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_MAX SEEK_END
43 /* And dynamically-tunable limits and defaults: */
44 struct files_stat_struct {
45 unsigned long nr_files; /* read only */
46 unsigned long nr_free_files; /* read only */
47 unsigned long max_files; /* tunable */
50 struct inodes_stat_t {
53 int dummy[5]; /* padding for sysctl ABI compatibility */
57 #define NR_FILE 8192 /* this can well be larger on a larger system */
66 #define MAY_NOT_BLOCK 128 /* called from RCU mode, don't block */
69 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
70 * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
73 /* file is open for reading */
74 #define FMODE_READ ((__force fmode_t)0x1)
75 /* file is open for writing */
76 #define FMODE_WRITE ((__force fmode_t)0x2)
77 /* file is seekable */
78 #define FMODE_LSEEK ((__force fmode_t)0x4)
79 /* file can be accessed using pread */
80 #define FMODE_PREAD ((__force fmode_t)0x8)
81 /* file can be accessed using pwrite */
82 #define FMODE_PWRITE ((__force fmode_t)0x10)
83 /* File is opened for execution with sys_execve / sys_uselib */
84 #define FMODE_EXEC ((__force fmode_t)0x20)
85 /* File is opened with O_NDELAY (only set for block devices) */
86 #define FMODE_NDELAY ((__force fmode_t)0x40)
87 /* File is opened with O_EXCL (only set for block devices) */
88 #define FMODE_EXCL ((__force fmode_t)0x80)
89 /* File is opened using open(.., 3, ..) and is writeable only for ioctls
90 (specialy hack for floppy.c) */
91 #define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
94 * Don't update ctime and mtime.
96 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
97 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
99 #define FMODE_NOCMTIME ((__force fmode_t)0x800)
101 /* Expect random access pattern */
102 #define FMODE_RANDOM ((__force fmode_t)0x1000)
104 /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
105 #define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
107 /* File is opened with O_PATH; almost nothing can be done with it */
108 #define FMODE_PATH ((__force fmode_t)0x4000)
110 /* File was opened by fanotify and shouldn't generate fanotify events */
111 #define FMODE_NONOTIFY ((__force fmode_t)0x1000000)
114 * The below are the various read and write types that we support. Some of
115 * them include behavioral modifiers that send information down to the
116 * block layer and IO scheduler. Terminology:
118 * The block layer uses device plugging to defer IO a little bit, in
119 * the hope that we will see more IO very shortly. This increases
120 * coalescing of adjacent IO and thus reduces the number of IOs we
121 * have to send to the device. It also allows for better queuing,
122 * if the IO isn't mergeable. If the caller is going to be waiting
123 * for the IO, then he must ensure that the device is unplugged so
124 * that the IO is dispatched to the driver.
126 * All IO is handled async in Linux. This is fine for background
127 * writes, but for reads or writes that someone waits for completion
128 * on, we want to notify the block layer and IO scheduler so that they
129 * know about it. That allows them to make better scheduling
130 * decisions. So when the below references 'sync' and 'async', it
131 * is referencing this priority hint.
133 * With that in mind, the available types are:
135 * READ A normal read operation. Device will be plugged.
136 * READ_SYNC A synchronous read. Device is not plugged, caller can
137 * immediately wait on this read without caring about
139 * READA Used for read-ahead operations. Lower priority, and the
140 * block layer could (in theory) choose to ignore this
141 * request if it runs into resource problems.
142 * WRITE A normal async write. Device will be plugged.
143 * WRITE_SYNC Synchronous write. Identical to WRITE, but passes down
144 * the hint that someone will be waiting on this IO
145 * shortly. The write equivalent of READ_SYNC.
146 * WRITE_ODIRECT Special case write for O_DIRECT only.
147 * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
148 * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
149 * non-volatile media on completion.
150 * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
151 * by a cache flush and data is guaranteed to be on
152 * non-volatile media on completion.
155 #define RW_MASK REQ_WRITE
156 #define RWA_MASK REQ_RAHEAD
159 #define WRITE RW_MASK
160 #define READA RWA_MASK
162 #define READ_SYNC (READ | REQ_SYNC)
163 #define READ_META (READ | REQ_META)
164 #define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE)
165 #define WRITE_ODIRECT (WRITE | REQ_SYNC)
166 #define WRITE_META (WRITE | REQ_META)
167 #define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
168 #define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
169 #define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
175 /* public flags for file_system_type */
176 #define FS_REQUIRES_DEV 1
177 #define FS_BINARY_MOUNTDATA 2
178 #define FS_HAS_SUBTYPE 4
179 #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
180 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move()
181 * during rename() internally.
185 * These are the fs-independent mount-flags: up to 32 flags are supported
187 #define MS_RDONLY 1 /* Mount read-only */
188 #define MS_NOSUID 2 /* Ignore suid and sgid bits */
189 #define MS_NODEV 4 /* Disallow access to device special files */
190 #define MS_NOEXEC 8 /* Disallow program execution */
191 #define MS_SYNCHRONOUS 16 /* Writes are synced at once */
192 #define MS_REMOUNT 32 /* Alter flags of a mounted FS */
193 #define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
194 #define MS_DIRSYNC 128 /* Directory modifications are synchronous */
195 #define MS_NOATIME 1024 /* Do not update access times. */
196 #define MS_NODIRATIME 2048 /* Do not update directory access times */
200 #define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
201 MS_VERBOSE is deprecated. */
202 #define MS_SILENT 32768
203 #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
204 #define MS_UNBINDABLE (1<<17) /* change to unbindable */
205 #define MS_PRIVATE (1<<18) /* change to private */
206 #define MS_SLAVE (1<<19) /* change to slave */
207 #define MS_SHARED (1<<20) /* change to shared */
208 #define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */
209 #define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
210 #define MS_I_VERSION (1<<23) /* Update inode I_version field */
211 #define MS_STRICTATIME (1<<24) /* Always perform atime updates */
212 #define MS_NOSEC (1<<28)
213 #define MS_BORN (1<<29)
214 #define MS_ACTIVE (1<<30)
215 #define MS_NOUSER (1<<31)
218 * Superblock flags that can be altered by MS_REMOUNT
220 #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK|MS_I_VERSION)
223 * Old magic mount flag and mask
225 #define MS_MGC_VAL 0xC0ED0000
226 #define MS_MGC_MSK 0xffff0000
228 /* Inode flags - they have nothing to superblock flags now */
230 #define S_SYNC 1 /* Writes are synced at once */
231 #define S_NOATIME 2 /* Do not update access times */
232 #define S_APPEND 4 /* Append-only file */
233 #define S_IMMUTABLE 8 /* Immutable file */
234 #define S_DEAD 16 /* removed, but still open directory */
235 #define S_NOQUOTA 32 /* Inode is not counted to quota */
236 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
237 #define S_NOCMTIME 128 /* Do not update file c/mtime */
238 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
239 #define S_PRIVATE 512 /* Inode is fs-internal */
240 #define S_IMA 1024 /* Inode has an associated IMA struct */
241 #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
242 #define S_NOSEC 4096 /* no suid or xattr security attributes */
245 * Note that nosuid etc flags are inode-specific: setting some file-system
246 * flags just means all the inodes inherit those flags by default. It might be
247 * possible to override it selectively if you really wanted to with some
248 * ioctl() that is not currently implemented.
250 * Exception: MS_RDONLY is always applied to the entire file system.
