_version field will be incremented.
1845 * The filesystem has to be mounted with i_version flag
1848 static inline void inode_inc_iversion(struct inode *inode)
1850 spin_lock(&inode->i_lock);
1852 spin_unlock(&inode->i_lock);
1855 enum file_time_flags {
1862 extern void touch_atime(struct path *);
1863 static inline void file_accessed(struct file *file)
1865 if (!(file->f_flags & O_NOATIME))
1866 touch_atime(&file->f_path);
1869 int sync_inode(struct inode *inode, struct writeback_control *wbc);
1870 int sync_inode_metadata(struct inode *inode, int wait);
1872 struct file_system_type {
1875 struct dentry *(*mount) (struct file_system_type *, int,
1876 const char *, void *);
1877 void (*kill_sb) (struct super_block *);
1878 struct module *owner;
1879 struct file_system_type * next;
1880 struct hlist_head fs_supers;
1882 struct lock_class_key s_lock_key;
1883 struct lock_class_key s_umount_key;
1884 struct lock_class_key s_vfs_rename_key;
1886 struct lock_class_key i_lock_key;
1887 struct lock_class_key i_mutex_key;
1888 struct lock_class_key i_mutex_dir_key;
1891 extern struct dentry *mount_ns(struct file_system_type *fs_type, int flags,
1892 void *data, int (*fill_super)(struct super_block *, void *, int));
1893 extern struct dentry *mount_bdev(struct file_system_type *fs_type,
1894 int flags, const char *dev_name, void *data,
1895 int (*fill_super)(struct super_block *, void *, int));
1896 extern struct dentry *mount_single(struct file_system_type *fs_type,
1897 int flags, void *data,
1898 int (*fill_super)(struct super_block *, void *, int));
1899 extern struct dentry *mount_nodev(struct file_system_type *fs_type,
1900 int flags, void *data,
1901 int (*fill_super)(struct super_block *, void *, int));
1902 extern struct dentry *mount_subtree(struct vfsmount *mnt, const char *path);
1903 void generic_shutdown_super(struct super_block *sb);
1904 void kill_block_super(struct super_block *sb);
1905 void kill_anon_super(struct super_block *sb);
1906 void kill_litter_super(struct super_block *sb);
1907 void deactivate_super(struct super_block *sb);
1908 void deactivate_locked_super(struct super_block *sb);
1909 int set_anon_super(struct super_block *s, void *data);
1910 int get_anon_bdev(dev_t *);
1911 void free_anon_bdev(dev_t);
1912 struct super_block *sget(struct file_system_type *type,
1913 int (*test)(struct super_block *,void *),
1914 int (*set)(struct super_block *,void *),
1916 extern struct dentry *mount_pseudo(struct file_system_type *, char *,
1917 const struct super_operations *ops,
1918 const struct dentry_operations *dops,
1921 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1922 #define fops_get(fops) \
1923 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1924 #define fops_put(fops) \
1925 do { if (fops) module_put((fops)->owner); } while(0)
1927 extern int register_filesystem(struct file_system_type *);
1928 extern int unregister_filesystem(struct file_system_type *);
1929 extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
1930 #define kern_mount(type) kern_mount_data(type, NULL)
1931 extern void kern_unmount(struct vfsmount *mnt);
1932 extern int may_umount_tree(struct vfsmount *);
1933 extern int may_umount(struct vfsmount *);
1934 extern long do_mount(char *, char *, char *, unsigned long, void *);
1935 extern struct vfsmount *collect_mounts(struct path *);
1936 extern void drop_collected_mounts(struct vfsmount *);
1937 extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
1939 extern int vfs_statfs(struct path *, struct kstatfs *);
1940 extern int user_statfs(const char __user *, struct kstatfs *);
1941 extern int fd_statfs(int, struct kstatfs *);
1942 extern int vfs_ustat(dev_t, struct kstatfs *);
1943 extern int freeze_super(struct super_block *super);
1944 extern int thaw_super(struct super_block *super);
1945 extern bool our_mnt(struct vfsmount *mnt);
1947 extern int current_umask(void);
1950 extern struct kobject *fs_kobj;
1952 #define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
1953 extern int rw_verify_area(int, struct file *, loff_t *, size_t);
1955 #define FLOCK_VERIFY_READ 1
1956 #define FLOCK_VERIFY_WRITE 2
1958 #ifdef CONFIG_FILE_LOCKING
1959 extern int locks_mandatory_locked(struct inode *);
1960 extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
1963 * Candidates for mandatory locking have the setgid bit set
1964 * but no group execute bit - an otherwise meaningless combination.