252 * Unfortunately, it is possible to change a filesystems flags with it mounted
253 * with files in use. This means that all of the inodes will not have their
254 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
255 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
257 #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
259 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
260 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
261 ((inode)->i_flags & S_SYNC))
262 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
263 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
264 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
265 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
266 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
268 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
269 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
270 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
271 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
273 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
274 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
275 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
276 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
277 #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
278 #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
279 #define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
281 /* the read-only stuff doesn't really belong here, but any other place is
282 probably as bad and I don't want to create yet another include file. */
284 #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
285 #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
286 #define BLKRRPART _IO(0x12,95) /* re-read partition table */
287 #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
288 #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
289 #define BLKRASET _IO(0x12,98) /* set read ahead for block device */
290 #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
291 #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
292 #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
293 #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
294 #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
295 #define BLKSSZGET _IO(0x12,104)/* get block device sector size */
297 #define BLKPG _IO(0x12,105)/* See blkpg.h */
299 /* Some people are morons. Do not use sizeof! */
301 #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
302 #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
303 /* This was here just to show that the number is taken -
304 probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
306 /* A jump here: 108-111 have been used for various private purposes. */
307 #define BLKBSZGET _IOR(0x12,112,size_t)
308 #define BLKBSZSET _IOW(0x12,113,size_t)
309 #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
310 #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
311 #define BLKTRACESTART _IO(0x12,116)
312 #define BLKTRACESTOP _IO(0x12,117)
313 #define BLKTRACETEARDOWN _IO(0x12,118)
314 #define BLKDISCARD _IO(0x12,119)
315 #define BLKIOMIN _IO(0x12,120)
316 #define BLKIOOPT _IO(0x12,121)
317 #define BLKALIGNOFF _IO(0x12,122)
318 #define BLKPBSZGET _IO(0x12,123)
319 #define BLKDISCARDZEROES _IO(0x12,124)
320 #define BLKSECDISCARD _IO(0x12,125)
322 #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
323 #define FIBMAP _IO(0x00,1) /* bmap access */
324 #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
325 #define FIFREEZE _IOWR('X', 119, int) /* Freeze */
326 #define FITHAW _IOWR('X', 120, int) /* Thaw */
327 #define FITRIM _IOWR('X', 121, struct fstrim_range) /* Trim */
329 #define FS_IOC_GETFLAGS _IOR('f', 1, long)
330 #define FS_IOC_SETFLAGS _IOW('f', 2, long)
331 #define FS_IOC_GETVERSION _IOR('v', 1, long)
332 #define FS_IOC_SETVERSION _IOW('v', 2, long)
333 #define FS_IOC_FIEMAP _IOWR('f', 11, struct fiemap)
334 #define FS_IOC32_GETFLAGS _IOR('f', 1, int)
335 #define FS_IOC32_SETFLAGS _IOW('f', 2, int)
336 #define FS_IOC32_GETVERSION _IOR('v', 1, int)
337 #define FS_IOC32_SETVERSION _IOW('v', 2, int)
340 * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS)
342 #define FS_SECRM_FL 0x00000001 /* Secure deletion */
343 #define FS_UNRM_FL 0x00000002 /* Undelete */
344 #define FS_COMPR_FL 0x00000004 /* Compress file */
345 #define FS_SYNC_FL 0x00000008 /* Synchronous updates */
346 #define FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
347 #define FS_APPEND_FL 0x00000020 /* writes to file may only append */
348 #define FS_NODUMP_FL 0x00000040 /* do not dump file */
349 #define FS_NOATIME_FL 0x00000080 /* do not update atime */
350 /* Reserved for compression usage... */
351 #define FS_DIRTY_FL 0x00000100
352 #define FS_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
353 #define FS_NOCOMP_FL 0x00000400 /* Don't compress */
354 #define FS_ECOMPR_FL 0x00000800 /* Compression error */
355 /* End compression flags --- maybe not all used */
356 #define FS_BTREE_FL 0x00001000 /* btree format dir */
357 #define FS_INDEX_FL 0x00001000 /* hash-indexed directory */
358 #define FS_IMAGIC_FL 0x00002000 /* AFS directory */
359 #define FS_JOURNAL_DATA_FL 0x00004000 /* Reserved for ext3 */
360 #define FS_NOTAIL_FL 0x00008000 /* file tail should not be merged */
361 #define FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
362 #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
363 #define FS_EXTENT_FL 0x00080000 /* Extents */
364 #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */
365 #define FS_NOCOW_FL 0x00800000 /* Do not cow file */
366 #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */
368 #define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
369 #define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
372 #define SYNC_FILE_RANGE_WAIT_BEFORE 1
373 #define SYNC_FILE_RANGE_WRITE 2
374 #define SYNC_FILE_RANGE_WAIT_AFTER 4
378 #include <linux/linkage.h>
379 #include <linux/wait.h>
380 #include <linux/types.h>
381 #include <linux/kdev_t.h>
382 #include <linux/dcache.h>
383 #include <linux/path.h>
384 #include <linux/stat.h>
385 #include <linux/cache.h>
386 #include <linux/list.h>
387 #include <linux/radix-tree.h>
388 #include <linux/prio_tree.h>
389 #include <linux/init.h>
390 #include <linux/pid.h>
391 #include <linux/mutex.h>
392 #include <linux/capability.h>
393 #include <linux/semaphore.h>
394 #include <linux/fiemap.h>
395 #include <linux/rculist_bl.h>
396 #include <linux/shrinker.h>
397 #include <linux/atomic.h>
399 #include <asm/byteorder.h>
401 struct export_operations;
407 struct pipe_inode_info;
408 struct poll_table_struct;
410 struct vm_area_struct;
414 extern void __init inode_init(void);
415 extern void __init inode_init_early(void);
416 extern void __init files_init(unsigned long);
418 extern struct files_stat_struct files_stat;
419 extern unsigned long get_max_files(void);
420 extern int sysctl_nr_open;
421 extern struct inodes_stat_t inodes_stat;
422 extern int leases_enable, lease_break_time;
425 typedef int (get_block_t)(struct inode *inode, sector_t iblock,
426 struct buffer_head *bh_result, int create);
427 typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
428 ssize_t bytes, void *private, int ret,
432 * Attribute flags. These should be or-ed together to figure out what
435 #define ATTR_MODE (1 << 0)
436 #define ATTR_UID (1 << 1)
437 #define ATTR_GID (1 << 2)
438 #define ATTR_SIZE (1 << 3)
439 #define ATTR_ATIME (1 << 4)
440 #define ATTR_MTIME (1 << 5)
441 #define ATTR_CTIME (1 << 6)
442 #define ATTR_ATIME_SET (1 << 7)
443 #define ATTR_MTIME_SET (1 << 8)
444 #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
445 #define ATTR_ATTR_FLAG (1 << 10)
446 #define ATTR_KILL_SUID (1 << 11)
447 #define ATTR_KILL_SGID (1 << 12)
448 #define ATTR_FILE (1 << 13)
449 #define ATTR_KILL_PRIV (1 << 14)
450 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
451 #define ATTR_TIMES_SET (1 << 16)
454 * This is the Inode Attributes structure, used for notify_change(). It
455 * uses the above definitions as flags, to know which values have changed.
456 * Also, in this manner, a Filesystem can look at only the values it cares
457 * about. Basically, these are the attributes that the VFS layer can
458 * request to change from the FS layer.
460 * Derek Atkins <warlord@MIT.EDU> 94-10-20
463 unsigned int ia_valid;
468 struct timespec ia_atime;
469 struct timespec ia_mtime;
470 struct timespec ia_ctime;
473 * Not an attribute, but an auxiliary info for filesystems wanting to
474 * implement an ftruncate() like method. NOTE: filesystem should
475 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
477 struct file *ia_file;
481 * Includes for diskquotas.
483 #include <linux/quota.h>
486 * enum positive_aop_returns - aop return codes with specific semantics
488 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
489 * completed, that the page is still locked, and
490 * should be considered active. The VM uses this hint
491 * to return the page to the active list -- it won't
492 * be a candidate for writeback again in the near
493 * future. Other callers must be careful to unlock
494 * the page if they get this return. Returned by
497 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
498 * unlocked it and the page might have been truncated.
499 * The caller should back up to acquiring a new page and
500 * trying again. The aop will be taking reasonable
501 * precautions not to livelock. If the caller held a page
502 * reference, it should drop it before retrying. Returned
505 * address_space_operation functions return these large constants to indicate
506 * special semantics to the caller. These are much larger than the bytes in a
507 * page to allow for functions that return the number of bytes operated on in a
511 enum positive_aop_returns {
512 AOP_WRITEPAGE_ACTIVATE = 0x80000,
513 AOP_TRUNCATED_PAGE = 0x80001,
516 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
517 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
518 #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
519 * helper code (eg buffer layer)
520 * to clear GFP_FS from alloc */
523 * oh the beauties of C type declarations.
526 struct address_space;
527 struct writeback_control;
530 const struct iovec *iov;
531 unsigned long nr_segs;
536 size_t iov_iter_copy_from_user_atomic(struct page *page,
537 struct iov_iter *i, unsigned long offset, size_t bytes);
538 size_t iov_iter_copy_from_user(struct page *page,
539 struct iov_iter *i, unsigned long offset, size_t bytes);
540 void iov_iter_advance(struct iov_iter *i, size_t bytes);
541 int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes);
542 size_t iov_iter_single_seg_count(struct iov_iter *i);
544 static inline void iov_iter_init(struct iov_iter *i,
545 const struct iovec *iov, unsigned long nr_segs,
546 size_t count, size_t written)
549 i->nr_segs = nr_segs;
551 i->count = count + written;
553 iov_iter_advance(i, written);
556 static inline size_t iov_iter_count(struct iov_iter *i)
562 * "descriptor" for what we're up to with a read.
563 * This allows us to use the same read code yet
564 * have multiple different users of the data that
565 * we read from a file.
567 * The simplest case just copies the data to user
580 typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
581 unsigned long, unsigned long);
583 struct address_space_operations {
584 int (*writepage)(struct page *page, struct writeback_control *wbc);
585 int (*readpage)(struct file *, struct page *);
587 /* Write back some dirty pages from this mapping. */
588 int (*writepages)(struct address_space *, struct writeback_control *);
590 /* Set a page dirty. Return true if this dirtied it */
591 int (*set_page_dirty)(struct page *page);
593 int (*readpages)(struct file *filp, struct address_space *mapping,
594 struct list_head *pages, unsigned nr_pages);
596 int (*write_begin)(struct file *, struct address_space *mapping,
597 loff_t pos, unsigned len, unsigned flags,
598 struct page **pagep, void **fsdata);
599 int (*write_end)(struct file *, struct address_space *mapping,
600 loff_t pos, unsigned len, unsigned copied,
601 struct page *page, void *fsdata);
603 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
604 sector_t (*bmap)(struct address_space *, sector_t);
605 void (*invalidatepage) (struct page *, unsigned long);
606 int (*releasepage) (struct page *, gfp_t);
607 void (*freepage)(struct page *);
608 ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
609 loff_t offset, unsigned long nr_segs);
610 int (*get_xip_mem)(struct address_space *, pgoff_t, int,
611 void **, unsigned long *);
612 /* migrate the contents of a page to the specified target */
613 int (*migratepage) (struct address_space *,
614 struct page *, struct page *);
615 int (*launder_page) (struct page *);
616 int (*is_partially_uptodate) (struct page *, read_descriptor_t *,
618 int (*error_remove_page)(struct address_space *, struct page *);
621 extern const struct address_space_operations empty_aops;
624 * pagecache_write_begin/pagecache_write_end must be used by general code
625 * to write into the pagecache.