1967 static inline int __mandatory_lock(struct inode *ino)
1969 return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
1973 * ... and these candidates should be on MS_MANDLOCK mounted fs,
1974 * otherwise these will be advisory locks
1977 static inline int mandatory_lock(struct inode *ino)
1979 return IS_MANDLOCK(ino) && __mandatory_lock(ino);
1982 static inline int locks_verify_locked(struct inode *inode)
1984 if (mandatory_lock(inode))
1985 return locks_mandatory_locked(inode);
1989 static inline int locks_verify_truncate(struct inode *inode,
1993 if (inode->i_flock && mandatory_lock(inode))
1994 return locks_mandatory_area(
1995 FLOCK_VERIFY_WRITE, inode, filp,
1996 size < inode->i_size ? size : inode->i_size,
1997 (size < inode->i_size ? inode->i_size - size
1998 : size - inode->i_size)
2003 static inline int break_lease(struct inode *inode, unsigned int mode)
2006 return __break_lease(inode, mode);
2009 #else /* !CONFIG_FILE_LOCKING */
2010 static inline int locks_mandatory_locked(struct inode *inode)
2015 static inline int locks_mandatory_area(int rw, struct inode *inode,
2016 struct file *filp, loff_t offset,
2022 static inline int __mandatory_lock(struct inode *inode)
2027 static inline int mandatory_lock(struct inode *inode)
2032 static inline int locks_verify_locked(struct inode *inode)
2037 static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
2043 static inline int break_lease(struct inode *inode, unsigned int mode)
2048 #endif /* CONFIG_FILE_LOCKING */
2052 extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
2054 extern int do_fallocate(struct file *file, int mode, loff_t offset,
2056 extern long do_sys_open(int dfd, const char __user *filename, int flags,
2058 extern struct file *filp_open(const char *, int, umode_t);
2059 extern struct file *file_open_root(struct dentry *, struct vfsmount *,
2061 extern struct file * dentry_open(struct dentry *, struct vfsmount *, int,
2062 const struct cred *);
2063 extern int filp_close(struct file *, fl_owner_t id);
2064 extern char * getname(const char __user *);
2068 extern int ioctl_preallocate(struct file *filp, void __user *argp);
2071 extern void __init vfs_caches_init_early(void);
2072 extern void __init vfs_caches_init(unsigned long);
2074 extern struct kmem_cache *names_cachep;
2076 #define __getname_gfp(gfp) kmem_cache_alloc(names_cachep, (gfp))
2077 #define __getname() __getname_gfp(GFP_KERNEL)
2078 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
2079 #ifndef CONFIG_AUDITSYSCALL
2080 #define putname(name) __putname(name)
2082 extern void putname(const char *name);
2086 extern int register_blkdev(unsigned int, const char *);
2087 extern void unregister_blkdev(unsigned int, const char *);
2088 extern struct block_device *bdget(dev_t);
2089 extern struct block_device *bdgrab(struct block_device *bdev);
2090 extern void bd_set_size(struct block_device *, loff_t size);
2091 extern sector_t blkdev_max_block(struct block_device *bdev);
2092 extern void bd_forget(struct inode *inode);
2093 extern void bdput(struct block_device *);
2094 extern void invalidate_bdev(struct block_device *);
2095 extern int sync_blockdev(struct block_device *bdev);
2096 extern void kill_bdev(struct block_device *);
2097 extern struct super_block *freeze_bdev(struct block_device *);
2098 extern void emergency_thaw_all(void);
2099 extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
2100 extern int fsync_bdev(struct block_device *);
2102 static inline void bd_forget(struct inode *inode) {}
2103 static inline int sync_blockdev(struct block_device *bdev) { return 0; }
2104 static inline void kill_bdev(struct block_device *bdev) {}
2105 static inline void invalidate_bdev(struct block_device *bdev) {}
2107 static inline struct super_block *freeze_bdev(struct block_device *sb)
2112 static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
2117 extern int sync_filesystem(struct super_block *);
2118 extern const struct file_operations def_blk_fops;
2119 extern const struct file_operations def_chr_fops;
2120 extern const struct file_operations bad_sock_fops;
2121 extern const struct file_operations def_fifo_fops;
2123 extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
2124 extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
2125 extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
2126 extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
2127 extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
2129 extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
2131 extern int blkdev_put(struct block_device *bdev, fmode_t mode);