627 int pagecache_write_begin(struct file *, struct address_space *mapping,
628 loff_t pos, unsigned len, unsigned flags,
629 struct page **pagep, void **fsdata);
631 int pagecache_write_end(struct file *, struct address_space *mapping,
632 loff_t pos, unsigned len, unsigned copied,
633 struct page *page, void *fsdata);
635 struct backing_dev_info;
636 struct address_space {
637 struct inode *host; /* owner: inode, block_device */
638 struct radix_tree_root page_tree; /* radix tree of all pages */
639 spinlock_t tree_lock; /* and lock protecting it */
640 unsigned int i_mmap_writable;/* count VM_SHARED mappings */
641 struct prio_tree_root i_mmap; /* tree of private and shared mappings */
642 struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
643 struct mutex i_mmap_mutex; /* protect tree, count, list */
644 /* Protected by tree_lock together with the radix tree */
645 unsigned long nrpages; /* number of total pages */
646 pgoff_t writeback_index;/* writeback starts here */
647 const struct address_space_operations *a_ops; /* methods */
648 unsigned long flags; /* error bits/gfp mask */
649 struct backing_dev_info *backing_dev_info; /* device readahead, etc */
650 spinlock_t private_lock; /* for use by the address_space */
651 struct list_head private_list; /* ditto */
652 struct address_space *assoc_mapping; /* ditto */
653 } __attribute__((aligned(sizeof(long))));
655 * On most architectures that alignment is already the case; but
656 * must be enforced here for CRIS, to let the least significant bit
657 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
660 struct block_device {
661 dev_t bd_dev; /* not a kdev_t - it's a search key */
663 struct inode * bd_inode; /* will die */
664 struct super_block * bd_super;
665 struct mutex bd_mutex; /* open/close mutex */
666 struct list_head bd_inodes;
670 bool bd_write_holder;
672 struct list_head bd_holder_disks;
674 struct block_device * bd_contains;
675 unsigned bd_block_size;
676 struct hd_struct * bd_part;
677 /* number of times partitions within this device have been opened. */
678 unsigned bd_part_count;
680 struct gendisk * bd_disk;
681 struct list_head bd_list;
683 * Private data. You must have bd_claim'ed the block_device
684 * to use this. NOTE: bd_claim allows an owner to claim
685 * the same device multiple times, the owner must take special
686 * care to not mess up bd_private for that case.
688 unsigned long bd_private;
690 /* The counter of freeze processes */
691 int bd_fsfreeze_count;
692 /* Mutex for freeze */
693 struct mutex bd_fsfreeze_mutex;
697 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
700 #define PAGECACHE_TAG_DIRTY 0
701 #define PAGECACHE_TAG_WRITEBACK 1
702 #define PAGECACHE_TAG_TOWRITE 2
704 int mapping_tagged(struct address_space *mapping, int tag);
707 * Might pages of this file be mapped into userspace?
709 static inline int mapping_mapped(struct address_space *mapping)
711 return !prio_tree_empty(&mapping->i_mmap) ||
712 !list_empty(&mapping->i_mmap_nonlinear);
716 * Might pages of this file have been modified in userspace?
717 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
718 * marks vma as VM_SHARED if it is shared, and the file was opened for
719 * writing i.e. vma may be mprotected writable even if now readonly.
721 static inline int mapping_writably_mapped(struct address_space *mapping)
723 return mapping->i_mmap_writable != 0;
727 * Use sequence counter to get consistent i_size on 32-bit processors.
729 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
730 #include <linux/seqlock.h>
731 #define __NEED_I_SIZE_ORDERED
732 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
734 #define i_size_ordered_init(inode) do { } while (0)
738 #define ACL_NOT_CACHED ((void *)(-1))
741 /* RCU path lookup touches following: */
745 const struct inode_operations *i_op;
746 struct super_block *i_sb;
748 spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
749 unsigned int i_flags;
750 unsigned long i_state;
751 #ifdef CONFIG_SECURITY
754 struct mutex i_mutex;
757 unsigned long dirtied_when; /* jiffies of first dirtying */
759 struct hlist_node i_hash;
760 struct list_head i_wb_list; /* backing dev IO list */
761 struct list_head i_lru; /* inode LRU list */
762 struct list_head i_sb_list;
764 struct list_head i_dentry;
765 struct rcu_head i_rcu;
769 unsigned int i_nlink;
771 unsigned int i_blkbits;
774 #ifdef __NEED_I_SIZE_ORDERED
775 seqcount_t i_size_seqcount;
777 struct timespec i_atime;
778 struct timespec i_mtime;
779 struct timespec i_ctime;
781 unsigned short i_bytes;
782 struct rw_semaphore i_alloc_sem;
783 const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
784 struct file_lock *i_flock;
785 struct address_space *i_mapping;
786 struct address_space i_data;
788 struct dquot *i_dquot[MAXQUOTAS];
790 struct list_head i_devices;
792 struct pipe_inode_info *i_pipe;
793 struct block_device *i_bdev;
799 #ifdef CONFIG_FSNOTIFY
800 __u32 i_fsnotify_mask; /* all events this inode cares about */
801 struct hlist_head i_fsnotify_marks;
805 atomic_t i_readcount; /* struct files open RO */
807 atomic_t i_writecount;
808 #ifdef CONFIG_FS_POSIX_ACL
809 struct posix_acl *i_acl;
810 struct posix_acl *i_default_acl;
812 void *i_private; /* fs or device private pointer */
815 static inline int inode_unhashed(struct inode *inode)
817 return hlist_unhashed(&inode->i_hash);
821 * inode->i_mutex nesting subclasses for the lock validator:
823 * 0: the object of the current VFS operation
828 * The locking order between these classes is
829 * parent -> child -> normal -> xattr -> quota
831 enum inode_i_mutex_lock_class
841 * NOTE: in a 32bit arch with a preemptable kernel and
842 * an UP compile the i_size_read/write must be atomic
843 * with respect to the local cpu (unlike with preempt disabled),
844 * but they don't need to be atomic with respect to other cpus like in
845 * true SMP (so they need either to either locally disable irq around
846 * the read or for example on x86 they can be still implemented as a
847 * cmpxchg8b without the need of the lock prefix). For SMP compiles
848 * and 64bit archs it makes no difference if preempt is enabled or not.
850 static inline loff_t i_size_read(const struct inode *inode)
852 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
857 seq = read_seqcount_begin(&inode->i_size_seqcount);
858 i_size = inode->i_size;
859 } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
861 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
865 i_size = inode->i_size;
869 return inode->i_size;
874 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
875 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
876 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
878 static inline void i_size_write(struct inode *inode, loff_t i_size)
880 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
881 write_seqcount_begin(&inode->i_size_seqcount);
882 inode->i_size = i_size;
883 write_seqcount_end(&inode->i_size_seqcount);
884 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
886 inode->i_size = i_size;
889 inode->i_size = i_size;
893 static inline unsigned iminor(const struct inode *inode)
895 return MINOR(inode->i_rdev);
898 static inline unsigned imajor(const struct inode *inode)
900 return MAJOR(inode->i_rdev);
903 extern struct block_device *I_BDEV(struct inode *inode);
906 rwlock_t lock; /* protects pid, uid, euid fields */
907 struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
908 enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
909 uid_t uid, euid; /* uid/euid of process setting the owner */
910 int signum; /* posix.1b rt signal to be delivered on IO */
914 * Track a single file's readahead state
916 struct file_ra_state {
917 pgoff_t start; /* where readahead started */
918 unsigned int size; /* # of readahead pages */
919 unsigned int async_size; /* do asynchronous readahead when
920 there are only # of pages ahead */
922 unsigned int ra_pages; /* Maximum readahead window */
923 unsigned int mmap_miss; /* Cache miss stat for mmap accesses */
924 loff_t prev_pos; /* Cache last read() position */
928 * Check if @index falls in the readahead windows.
930 static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
932 return (index >= ra->start &&
933 index < ra->start + ra->size);
936 #define FILE_MNT_WRITE_TAKEN 1
937 #define FILE_MNT_WRITE_RELEASED 2
941 * fu_list becomes invalid after file_free is called and queued via
942 * fu_rcuhead for RCU freeing
945 struct list_head fu_list;
946 struct rcu_head fu_rcuhead;
949 #define f_dentry f_path.dentry
950 #define f_vfsmnt f_path.mnt
951 const struct file_operations *f_op;
952 spinlock_t f_lock; /* f_ep_links, f_flags, no IRQ */
956 atomic_long_t f_count;
957 unsigned int f_flags;
960 struct fown_struct f_owner;
961 const struct cred *f_cred;
962 struct file_ra_state f_ra;
965 #ifdef CONFIG_SECURITY
968 /* needed for tty driver, and maybe others */
972 /* Used by fs/eventpoll.c to link all the hooks to this file */
973 struct list_head f_ep_links;
974 #endif /* #ifdef CONFIG_EPOLL */
975 struct address_space *f_mapping;
976 #ifdef CONFIG_DEBUG_WRITECOUNT
977 unsigned long f_mnt_write_state;
984 /* file identifier */
985 unsigned char f_handle[0];
988 #define get_file(x) atomic_long_inc(&(x)->f_count)
989 #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
990 #define file_count(x) atomic_long_read(&(x)->f_count)
992 #ifdef CONFIG_DEBUG_WRITECOUNT
993 static inline void file_take_write(struct file *f)
995 WARN_ON(f->f_mnt_write_state != 0);
996 f->f_mnt_write_state = FILE_MNT_WRITE_TAKEN;
998 static inline void file_release_write(struct file *f)
1000 f->f_mnt_write_state |= FILE_MNT_WRITE_RELEASED;
1002 static inline void file_reset_write(struct file *f)
1004 f->f_mnt_write_state = 0;
1006 static inline void file_check_state(struct file *f)
1009 * At this point, either both or neither of these bits
1012 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN);
1013 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_RELEASED);
1015 static inline int file_check_writeable(struct file *f)
1017 if (f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN)
1019 printk(KERN_WARNING "writeable file with no "
1020 "mnt_want_write()\n");
1024 #else /* !CONFIG_DEBUG_WRITECOUNT */
1025 static inline void file_take_write(struct file *filp) {}
1026 static inline void file_release_write(struct file *filp) {}
1027 static inline void file_reset_write(struct file *filp) {}
1028 static inline void file_check_state(struct file *filp) {}
1029 static inline int file_check_writeable(struct file *filp)
1033 #endif /* CONFIG_DEBUG_WRITECOUNT */
1035 #define MAX_NON_LFS ((1UL<<31) - 1)
1037 /* Page cache limit. The filesystems should put that into their s_maxbytes
1038 limits, otherwise bad things can happen in VM. */
1039 #if BITS_PER_LONG==32
1040 #define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
1041 #elif BITS_PER_LONG==64
1042 #define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
1047 #define FL_ACCESS 8 /* not trying to lock, just looking */
1048 #define FL_EXISTS 16 /* when unlocking, test for existence */
1049 #define FL_LEASE 32 /* lease held on this file */
1050 #define FL_CLOSE 64 /* unlock on close */
1051 #define FL_SLEEP 128 /* A blocking lock */
1054 * Special return value from posix_lock_file() and vfs_lock_file() for
1055 * asynchronous locking.