2133 extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
2134 extern void bd_unlink_disk_holder(struct block_device *bdev,
2135 struct gendisk *disk);
2137 static inline int bd_link_disk_holder(struct block_device *bdev,
2138 struct gendisk *disk)
2142 static inline void bd_unlink_disk_holder(struct block_device *bdev,
2143 struct gendisk *disk)
2150 #define CHRDEV_MAJOR_HASH_SIZE 255
2151 extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
2152 extern int register_chrdev_region(dev_t, unsigned, const char *);
2153 extern int __register_chrdev(unsigned int major, unsigned int baseminor,
2154 unsigned int count, const char *name,
2155 const struct file_operations *fops);
2156 extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
2157 unsigned int count, const char *name);
2158 extern void unregister_chrdev_region(dev_t, unsigned);
2159 extern void chrdev_show(struct seq_file *,off_t);
2161 static inline int register_chrdev(unsigned int major, const char *name,
2162 const struct file_operations *fops)
2164 return __register_chrdev(major, 0, 256, name, fops);
2167 static inline void unregister_chrdev(unsigned int major, const char *name)
2169 __unregister_chrdev(major, 0, 256, name);
2172 /* fs/block_dev.c */
2173 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
2174 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
2177 #define BLKDEV_MAJOR_HASH_SIZE 255
2178 extern const char *__bdevname(dev_t, char *buffer);
2179 extern const char *bdevname(struct block_device *bdev, char *buffer);
2180 extern struct block_device *lookup_bdev(const char *);
2181 extern void blkdev_show(struct seq_file *,off_t);
2184 #define BLKDEV_MAJOR_HASH_SIZE 0
2187 extern void init_special_inode(struct inode *, umode_t, dev_t);
2189 /* Invalid inode operations -- fs/bad_inode.c */
2190 extern void make_bad_inode(struct inode *);
2191 extern int is_bad_inode(struct inode *);
2193 extern const struct file_operations read_pipefifo_fops;
2194 extern const struct file_operations write_pipefifo_fops;
2195 extern const struct file_operations rdwr_pipefifo_fops;
2199 * return READ, READA, or WRITE
2201 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2204 * return data direction, READ or WRITE
2206 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
2208 extern void check_disk_size_change(struct gendisk *disk,
2209 struct block_device *bdev);
2210 extern int revalidate_disk(struct gendisk *);
2211 extern int check_disk_change(struct block_device *);
2212 extern int __invalidate_device(struct block_device *, bool);
2213 extern int invalidate_partition(struct gendisk *, int);
2215 unsigned long invalidate_mapping_pages(struct address_space *mapping,
2216 pgoff_t start, pgoff_t end);
2218 static inline void invalidate_remote_inode(struct inode *inode)
2220 if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
2221 S_ISLNK(inode->i_mode))
2222 invalidate_mapping_pages(inode->i_mapping, 0, -1);
2224 extern int invalidate_inode_pages2(struct address_space *mapping);
2225 extern int invalidate_inode_pages2_range(struct address_space *mapping,
2226 pgoff_t start, pgoff_t end);
2227 extern int write_inode_now(struct inode *, int);
2228 extern int filemap_fdatawrite(struct address_space *);
2229 extern int filemap_flush(struct address_space *);
2230 extern int filemap_fdatawait(struct address_space *);
2231 extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
2233 extern int filemap_write_and_wait(struct address_space *mapping);
2234 extern int filemap_write_and_wait_range(struct address_space *mapping,
2235 loff_t lstart, loff_t lend);
2236 extern int __filemap_fdatawrite_range(struct address_space *mapping,
2237 loff_t start, loff_t end, int sync_mode);
2238 extern int filemap_fdatawrite_range(struct address_space *mapping,
2239 loff_t start, loff_t end);
2241 extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
2243 extern int vfs_fsync(struct file *file, int datasync);
2244 extern int generic_write_sync(struct file *file, loff_t pos, loff_t count);
2245 extern void sync_supers(void);
2246 extern void emergency_sync(void);
2247 extern void emergency_remount(void);
2249 extern sector_t bmap(struct inode *, sector_t);
2251 extern int notify_change(struct dentry *, struct iattr *);
2252 extern int inode_permission(struct inode *, int);
2253 extern int generic_permission(struct inode *, int);
2255 static inline bool execute_ok(struct inode *inode)
2257 return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
2261 * get_write_access() gets write permission for a file.