1057 #define FILE_LOCK_DEFERRED 1
1060 * The POSIX file lock owner is determined by
1061 * the "struct files_struct" in the thread group
1062 * (or NULL for no owner - BSD locks).
1064 * Lockd stuffs a "host" pointer into this.
1066 typedef struct files_struct *fl_owner_t;
1068 struct file_lock_operations {
1069 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
1070 void (*fl_release_private)(struct file_lock *);
1073 struct lock_manager_operations {
1074 int (*fl_compare_owner)(struct file_lock *, struct file_lock *);
1075 void (*fl_notify)(struct file_lock *); /* unblock callback */
1076 int (*fl_grant)(struct file_lock *, struct file_lock *, int);
1077 void (*fl_release_private)(struct file_lock *);
1078 void (*fl_break)(struct file_lock *);
1079 int (*fl_change)(struct file_lock **, int);
1082 struct lock_manager {
1083 struct list_head list;
1086 void locks_start_grace(struct lock_manager *);
1087 void locks_end_grace(struct lock_manager *);
1088 int locks_in_grace(void);
1090 /* that will die - we need it for nfs_lock_info */
1091 #include <linux/nfs_fs_i.h>
1094 struct file_lock *fl_next; /* singly linked list for this inode */
1095 struct list_head fl_link; /* doubly linked list of all locks */
1096 struct list_head fl_block; /* circular list of blocked processes */
1097 fl_owner_t fl_owner;
1098 unsigned char fl_flags;
1099 unsigned char fl_type;
1100 unsigned int fl_pid;
1101 struct pid *fl_nspid;
1102 wait_queue_head_t fl_wait;
1103 struct file *fl_file;
1107 struct fasync_struct * fl_fasync; /* for lease break notifications */
1108 unsigned long fl_break_time; /* for nonblocking lease breaks */
1110 const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
1111 const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
1113 struct nfs_lock_info nfs_fl;
1114 struct nfs4_lock_info nfs4_fl;
1116 struct list_head link; /* link in AFS vnode's pending_locks list */
1117 int state; /* state of grant or error if -ve */
1122 /* The following constant reflects the upper bound of the file/locking space */
1124 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
1125 #define OFFSET_MAX INT_LIMIT(loff_t)
1126 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
1129 #include <linux/fcntl.h>
1131 extern void send_sigio(struct fown_struct *fown, int fd, int band);
1133 #ifdef CONFIG_FILE_LOCKING
1134 extern int fcntl_getlk(struct file *, struct flock __user *);
1135 extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
1136 struct flock __user *);
1138 #if BITS_PER_LONG == 32
1139 extern int fcntl_getlk64(struct file *, struct flock64 __user *);
1140 extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
1141 struct flock64 __user *);
1144 extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
1145 extern int fcntl_getlease(struct file *filp);
1148 void locks_free_lock(struct file_lock *fl);
1149 extern void locks_init_lock(struct file_lock *);
1150 extern struct file_lock * locks_alloc_lock(void);
1151 extern void locks_copy_lock(struct file_lock *, struct file_lock *);
1152 extern void __locks_copy_lock(struct file_lock *, const struct file_lock *);
1153 extern void locks_remove_posix(struct file *, fl_owner_t);
1154 extern void locks_remove_flock(struct file *);
1155 extern void locks_release_private(struct file_lock *);
1156 extern void posix_test_lock(struct file *, struct file_lock *);
1157 extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
1158 extern int posix_lock_file_wait(struct file *, struct file_lock *);
1159 extern int posix_unblock_lock(struct file *, struct file_lock *);
1160 extern int vfs_test_lock(struct file *, struct file_lock *);
1161 extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
1162 extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
1163 extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
1164 extern int __break_lease(struct inode *inode, unsigned int flags);
1165 extern void lease_get_mtime(struct inode *, struct timespec *time);
1166 extern int generic_setlease(struct file *, long, struct file_lock **);
1167 extern int vfs_setlease(struct file *, long, struct file_lock **);
1168 extern int lease_modify(struct file_lock **, int);
1169 extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
1170 extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
1171 extern void lock_flocks(void);
1172 extern void unlock_flocks(void);
1173 #else /* !CONFIG_FILE_LOCKING */
1174 static inline int fcntl_getlk(struct file *file, struct flock __user *user)
1179 static inline int fcntl_setlk(unsigned int fd, struct file *file,
1180 unsigned int cmd, struct flock __user *user)
1185 #if BITS_PER_LONG == 32
1186 static inline int fcntl_getlk64(struct file *file, struct flock64 __user *user)
1191 static inline int fcntl_setlk64(unsigned int fd, struct file *file,
1192 unsigned int cmd, struct flock64 __user *user)
1197 static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
1202 static inline int fcntl_getlease(struct file *filp)
1207 static inline void locks_init_lock(struct file_lock *fl)
1212 static inline void __locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1217 static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1222 static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
1227 static inline void locks_remove_flock(struct file *filp)
1232 static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
1237 static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
1238 struct file_lock *conflock)
1243 static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
1248 static inline int posix_unblock_lock(struct file *filp,
1249 struct file_lock *waiter)
1254 static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
1259 static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
1260 struct file_lock *fl, struct file_lock *conf)
1265 static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
1270 static inline int flock_lock_file_wait(struct file *filp,
1271 struct file_lock *request)
1276 static inline int __break_lease(struct inode *inode, unsigned int mode)
1281 static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
1286 static inline int generic_setlease(struct file *filp, long arg,
1287 struct file_lock **flp)
1292 static inline int vfs_setlease(struct file *filp, long arg,
1293 struct file_lock **lease)
1298 static inline int lease_modify(struct file_lock **before, int arg)
1303 static inline int lock_may_read(struct inode *inode, loff_t start,
1309 static inline int lock_may_write(struct inode *inode, loff_t start,
1315 static inline void lock_flocks(void)
1319 static inline void unlock_flocks(void)
1323 #endif /* !CONFIG_FILE_LOCKING */
1326 struct fasync_struct {
1330 struct fasync_struct *fa_next; /* singly linked list */
1331 struct file *fa_file;
1332 struct rcu_head fa_rcu;
1335 #define FASYNC_MAGIC 0x4601
1337 /* SMP safe fasync helpers: */
1338 extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
1339 extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
1340 extern int fasync_remove_entry(struct file *, struct fasync_struct **);
1341 extern struct fasync_struct *fasync_alloc(void);
1342 extern void fasync_free(struct fasync_struct *);
1344 /* can be called from interrupts */
1345 extern void kill_fasync(struct fasync_struct **, int, int);
1347 extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
1348 extern int f_setown(struct file *filp, unsigned long arg, int force);
1349 extern void f_delown(struct file *filp);
1350 extern pid_t f_getown(struct file *filp);
1351 extern int send_sigurg(struct fown_struct *fown);
1357 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1358 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1359 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1360 #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1361 #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
1363 extern struct list_head super_blocks;
1364 extern spinlock_t sb_lock;
1366 struct super_block {
1367 struct list_head s_list; /* Keep this first */
1368 dev_t s_dev; /* search index; _not_ kdev_t */
1369 unsigned char s_dirt;
1370 unsigned char s_blocksize_bits;
1371 unsigned long s_blocksize;
1372 loff_t s_maxbytes; /* Max file size */
1373 struct file_system_type *s_type;
1374 const struct super_operations *s_op;
1375 const struct dquot_operations *dq_op;
1376 const struct quotactl_ops *s_qcop;
1377 const struct export_operations *s_export_op;
1378 unsigned long s_flags;
1379 unsigned long s_magic;
1380 struct dentry *s_root;
1381 struct rw_semaphore s_umount;
1382 struct mutex s_lock;
1385 #ifdef CONFIG_SECURITY
1388 const struct xattr_handler **s_xattr;
1390 struct list_head s_inodes; /* all inodes */
1391 struct hlist_bl_head s_anon; /* anonymous dentries for (nfs) exporting */
1393 struct list_head __percpu *s_files;
1395 struct list_head s_files;
1397 /* s_dentry_lru, s_nr_dentry_unused protected by dcache.c lru locks */
1398 struct list_head s_dentry_lru; /* unused dentry lru */
1399 int s_nr_dentry_unused; /* # of dentry on lru */
1401 /* s_inode_lru_lock protects s_inode_lru and s_nr_inodes_unused */
1402 spinlock_t s_inode_lru_lock ____cacheline_aligned_in_smp;
1403 struct list_head s_inode_lru; /* unused inode lru */
1404 int s_nr_inodes_unused; /* # of inodes on lru */
1406 struct block_device *s_bdev;
1407 struct backing_dev_info *s_bdi;
1408 struct mtd_info *s_mtd;
1409 struct list_head s_instances;
1410 struct quota_info s_dquot; /* Diskquota specific options */
1413 wait_queue_head_t s_wait_unfrozen;
1415 char s_id[32]; /* Informational name */
1416 u8 s_uuid[16]; /* UUID */
1418 void *s_fs_info; /* Filesystem private info */
1421 /* Granularity of c/m/atime in ns.