2262 * put_write_access() releases this write permission.
2263 * This is used for regular files.
2264 * We cannot support write (and maybe mmap read-write shared) accesses and
2265 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2266 * can have the following values:
2267 * 0: no writers, no VM_DENYWRITE mappings
2268 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2269 * > 0: (i_writecount) users are writing to the file.
2271 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2272 * except for the cases where we don't hold i_writecount yet. Then we need to
2273 * use {get,deny}_write_access() - these functions check the sign and refuse
2274 * to do the change if sign is wrong.
2276 static inline int get_write_access(struct inode *inode)
2278 return atomic_inc_unless_negative(&inode->i_writecount) ? 0 : -ETXTBSY;
2280 static inline int deny_write_access(struct file *file)
2282 struct inode *inode = file->f_path.dentry->d_inode;
2283 return atomic_dec_unless_positive(&inode->i_writecount) ? 0 : -ETXTBSY;
2285 static inline void put_write_access(struct inode * inode)
2287 atomic_dec(&inode->i_writecount);
2289 static inline void allow_write_access(struct file *file)
2292 atomic_inc(&file->f_path.dentry->d_inode->i_writecount);
2295 static inline void i_readcount_dec(struct inode *inode)
2297 BUG_ON(!atomic_read(&inode->i_readcount));
2298 atomic_dec(&inode->i_readcount);
2300 static inline void i_readcount_inc(struct inode *inode)
2302 atomic_inc(&inode->i_readcount);
2305 static inline void i_readcount_dec(struct inode *inode)
2309 static inline void i_readcount_inc(struct inode *inode)
2314 extern int do_pipe_flags(int *, int);
2315 extern struct file *create_read_pipe(struct file *f, int flags);
2316 extern struct file *create_write_pipe(int flags);
2317 extern void free_write_pipe(struct file *);
2319 extern int kernel_read(struct file *, loff_t, char *, unsigned long);
2320 extern struct file * open_exec(const char *);
2322 /* fs/dcache.c -- generic fs support functions */
2323 extern int is_subdir(struct dentry *, struct dentry *);
2324 extern int path_is_under(struct path *, struct path *);
2325 extern ino_t find_inode_number(struct dentry *, struct qstr *);
2327 #include <linux/err.h>
2329 /* needed for stackable file system support */
2330 extern loff_t default_llseek(struct file *file, loff_t offset, int origin);
2332 extern loff_t vfs_llseek(struct file *file, loff_t offset, int origin);
2334 extern int inode_init_always(struct super_block *, struct inode *);
2335 extern void inode_init_once(struct inode *);
2336 extern void address_space_init_once(struct address_space *mapping);
2337 extern void ihold(struct inode * inode);
2338 extern void iput(struct inode *);
2339 extern struct inode * igrab(struct inode *);
2340 extern ino_t iunique(struct super_block *, ino_t);
2341 extern int inode_needs_sync(struct inode *inode);
2342 extern int generic_delete_inode(struct inode *inode);
2343 static inline int generic_drop_inode(struct inode *inode)
2345 return !inode->i_nlink || inode_unhashed(inode);
2348 extern struct inode *ilookup5_nowait(struct super_block *sb,
2349 unsigned long hashval, int (*test)(struct inode *, void *),
2351 extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
2352 int (*test)(struct inode *, void *), void *data);
2353 extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
2355 extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
2356 extern struct inode * iget_locked(struct super_block *, unsigned long);
2357 extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
2358 extern int insert_inode_locked(struct inode *);
2359 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2360 extern void lockdep_annotate_inode_mutex_key(struct inode *inode);
2362 static inline void lockdep_annotate_inode_mutex_key(struct inode *inode) { };
2364 extern void unlock_new_inode(struct inode *);
2365 extern unsigned int get_next_ino(void);
2367 extern void __iget(struct inode * inode);
2368 extern void iget_failed(struct inode *);
2369 extern void clear_inode(struct inode *);
2370 extern void __destroy_inode(struct inode *);
2371 extern struct inode *new_inode_pseudo(struct super_block *sb);
2372 extern struct inode *new_inode(struct super_block *sb);
2373 extern void free_inode_nonrcu(struct inode *inode);
2374 extern int should_remove_suid(struct dentry *);
2375 extern int file_remove_suid(struct file *);
2377 extern void __insert_inode_hash(struct inode *, unsigned long hashval);
2378 static inline void insert_inode_hash(struct inode *inode)
2380 __insert_inode_hash(inode, inode->i_ino);
2383 extern void __remove_inode_hash(struct inode *);
2384 static inline void remove_inode_hash(struct inode *inode)
2386 if (!inode_unhashed(inode))
2387 __remove_inode_hash(inode);
2390 extern void inode_sb_list_add(struct inode *inode);
2393 extern void submit_bio(int, struct bio *);