1422 Cannot be worse than a second */
1426 * The next field is for VFS *only*. No filesystems have any business
1427 * even looking at it. You had been warned.
1429 struct mutex s_vfs_rename_mutex; /* Kludge */
1432 * Filesystem subtype. If non-empty the filesystem type field
1433 * in /proc/mounts will be "type.subtype"
1438 * Saved mount options for lazy filesystems using
1439 * generic_show_options()
1441 char __rcu *s_options;
1442 const struct dentry_operations *s_d_op; /* default d_op for dentries */
1445 * Saved pool identifier for cleancache (-1 means none)
1447 int cleancache_poolid;
1449 struct shrinker s_shrink; /* per-sb shrinker handle */
1452 /* superblock cache pruning functions */
1453 extern void prune_icache_sb(struct super_block *sb, int nr_to_scan);
1454 extern void prune_dcache_sb(struct super_block *sb, int nr_to_scan);
1456 extern struct timespec current_fs_time(struct super_block *sb);
1459 * Snapshotting support.
1463 SB_FREEZE_WRITE = 1,
1464 SB_FREEZE_TRANS = 2,
1467 #define vfs_check_frozen(sb, level) \
1468 wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
1470 #define get_fs_excl() atomic_inc(¤t->fs_excl)
1471 #define put_fs_excl() atomic_dec(¤t->fs_excl)
1472 #define has_fs_excl() atomic_read(¤t->fs_excl)
1475 * until VFS tracks user namespaces for inodes, just make all files
1476 * belong to init_user_ns
1478 extern struct user_namespace init_user_ns;
1479 #define inode_userns(inode) (&init_user_ns)
1480 extern bool inode_owner_or_capable(const struct inode *inode);
1482 /* not quite ready to be deprecated, but... */
1483 extern void lock_super(struct super_block *);
1484 extern void unlock_super(struct super_block *);
1487 * VFS helper functions..
1489 extern int vfs_create(struct inode *, struct dentry *, int, struct nameidata *);
1490 extern int vfs_mkdir(struct inode *, struct dentry *, int);
1491 extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t);
1492 extern int vfs_symlink(struct inode *, struct dentry *, const char *);
1493 extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
1494 extern int vfs_rmdir(struct inode *, struct dentry *);
1495 extern int vfs_unlink(struct inode *, struct dentry *);
1496 extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
1499 * VFS dentry helper functions.
1501 extern void dentry_unhash(struct dentry *dentry);
1504 * VFS file helper functions.
1506 extern void inode_init_owner(struct inode *inode, const struct inode *dir,
1509 * VFS FS_IOC_FIEMAP helper definitions.
1511 struct fiemap_extent_info {
1512 unsigned int fi_flags; /* Flags as passed from user */
1513 unsigned int fi_extents_mapped; /* Number of mapped extents */
1514 unsigned int fi_extents_max; /* Size of fiemap_extent array */
1515 struct fiemap_extent __user *fi_extents_start; /* Start of
1516 fiemap_extent array */
1518 int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
1519 u64 phys, u64 len, u32 flags);
1520 int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
1525 * NOTE! These match bits 12..15 of stat.st_mode
1526 * (ie "(i_mode >> 12) & 15").
1528 #define DT_UNKNOWN 0
1539 * This is the "filldir" function type, used by readdir() to let
1540 * the kernel specify what kind of dirent layout it wants to have.
1541 * This allows the kernel to read directories into kernel space or
1542 * to have different dirent layouts depending on the binary type.
1544 typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned);
1545 struct block_device_operations;
1547 /* These macros are for out of kernel modules to test that
1548 * the kernel supports the unlocked_ioctl and compat_ioctl
1549 * fields in struct file_operations. */
1550 #define HAVE_COMPAT_IOCTL 1
1551 #define HAVE_UNLOCKED_IOCTL 1
1555 * all file operations except setlease can be called without
1556 * the big kernel lock held in all filesystems.
1558 struct file_operations {
1559 struct module *owner;
1560 loff_t (*llseek) (struct file *, loff_t, int);
1561 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1562 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
1563 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1564 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1565 int (*readdir) (struct file *, void *, filldir_t);
1566 unsigned int (*poll) (struct file *, struct poll_table_struct *);
1567 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1568 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1569 int (*mmap) (struct file *, struct vm_area_struct *);
1570 int (*open) (struct inode *, struct file *);
1571 int (*flush) (struct file *, fl_owner_t id);
1572 int (*release) (struct inode *, struct file *);
1573 int (*fsync) (struct file *, int datasync);
1574 int (*aio_fsync) (struct kiocb *, int datasync);
1575 int (*fasync) (int, struct file *, int);
1576 int (*lock) (struct file *, int, struct file_lock *);
1577 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1578 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1579 int (*check_flags)(int);
1580 int (*flock) (struct file *, int, struct file_lock *);
1581 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1582 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
1583 int (*setlease)(struct file *, long, struct file_lock **);
1584 long (*fallocate)(struct file *file, int mode, loff_t offset,
1588 struct inode_operations {
1589 struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
1590 void * (*follow_link) (struct dentry *, struct nameidata *);
1591 int (*permission) (struct inode *, int);
1592 int (*check_acl)(struct inode *, int);
1594 int (*readlink) (struct dentry *, char __user *,int);
1595 void (*put_link) (struct dentry *, struct nameidata *, void *);
1597 int (*create) (struct inode *,struct dentry *,int, struct nameidata *);
1598 int (*link) (struct dentry *,struct inode *,struct dentry *);
1599 int (*unlink) (struct inode *,struct dentry *);
1600 int (*symlink) (struct inode *,struct dentry *,const char *);
1601 int (*mkdir) (struct inode *,struct dentry *,int);
1602 int (*rmdir) (struct inode *,struct dentry *);
1603 int (*mknod) (struct inode *,struct dentry *,int,dev_t);
1604 int (*rename) (struct inode *, struct dentry *,
1605 struct inode *, struct dentry *);
1606 void (*truncate) (struct inode *);
1607 int (*setattr) (struct dentry *, struct iattr *);
1608 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1609 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1610 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1611 ssize_t (*listxattr) (struct dentry *, char *, size_t);
1612 int (*removexattr) (struct dentry *, const char *);
1613 void (*truncate_range)(struct inode *, loff_t, loff_t);
1614 int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
1616 } ____cacheline_aligned;
1620 ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
1621 unsigned long nr_segs, unsigned long fast_segs,
1622 struct iovec *fast_pointer,
1623 struct iovec **ret_pointer);
1625 extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1626 extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1627 extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1628 unsigned long, loff_t *);
1629 extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1630 unsigned long, loff_t *);
1632 struct super_operations {
1633 struct inode *(*alloc_inode)(struct super_block *sb);
1634 void (*destroy_inode)(struct inode *);
1636 void (*dirty_inode) (struct inode *, int flags);
1637 int (*write_inode) (struct inode *, struct writeback_control *wbc);
1638 int (*drop_inode) (struct inode *);
1639 void (*evict_inode) (struct inode *);
1640 void (*put_super) (struct super_block *);
1641 void (*write_super) (struct super_block *);
1642 int (*sync_fs)(struct super_block *sb, int wait);
1643 int (*freeze_fs) (struct super_block *);
1644 int (*unfreeze_fs) (struct super_block *);
1645 int (*statfs) (struct dentry *, struct kstatfs *);
1646 int (*remount_fs) (struct super_block *, int *, char *);
1647 void (*umount_begin) (struct super_block *);
1649 int (*show_options)(struct seq_file *, struct vfsmount *);
1650 int (*show_devname)(struct seq_file *, struct vfsmount *);
1651 int (*show_path)(struct seq_file *, struct vfsmount *);
1652 int (*show_stats)(struct seq_file *, struct vfsmount *);
1654 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1655 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
1657 int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
1661 * Inode state bits. Protected by inode->i_lock
1663 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1664 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1666 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1667 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1668 * various stages of removing an inode.
1670 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1672 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1673 * fdatasync(). i_atime is the usual cause.
1674 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1675 * these changes separately from I_DIRTY_SYNC so that we
1676 * don't have to write inode on fdatasync() when only
1677 * mtime has changed in it.
1678 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1679 * I_NEW Serves as both a mutex and completion notification.
1680 * New inodes set I_NEW. If two processes both create
1681 * the same inode, one of them will release its inode and
1682 * wait for I_NEW to be released before returning.
1683 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1684 * also cause waiting on I_NEW, without I_NEW actually
1685 * being set. find_inode() uses this to prevent returning
1686 * nearly-dead inodes.
1687 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1688 * is zero. I_FREEING must be set when I_WILL_FREE is
1690 * I_FREEING Set when inode is about to be freed but still has dirty
1691 * pages or buffers attached or the inode itself is still
1693 * I_CLEAR Added by end_writeback(). In this state the inode is clean
1694 * and can be destroyed. Inode keeps I_FREEING.
1696 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1697 * prohibited for many purposes. iget() must wait for
1698 * the inode to be completely released, then create it
1699 * anew. Other functions will just ignore such inodes,
1700 * if appropriate. I_NEW is used for waiting.
1702 * I_SYNC Synchonized write of dirty inode data. The bits is
1703 * set during data writeback, and cleared with a wakeup
1704 * on the bit address once it is done.