2394 extern int bdev_read_only(struct block_device *);
2396 extern int set_blocksize(struct block_device *, int);
2397 extern int sb_set_blocksize(struct super_block *, int);
2398 extern int sb_min_blocksize(struct super_block *, int);
2400 extern int generic_file_mmap(struct file *, struct vm_area_struct *);
2401 extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
2402 extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
2403 int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
2404 extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2405 extern ssize_t __generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long,
2407 extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2408 extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
2409 unsigned long *, loff_t, loff_t *, size_t, size_t);
2410 extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
2411 unsigned long, loff_t, loff_t *, size_t, ssize_t);
2412 extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
2413 extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
2414 extern int generic_segment_checks(const struct iovec *iov,
2415 unsigned long *nr_segs, size_t *count, int access_flags);
2417 /* fs/block_dev.c */
2418 extern ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
2419 unsigned long nr_segs, loff_t pos);
2420 extern int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
2422 extern void block_sync_page(struct page *page);
2425 extern ssize_t generic_file_splice_read(struct file *, loff_t *,
2426 struct pipe_inode_info *, size_t, unsigned int);
2427 extern ssize_t default_file_splice_read(struct file *, loff_t *,
2428 struct pipe_inode_info *, size_t, unsigned int);
2429 extern ssize_t generic_file_splice_write(struct pipe_inode_info *,
2430 struct file *, loff_t *, size_t, unsigned int);
2431 extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
2432 struct file *out, loff_t *, size_t len, unsigned int flags);
2433 extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
2434 size_t len, unsigned int flags);
2437 file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
2438 extern loff_t noop_llseek(struct file *file, loff_t offset, int origin);
2439 extern loff_t no_llseek(struct file *file, loff_t offset, int origin);
2440 extern loff_t generic_file_llseek(struct file *file, loff_t offset, int origin);
2441 extern loff_t generic_file_llseek_size(struct file *file, loff_t offset,
2442 int origin, loff_t maxsize);
2443 extern int generic_file_open(struct inode * inode, struct file * filp);
2444 extern int nonseekable_open(struct inode * inode, struct file * filp);
2446 #ifdef CONFIG_FS_XIP
2447 extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
2449 extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
2450 extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
2451 size_t len, loff_t *ppos);
2452 extern int xip_truncate_page(struct address_space *mapping, loff_t from);
2454 static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
2461 typedef void (dio_submit_t)(int rw, struct bio *bio, struct inode *inode,
2462 loff_t file_offset);
2465 /* need locking between buffered and direct access */
2468 /* filesystem does not support filling holes */
2469 DIO_SKIP_HOLES = 0x02,
2472 void dio_end_io(struct bio *bio, int error);
2474 ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
2475 struct block_device *bdev, const struct iovec *iov, loff_t offset,
2476 unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
2477 dio_submit_t submit_io, int flags);
2479 static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
2480 struct inode *inode, const struct iovec *iov, loff_t offset,
2481 unsigned long nr_segs, get_block_t get_block)
2483 return __blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
2484 offset, nr_segs, get_block, NULL, NULL,
2485 DIO_LOCKING | DIO_SKIP_HOLES);
2489 void inode_dio_wait(struct inode *inode);
2490 void inode_dio_done(struct inode *inode);
2492 extern const struct file_operations generic_ro_fops;
2494 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2496 extern int vfs_readlink(struct dentry *, char __user *, int, const char *);
2497 extern int vfs_follow_link(struct nameidata *, const char *);
2498 extern int page_readlink(struct dentry *, char __user *, int);
2499 extern void *page_follow_link_light(struct dentry *, struct nameidata *);
2500 extern void page_put_link(struct dentry *, struct nameidata *, void *);
2501 extern int __page_symlink(struct inode *inode, const char *symname, int len,
2503 extern int page_symlink(struct inode *inode, const char *symname, int len);
2504 extern const struct inode_operations page_symlink_inode_operations;
2505 extern int generic_readlink(struct dentry *, char __user *, int);
2506 extern void generic_fillattr(struct inode *, struct kstat *);