1706 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1708 #define I_DIRTY_SYNC (1 << 0)
1709 #define I_DIRTY_DATASYNC (1 << 1)
1710 #define I_DIRTY_PAGES (1 << 2)
1712 #define I_NEW (1 << __I_NEW)
1713 #define I_WILL_FREE (1 << 4)
1714 #define I_FREEING (1 << 5)
1715 #define I_CLEAR (1 << 6)
1717 #define I_SYNC (1 << __I_SYNC)
1718 #define I_REFERENCED (1 << 8)
1720 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1722 extern void __mark_inode_dirty(struct inode *, int);
1723 static inline void mark_inode_dirty(struct inode *inode)
1725 __mark_inode_dirty(inode, I_DIRTY);
1728 static inline void mark_inode_dirty_sync(struct inode *inode)
1730 __mark_inode_dirty(inode, I_DIRTY_SYNC);
1734 * inc_nlink - directly increment an inode's link count
1737 * This is a low-level filesystem helper to replace any
1738 * direct filesystem manipulation of i_nlink. Currently,
1739 * it is only here for parity with dec_nlink().
1741 static inline void inc_nlink(struct inode *inode)
1746 static inline void inode_inc_link_count(struct inode *inode)
1749 mark_inode_dirty(inode);
1753 * drop_nlink - directly drop an inode's link count
1756 * This is a low-level filesystem helper to replace any
1757 * direct filesystem manipulation of i_nlink. In cases
1758 * where we are attempting to track writes to the
1759 * filesystem, a decrement to zero means an imminent
1760 * write when the file is truncated and actually unlinked
1761 * on the filesystem.
1763 static inline void drop_nlink(struct inode *inode)
1769 * clear_nlink - directly zero an inode's link count
1772 * This is a low-level filesystem helper to replace any
1773 * direct filesystem manipulation of i_nlink. See
1774 * drop_nlink() for why we care about i_nlink hitting zero.
1776 static inline void clear_nlink(struct inode *inode)
1781 static inline void inode_dec_link_count(struct inode *inode)
1784 mark_inode_dirty(inode);
1788 * inode_inc_iversion - increments i_version
1789 * @inode: inode that need to be updated
1791 * Every time the inode is modified, the i_version field will be incremented.
1792 * The filesystem has to be mounted with i_version flag
1795 static inline void inode_inc_iversion(struct inode *inode)
1797 spin_lock(&inode->i_lock);
1799 spin_unlock(&inode->i_lock);
1802 extern void touch_atime(struct vfsmount *mnt, struct dentry *dentry);
1803 static inline void file_accessed(struct file *file)
1805 if (!(file->f_flags & O_NOATIME))
1806 touch_atime(file->f_path.mnt, file->f_path.dentry);
1809 int sync_inode(struct inode *inode, struct writeback_control *wbc);
1810 int sync_inode_metadata(struct inode *inode, int wait);
1812 struct file_system_type {
1815 struct dentry *(*mount) (struct file_system_type *, int,
1816 const char *, void *);
1817 void (*kill_sb) (struct super_block *);
1818 struct module *owner;
1819 struct file_system_type * next;
1820 struct list_head fs_supers;
1822 struct lock_class_key s_lock_key;
1823 struct lock_class_key s_umount_key;
1824 struct lock_class_key s_vfs_rename_key;
1826 struct lock_class_key i_lock_key;
1827 struct lock_class_key i_mutex_key;
1828 struct lock_class_key i_mutex_dir_key;
1829 struct lock_class_key i_alloc_sem_key;
1832 extern struct dentry *mount_ns(struct file_system_type *fs_type, int flags,
1833 void *data, int (*fill_super)(struct super_block *, void *, int));
1834 extern struct dentry *mount_bdev(struct file_system_type *fs_type,
1835 int flags, const char *dev_name, void *data,
1836 int (*fill_super)(struct super_block *, void *, int));
1837 extern struct dentry *mount_single(struct file_system_type *fs_type,
1838 int flags, void *data,
1839 int (*fill_super)(struct super_block *, void *, int));
1840 extern struct dentry *mount_nodev(struct file_system_type *fs_type,
1841 int flags, void *data,
1842 int (*fill_super)(struct super_block *, void *, int));
1843 void generic_shutdown_super(struct super_block *sb);
1844 void kill_block_super(struct super_block *sb);
1845 void kill_anon_super(struct super_block *sb);
1846 void kill_litter_super(struct super_block *sb);
1847 void deactivate_super(struct super_block *sb);
1848 void deactivate_locked_super(struct super_block *sb);
1849 int set_anon_super(struct super_block *s, void *data);
1850 int get_anon_bdev(dev_t *);
1851 void free_anon_bdev(dev_t);
1852 struct super_block *sget(struct file_system_type *type,
1853 int (*test)(struct super_block *,void *),
1854 int (*set)(struct super_block *,void *),
1856 extern struct dentry *mount_pseudo(struct file_system_type *, char *,
1857 const struct super_operations *ops,
1858 const struct dentry_operations *dops,
1861 static inline void sb_mark_dirty(struct super_block *sb)
1865 static inline void sb_mark_clean(struct super_block *sb)
1869 static inline int sb_is_dirty(struct super_block *sb)
1874 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1875 #define fops_get(fops) \
1876 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1877 #define fops_put(fops) \
1878 do { if (fops) module_put((fops)->owner); } while(0)
1880 extern int register_filesystem(struct file_system_type *);
1881 extern int unregister_filesystem(struct file_system_type *);
1882 extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
1883 #define kern_mount(type) kern_mount_data(type, NULL)
1884 extern int may_umount_tree(struct vfsmount *);
1885 extern int may_umount(struct vfsmount *);
1886 extern long do_mount(char *, char *, char *, unsigned long, void *);
1887 extern struct vfsmount *collect_mounts(struct path *);
1888 extern void drop_collected_mounts(struct vfsmount *);
1889 extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
1891 extern int vfs_statfs(struct path *, struct kstatfs *);
1892 extern int user_statfs(const char __user *, struct kstatfs *);
1893 extern int fd_statfs(int, struct kstatfs *);
1894 extern int statfs_by_dentry(struct dentry *, struct kstatfs *);
1895 extern int freeze_super(struct super_block *super);
1896 extern int thaw_super(struct super_block *super);
1898 extern int current_umask(void);
1901 extern struct kobject *fs_kobj;
1903 #define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
1904 extern int rw_verify_area(int, struct file *, loff_t *, size_t);
1906 #define FLOCK_VERIFY_READ 1
1907 #define FLOCK_VERIFY_WRITE 2
1909 #ifdef CONFIG_FILE_LOCKING
1910 extern int locks_mandatory_locked(struct inode *);
1911 extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
1914 * Candidates for mandatory locking have the setgid bit set
1915 * but no group execute bit - an otherwise meaningless combination.
1918 static inline int __mandatory_lock(struct inode *ino)
1920 return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
1924 * ... and these candidates should be on MS_MANDLOCK mounted fs,
1925 * otherwise these will be advisory locks
1928 static inline int mandatory_lock(struct inode *ino)
1930 return IS_MANDLOCK(ino) && __mandatory_lock(ino);
1933 static inline int locks_verify_locked(struct inode *inode)
1935 if (mandatory_lock(inode))
1936 return locks_mandatory_locked(inode);
1940 static inline int locks_verify_truncate(struct inode *inode,
1944 if (inode->i_flock && mandatory_lock(inode))
1945 return locks_mandatory_area(
1946 FLOCK_VERIFY_WRITE, inode, filp,
1947 size < inode->i_size ? size : inode->i_size,
1948 (size < inode->i_size ? inode->i_size - size
1949 : size - inode->i_size)
1954 static inline int break_lease(struct inode *inode, unsigned int mode)
1957 return __break_lease(inode, mode);
1960 #else /* !CONFIG_FILE_LOCKING */
1961 static inline int locks_mandatory_locked(struct inode *inode)
1966 static inline int locks_mandatory_area(int rw, struct inode *inode,
1967 struct file *filp, loff_t offset,
1973 static inline int __mandatory_lock(struct inode *inode)
1978 static inline int mandatory_lock(struct inode *inode)
1983 static inline int locks_verify_locked(struct inode *inode)
1988 static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
1994 static inline int break_lease(struct inode *inode, unsigned int mode)
1999 #endif /* CONFIG_FILE_LOCKING */
2003 extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
2005 extern int do_fallocate(struct file *file, int mode, loff_t offset,
2007 extern long do_sys_open(int dfd, const char __user *filename, int flags,
2009 extern struct file *filp_open(const char *, int, int);
2010 extern struct file *file_open_root(struct dentry *, struct vfsmount *,
2012 extern struct file * dentry_open(struct dentry *, struct vfsmount *, int,
2013 const struct cred *);
2014 extern int filp_close(struct file *, fl_owner_t id);
2015 extern char * getname(const char __user *);
2019 extern int ioctl_preallocate(struct file *filp, void __user *argp);
2022 extern void __init vfs_caches_init_early(void);
2023 extern void __init vfs_caches_init(unsigned long);
2025 extern struct kmem_cache *names_cachep;
2027 #define __getname_gfp(gfp) kmem_cache_alloc(names_cachep, (gfp))
2028 #define __getname() __getname_gfp(GFP_KERNEL)
2029 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
2030 #ifndef CONFIG_AUDITSYSCALL
2031 #define putname(name) __putname(name)
2033 extern void putname(const char *name);
2037 extern int register_blkdev(unsigned int, const char *);
2038 extern void unregister_blkdev(unsigned int, const char *);
2039 extern struct block_device *bdget(dev_t);
2040 extern struct block_device *bdgrab(struct block_device *bdev);
2041 extern void bd_set_size(struct block_device *, loff_t size);
2042 extern void bd_forget(struct inode *inode);
2043 extern void bdput(struct block_device *);
2044 extern void invalidate_bdev(struct block_device *);
2045 extern int sync_blockdev(struct block_device *bdev);
2046 extern struct super_block *freeze_bdev(struct block_device *);
2047 extern void emergency_thaw_all(void);
2048 extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
2049 extern int fsync_bdev(struct block_device *);
2051 static inline void bd_forget(struct inode *inode) {}
2052 static inline int sync_blockdev(struct block_device *bdev) { return 0; }
2053 static inline void invalidate_bdev(struct block_device *bdev) {}
2055 static inline struct super_block *freeze_bdev(struct block_device *sb)
2060 static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
2065 extern int sync_filesystem(struct super_block *);
2066 extern const struct file_operations def_blk_fops;