2507 extern int vfs_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2508 void __inode_add_bytes(struct inode *inode, loff_t bytes);
2509 void inode_add_bytes(struct inode *inode, loff_t bytes);
2510 void inode_sub_bytes(struct inode *inode, loff_t bytes);
2511 loff_t inode_get_bytes(struct inode *inode);
2512 void inode_set_bytes(struct inode *inode, loff_t bytes);
2514 extern int vfs_readdir(struct file *, filldir_t, void *);
2516 extern int vfs_stat(const char __user *, struct kstat *);
2517 extern int vfs_lstat(const char __user *, struct kstat *);
2518 extern int vfs_fstat(unsigned int, struct kstat *);
2519 extern int vfs_fstatat(int , const char __user *, struct kstat *, int);
2521 extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
2523 extern int __generic_block_fiemap(struct inode *inode,
2524 struct fiemap_extent_info *fieinfo,
2525 loff_t start, loff_t len,
2526 get_block_t *get_block);
2527 extern int generic_block_fiemap(struct inode *inode,
2528 struct fiemap_extent_info *fieinfo, u64 start,
2529 u64 len, get_block_t *get_block);
2531 extern void get_filesystem(struct file_system_type *fs);
2532 extern void put_filesystem(struct file_system_type *fs);
2533 extern struct file_system_type *get_fs_type(const char *name);
2534 extern struct super_block *get_super(struct block_device *);
2535 extern struct super_block *get_super_thawed(struct block_device *);
2536 extern struct super_block *get_active_super(struct block_device *bdev);
2537 extern void drop_super(struct super_block *sb);
2538 extern void iterate_supers(void (*)(struct super_block *, void *), void *);
2539 extern void iterate_supers_type(struct file_system_type *,
2540 void (*)(struct super_block *, void *), void *);
2542 extern int dcache_dir_open(struct inode *, struct file *);
2543 extern int dcache_dir_close(struct inode *, struct file *);
2544 extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
2545 extern int dcache_readdir(struct file *, void *, filldir_t);
2546 extern int simple_setattr(struct dentry *, struct iattr *);
2547 extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2548 extern int simple_statfs(struct dentry *, struct kstatfs *);
2549 extern int simple_open(struct inode *inode, struct file *file);
2550 extern int simple_link(struct dentry *, struct inode *, struct dentry *);
2551 extern int simple_unlink(struct inode *, struct dentry *);
2552 extern int simple_rmdir(struct inode *, struct dentry *);
2553 extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
2554 extern int noop_fsync(struct file *, loff_t, loff_t, int);
2555 extern int simple_empty(struct dentry *);
2556 extern int simple_readpage(struct file *file, struct page *page);
2557 extern int simple_write_begin(struct file *file, struct address_space *mapping,
2558 loff_t pos, unsigned len, unsigned flags,
2559 struct page **pagep, void **fsdata);
2560 extern int simple_write_end(struct file *file, struct address_space *mapping,
2561 loff_t pos, unsigned len, unsigned copied,
2562 struct page *page, void *fsdata);
2564 extern struct dentry *simple_lookup(struct inode *, struct dentry *, struct nameidata *);
2565 extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
2566 extern const struct file_operations simple_dir_operations;
2567 extern const struct inode_operations simple_dir_inode_operations;
2568 struct tree_descr { char *name; const struct file_operations *ops; int mode; };
2569 struct dentry *d_alloc_name(struct dentry *, const char *);
2570 extern int simple_fill_super(struct super_block *, unsigned long, struct tree_descr *);
2571 extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
2572 extern void simple_release_fs(struct vfsmount **mount, int *count);
2574 extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
2575 loff_t *ppos, const void *from, size_t available);
2576 extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
2577 const void __user *from, size_t count);
2579 extern int generic_file_fsync(struct file *, loff_t, loff_t, int);
2581 extern int generic_check_addressable(unsigned, u64);
2583 #ifdef CONFIG_MIGRATION
2584 extern int buffer_migrate_page(struct address_space *,
2585 struct page *, struct page *,
2588 #define buffer_migrate_page NULL
2591 extern int inode_change_ok(const struct inode *, struct iattr *);
2592 extern int inode_newsize_ok(const struct inode *, loff_t offset);
2593 extern void setattr_copy(struct inode *inode, const struct iattr *attr);
2595 extern int file_update_time(struct file *file);
2597 extern int generic_show_options(struct seq_file *m, struct dentry *root);
2598 extern void save_mount_options(struct super_block *sb, char *options);
2599 extern void replace_mount_options(struct super_block *sb, char *options);
2601 static inline ino_t parent_ino(struct dentry *dentry)
2606 * Don't strictly need d_lock here? If the parent ino could change
2607 * then surely we'd have a deeper race in the caller?
2609 spin_lock(&dentry->d_lock);
2610 res = dentry->d_parent->d_inode->i_ino;
2611 spin_unlock(&dentry->d_lock);
2615 /* Transaction based IO helpers */
2618 * An argresp is stored in an allocated page and holds the
2619 * size of the argument or response, along with its content
2621 struct simple_transaction_argresp {
2626 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2628 char *simple_transaction_get(struct file *file, const char __user *buf,
2630 ssize_t simple_transaction_read(struct file *file, char __user *buf,
2631 size_t size, loff_t *pos);
2632 int simple_transaction_release(struct inode *inode, struct file *file);
2634 void simple_transaction_set(struct file *file, size_t n);
2637 * simple attribute files
2639 * These attributes behave similar to those in sysfs:
2641 * Writing to an attribute immediately sets a value, an open file can be
2642 * written to multiple times.
2644 * Reading from an attribute creates a buffer from the value that might get
2645 * read with multiple read calls. When the attribute has been read
2646 * completely, no further read calls are possible until the file is opened
2649 * All attributes contain a text representation of a numeric value
2650 * that are accessed with the get() and set() functions.
2652 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2653 static int __fops ## _open(struct inode *inode, struct file *file) \
2655 __simple_attr_check_format(__fmt, 0ull); \
2656 return simple_attr_open(inode, file, __get, __set, __fmt); \
2658 static const struct file_operations __fops = { \
2659 .owner = THIS_MODULE, \
2660 .open = __fops ## _open, \
2661 .release = simple_attr_release, \
2662 .read = simple_attr_read, \
2663 .write = simple_attr_write, \
2664 .llseek = generic_file_llseek, \
2667 static inline __printf(1, 2)
2668 void __simple_attr_check_format(const char *fmt, ...)
2670 /* don't do anything, just let the compiler check the arguments; */
2673 int simple_attr_open(struct inode *inode, struct file *file,
2674 int (*get)(void *, u64 *), int (*set)(void *, u64),
2676 int simple_attr_release(struct inode *inode, struct file *file);
2677 ssize_t simple_attr_read(struct file *file, char __user *buf,
2678 size_t len, loff_t *ppos);
2679 ssize_t simple_attr_write(struct file *file, const char __user *buf,
2680 size_t len, loff_t *ppos);
2683 int proc_nr_files(struct ctl_table *table, int write,
2684 void __user *buffer, size_t *lenp, loff_t *ppos);
2685 int proc_nr_dentry(struct ctl_table *table, int write,
2686 void __user *buffer, size_t *lenp, loff_t *ppos);
2687 int proc_nr_inodes(struct ctl_table *table, int write,
2688 void __user *buffer, size_t *lenp, loff_t *ppos);
2689 int __init get_filesystem_list(char *buf);
2691 #define __FMODE_EXEC ((__force int) FMODE_EXEC)
2692 #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
2694 #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
2695 #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
2696 (flag & __FMODE_NONOTIFY)))
2698 static inline int is_sxid(umode_t mode)
2700 return (mode & S_ISUID) || ((mode & S_ISGID) && (mode & S_IXGRP));
2703 static inline void inode_has_no_xattr(struct inode *inode)
2705 if (!is_sxid(inode->i_mode) && (inode->i_sb->s_flags & MS_NOSEC))
2706 inode->i_flags |= S_NOSEC;
2709 #endif /* __KERNEL__ */
2710 #endif /* _LINUX_FS_H */