2067 extern const struct file_operations def_chr_fops;
2068 extern const struct file_operations bad_sock_fops;
2069 extern const struct file_operations def_fifo_fops;
2071 extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
2072 extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
2073 extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
2074 extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
2075 extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
2077 extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
2079 extern int blkdev_put(struct block_device *bdev, fmode_t mode);
2081 extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
2082 extern void bd_unlink_disk_holder(struct block_device *bdev,
2083 struct gendisk *disk);
2085 static inline int bd_link_disk_holder(struct block_device *bdev,
2086 struct gendisk *disk)
2090 static inline void bd_unlink_disk_holder(struct block_device *bdev,
2091 struct gendisk *disk)
2098 #define CHRDEV_MAJOR_HASH_SIZE 255
2099 extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
2100 extern int register_chrdev_region(dev_t, unsigned, const char *);
2101 extern int __register_chrdev(unsigned int major, unsigned int baseminor,
2102 unsigned int count, const char *name,
2103 const struct file_operations *fops);
2104 extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
2105 unsigned int count, const char *name);
2106 extern void unregister_chrdev_region(dev_t, unsigned);
2107 extern void chrdev_show(struct seq_file *,off_t);
2109 static inline int register_chrdev(unsigned int major, const char *name,
2110 const struct file_operations *fops)
2112 return __register_chrdev(major, 0, 256, name, fops);
2115 static inline void unregister_chrdev(unsigned int major, const char *name)
2117 __unregister_chrdev(major, 0, 256, name);
2120 /* fs/block_dev.c */
2121 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
2122 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
2125 #define BLKDEV_MAJOR_HASH_SIZE 255
2126 extern const char *__bdevname(dev_t, char *buffer);
2127 extern const char *bdevname(struct block_device *bdev, char *buffer);
2128 extern struct block_device *lookup_bdev(const char *);
2129 extern void blkdev_show(struct seq_file *,off_t);
2132 #define BLKDEV_MAJOR_HASH_SIZE 0
2135 extern void init_special_inode(struct inode *, umode_t, dev_t);
2137 /* Invalid inode operations -- fs/bad_inode.c */
2138 extern void make_bad_inode(struct inode *);
2139 extern int is_bad_inode(struct inode *);
2141 extern const struct file_operations read_pipefifo_fops;
2142 extern const struct file_operations write_pipefifo_fops;
2143 extern const struct file_operations rdwr_pipefifo_fops;
2145 extern int fs_may_remount_ro(struct super_block *);
2149 * return READ, READA, or WRITE
2151 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2154 * return data direction, READ or WRITE
2156 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
2158 extern void check_disk_size_change(struct gendisk *disk,
2159 struct block_device *bdev);
2160 extern int revalidate_disk(struct gendisk *);
2161 extern int check_disk_change(struct block_device *);
2162 extern int __invalidate_device(struct block_device *, bool);
2163 extern int invalidate_partition(struct gendisk *, int);
2165 unsigned long invalidate_mapping_pages(struct address_space *mapping,
2166 pgoff_t start, pgoff_t end);
2168 static inline void invalidate_remote_inode(struct inode *inode)
2170 if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
2171 S_ISLNK(inode->i_mode))
2172 invalidate_mapping_pages(inode->i_mapping, 0, -1);
2174 extern int invalidate_inode_pages2(struct address_space *mapping);
2175 extern int invalidate_inode_pages2_range(struct address_space *mapping,
2176 pgoff_t start, pgoff_t end);
2177 extern int write_inode_now(struct inode *, int);
2178 extern int filemap_fdatawrite(struct address_space *);
2179 extern int filemap_flush(struct address_space *);
2180 extern int filemap_fdatawait(struct address_space *);
2181 extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
2183 extern int filemap_write_and_wait(struct address_space *mapping);
2184 extern int filemap_write_and_wait_range(struct address_space *mapping,
2185 loff_t lstart, loff_t lend);
2186 extern int __filemap_fdatawrite_range(struct address_space *mapping,
2187 loff_t start, loff_t end, int sync_mode);
2188 extern int filemap_fdatawrite_range(struct address_space *mapping,
2189 loff_t start, loff_t end);
2191 extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
2193 extern int vfs_fsync(struct file *file, int datasync);
2194 extern int generic_write_sync(struct file *file, loff_t pos, loff_t count);
2195 extern void sync_supers(void);
2196 extern void emergency_sync(void);
2197 extern void emergency_remount(void);
2199 extern sector_t bmap(struct inode *, sector_t);
2201 extern int notify_change(struct dentry *, struct iattr *);
2202 extern int inode_permission(struct inode *, int);
2203 extern int generic_permission(struct inode *, int);
2205 static inline bool execute_ok(struct inode *inode)
2207 return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
2211 * get_write_access() gets write permission for a file.
2212 * put_write_access() releases this write permission.
2213 * This is used for regular files.
2214 * We cannot support write (and maybe mmap read-write shared) accesses and
2215 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2216 * can have the following values:
2217 * 0: no writers, no VM_DENYWRITE mappings
2218 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2219 * > 0: (i_writecount) users are writing to the file.
2221 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2222 * except for the cases where we don't hold i_writecount yet. Then we need to
2223 * use {get,deny}_write_access() - these functions check the sign and refuse
2224 * to do the change if sign is wrong.
2226 static inline int get_write_access(struct inode *inode)
2228 return atomic_inc_unless_negative(&inode->i_writecount) ? 0 : -ETXTBSY;
2230 static inline int deny_write_access(struct file *file)
2232 struct inode *inode = file->f_path.dentry->d_inode;
2233 return atomic_dec_unless_positive(&inode->i_writecount) ? 0 : -ETXTBSY;
2235 static inline void put_write_access(struct inode * inode)
2237 atomic_dec(&inode->i_writecount);
2239 static inline void allow_write_access(struct file *file)
2242 atomic_inc(&file->f_path.dentry->d_inode->i_writecount);
2245 static inline void i_readcount_dec(struct inode *inode)
2247 BUG_ON(!atomic_read(&inode->i_readcount));
2248 atomic_dec(&inode->i_readcount);
2250 static inline void i_readcount_inc(struct inode *inode)
2252 atomic_inc(&inode->i_readcount);
2255 static inline void i_readcount_dec(struct inode *inode)
2259 static inline void i_readcount_inc(struct inode *inode)
2264 extern int do_pipe_flags(int *, int);
2265 extern struct file *create_read_pipe(struct file *f, int flags);
2266 extern struct file *create_write_pipe(int flags);
2267 extern void free_write_pipe(struct file *);
2269 extern int kernel_read(struct file *, loff_t, char *, unsigned long);
2270 extern struct file * open_exec(const char *);
2272 /* fs/dcache.c -- generic fs support functions */
2273 extern int is_subdir(struct dentry *, struct dentry *);
2274 extern int path_is_under(struct path *, struct path *);
2275 extern ino_t find_inode_number(struct dentry *, struct qstr *);
2277 #include <linux/err.h>
2279 /* needed for stackable file system support */
2280 extern loff_t default_llseek(struct file *file, loff_t offset, int origin);
2282 extern loff_t vfs_llseek(struct file *file, loff_t offset, int origin);
2284 extern int inode_init_always(struct super_block *, struct inode *);
2285 extern void inode_init_once(struct inode *);
2286 extern void address_space_init_once(struct address_space *mapping);
2287 extern void ihold(struct inode * inode);
2288 extern void iput(struct inode *);
2289 extern struct inode * igrab(struct inode *);
2290 extern ino_t iunique(struct super_block *, ino_t);
2291 extern int inode_needs_sync(struct inode *inode);
2292 extern int generic_delete_inode(struct inode *inode);
2293 extern int generic_drop_inode(struct inode *inode);
2295 extern struct inode *ilookup5_nowait(struct super_block *sb,
2296 unsigned long hashval, int (*test)(struct inode *, void *),
2298 extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
2299 int (*test)(struct inode *, void *), void *data);
2300 extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
2302 extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
2303 extern struct inode * iget_locked(struct super_block *, unsigned long);
2304 extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
2305 extern int insert_inode_locked(struct inode *);
2306 extern void unlock_new_inode(struct inode *);
2307 extern unsigned int get_next_ino(void);
2309 extern void __iget(struct inode * inode);
2310 extern void iget_failed(struct inode *);
2311 extern void end_writeback(struct inode *);
2312 extern void __destroy_inode(struct inode *);
2313 extern struct inode *new_inode(struct super_block *);
2314 extern void free_inode_nonrcu(struct inode *inode);
2315 extern int should_remove_suid(struct dentry *);
2316 extern int file_remove_suid(struct file *);
2318 extern void __insert_inode_hash(struct inode *, unsigned long hashval);
2319 extern void remove_inode_hash(struct inode *);
2320 static inline void insert_inode_hash(struct inode *inode)
2322 __insert_inode_hash(inode, inode->i_ino);
2324 extern void inode_sb_list_add(struct inode *inode);
2327 extern void submit_bio(int, struct bio *);
2328 extern int bdev_read_only(struct block_device *);
2330 extern int set_blocksize(struct block_device *, int);
2331 extern int sb_set_blocksize(struct super_block *, int);
2332 extern int sb_min_blocksize(struct super_block *, int);
2334 extern int generic_file_mmap(struct file *, struct vm_area_struct *);
2335 extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
2336 extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
2337 int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
2338 extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2339 extern ssize_t __generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long,
2341 extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2342 extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
2343 unsigned long *, loff_t, loff_t *, size_t, size_t);
2344 extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
2345 unsigned long, loff_t, loff_t *, size_t, ssize_t);
2346 extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
2347 extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
2348 extern int generic_segment_checks(const struct iovec *iov,
2349 unsigned long *nr_segs, size_t *count, int access_flags);
2351 /* fs/block_dev.c */
2352 extern ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
2353 unsigned long nr_segs, loff_t pos);
2354 extern int blkdev_fsync(struct file *filp, int datasync);
2357 extern ssize_t generic_file_splice_read(struct file *, loff_t *,
2358 struct pipe_inode_info *, size_t, unsigned int);
2359 extern ssize_t default_file_splice_read(struct file *, loff_t *,
2360 struct pipe_inode_info *, size_t, unsigned int);
2361 extern ssize_t generic_file_splice_write(struct pipe_inode_info *,
2362 struct file *, loff_t *, size_t, unsigned int);
2363 extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
2364 struct file *out, loff_t *, size_t len, unsigned int flags);
2365 extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
2366 size_t len, unsigned int flags);
2369 file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
2370 extern loff_t noop_llseek(struct file *file, loff_t offset, int origin);
2371 extern loff_t no_llseek(struct file *file, loff_t offset, int origin);
2372 extern loff_t generic_file_llseek(struct file *file, loff_t offset, int origin);
2373 extern loff_t generic_file_llseek_unlocked(struct file *file, loff_t offset,
2375 extern int generic_file_open(struct inode * inode, struct file * filp);
2376 extern int nonseekable_open(struct inode * inode, struct file * filp);
2378 #ifdef CONFIG_FS_XIP
2379 extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
2381 extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
2382 extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
2383 size_t len, loff_t *ppos);
2384 extern int xip_truncate_page(struct address_space *mapping, loff_t from);
2386 static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
2393 typedef void (dio_submit_t)(int rw, struct bio *bio, struct inode *inode,
2394 loff_t file_offset);
2397 /* need locking between buffered and direct access */
2400 /* filesystem does not support filling holes */
2401 DIO_SKIP_HOLES = 0x02,
2404 void dio_end_io(struct bio *bio, int error);
2406 ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
2407 struct block_device *bdev, const struct iovec *iov, loff_t offset,
2408 unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
2409 dio_submit_t submit_io, int flags);
2411 static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
2412 struct inode *inode, struct block_device *bdev, const struct iovec *iov,
2413 loff_t offset, unsigned long nr_segs, get_block_t get_block,
2414 dio_iodone_t end_io)
2416 return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset,
2417 nr_segs, get_block, end_io, NULL,
2418 DIO_LOCKING | DIO_SKIP_HOLES);
2422 extern const struct file_operations generic_ro_fops;
2424 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2426 extern int vfs_readlink(struct dentry *, char __user *, int, const char *);
2427 extern int vfs_follow_link(struct nameidata *, const char *);
2428 extern int page_readlink(struct dentry *, char __user *, int);
2429 extern void *page_follow_link_light(struct dentry *, struct nameidata *);
2430 extern void page_put_link(struct dentry *, struct nameidata *, void *);
2431 extern int __page_symlink(struct inode *inode, const char *symname, int len,
2433 extern int page_symlink(struct inode *inode, const char *symname, int len);
2434 extern const struct inode_operations page_symlink_inode_operations;
2435 extern int generic_readlink(struct dentry *, char __user *, int);
2436 extern void generic_fillattr(struct inode *, struct kstat *);
2437 extern int vfs_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2438 void __inode_add_bytes(struct inode *inode, loff_t bytes);
2439 void inode_add_bytes(struct inode *inode, loff_t bytes);
2440 void inode_sub_bytes(struct inode *inode, loff_t bytes);
2441 loff_t inode_get_bytes(struct inode *inode);
2442 void inode_set_bytes(struct inode *inode, loff_t bytes);
2444 extern int vfs_readdir(struct file *, filldir_t, void *);
2446 extern int vfs_stat(const char __user *, struct kstat *);
2447 extern int vfs_lstat(const char __user *, struct kstat *);
2448 extern int vfs_fstat(unsigned int, struct kstat *);
2449 extern int vfs_fstatat(int , const char __user *, struct kstat *, int);
2451 extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
2453 extern int __generic_block_fiemap(struct inode *inode,
2454 struct fiemap_extent_info *fieinfo,
2455 loff_t start, loff_t len,
2456 get_block_t *get_block);
2457 extern int generic_block_fiemap(struct inode *inode,
2458 struct fiemap_extent_info *fieinfo, u64 start,
2459 u64 len, get_block_t *get_block);
2461 extern void get_filesystem(struct file_system_type *fs);
2462 extern void put_filesystem(struct file_system_type *fs);
2463 extern struct file_system_type *get_fs_type(const char *name);
2464 extern struct super_block *get_super(struct block_device *);
2465 extern struct super_block *get_active_super(struct block_device *bdev);
2466 extern struct super_block *user_get_super(dev_t);
2467 extern void drop_super(struct super_block *sb);
2468 extern void iterate_supers(void (*)(struct super_block *, void *), void *);
2469 extern void iterate_supers_type(struct file_system_type *,
2470 void (*)(struct super_block *, void *), void *);
2472 extern int dcache_dir_open(struct inode *, struct file *);
2473 extern int dcache_dir_close(struct inode *, struct file *);
2474 extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
2475 extern int dcache_readdir(struct file *, void *, filldir_t);
2476 extern int simple_setattr(struct dentry *, struct iattr *);
2477 extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2478 extern int simple_statfs(struct dentry *, struct kstatfs *);
2479 extern int simple_link(struct dentry *, struct inode *, struct dentry *);
2480 extern int simple_unlink(struct inode *, struct dentry *);
2481 extern int simple_rmdir(struct inode *, struct dentry *);
2482 extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
2483 extern int noop_fsync(struct file *, int);
2484 extern int simple_empty(struct dentry *);
2485 extern int simple_readpage(struct file *file, struct page *page);
2486 extern int simple_write_begin(struct file *file, struct address_space *mapping,
2487 loff_t pos, unsigned len, unsigned flags,
2488 struct page **pagep, void **fsdata);
2489 extern int simple_write_end(struct file *file, struct address_space *mapping,
2490 loff_t pos, unsigned len, unsigned copied,
2491 struct page *page, void *fsdata);
2493 extern struct dentry *simple_lookup(struct inode *, struct dentry *, struct nameidata *);
2494 extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
2495 extern const struct file_operations simple_dir_operations;
2496 extern const struct inode_operations simple_dir_inode_operations;
2497 struct tree_descr { char *name; const struct file_operations *ops; int mode; };
2498 struct dentry *d_alloc_name(struct dentry *, const char *);
2499 extern int simple_fill_super(struct super_block *, unsigned long, struct tree_descr *);
2500 extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
2501 extern void simple_release_fs(struct vfsmount **mount, int *count);
2503 extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
2504 loff_t *ppos, const void *from, size_t available);
2505 extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
2506 const void __user *from, size_t count);
2508 extern int generic_file_fsync(struct file *, int);
2510 extern int generic_check_addressable(unsigned, u64);
2512 #ifdef CONFIG_MIGRATION
2513 extern int buffer_migrate_page(struct address_space *,
2514 struct page *, struct page *);
2516 #define buffer_migrate_page NULL
2519 extern int inode_change_ok(const struct inode *, struct iattr *);
2520 extern int inode_newsize_ok(const struct inode *, loff_t offset);
2521 extern void setattr_copy(struct inode *inode, const struct iattr *attr);
2523 extern void file_update_time(struct file *file);
2525 extern int generic_show_options(struct seq_file *m, struct vfsmount *mnt);
2526 extern void save_mount_options(struct super_block *sb, char *options);
2527 extern void replace_mount_options(struct super_block *sb, char *options);
2529 static inline ino_t parent_ino(struct dentry *dentry)
2534 * Don't strictly need d_lock here? If the parent ino could change
2535 * then surely we'd have a deeper race in the caller?
2537 spin_lock(&dentry->d_lock);
2538 res = dentry->d_parent->d_inode->i_ino;
2539 spin_unlock(&dentry->d_lock);
2543 /* Transaction based IO helpers */
2546 * An argresp is stored in an allocated page and holds the
2547 * size of the argument or response, along with its content
2549 struct simple_transaction_argresp {
2554 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2556 char *simple_transaction_get(struct file *file, const char __user *buf,
2558 ssize_t simple_transaction_read(struct file *file, char __user *buf,
2559 size_t size, loff_t *pos);
2560 int simple_transaction_release(struct inode *inode, struct file *file);
2562 void simple_transaction_set(struct file *file, size_t n);
2565 * simple attribute files
2567 * These attributes behave similar to those in sysfs:
2569 * Writing to an attribute immediately sets a value, an open file can be
2570 * written to multiple times.
2572 * Reading from an attribute creates a buffer from the value that might get
2573 * read with multiple read calls. When the attribute has been read
2574 * completely, no further read calls are possible until the file is opened
2577 * All attributes contain a text representation of a numeric value
2578 * that are accessed with the get() and set() functions.
2580 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2581 static int __fops ## _open(struct inode *inode, struct file *file) \
2583 __simple_attr_check_format(__fmt, 0ull); \
2584 return simple_attr_open(inode, file, __get, __set, __fmt); \
2586 static const struct file_operations __fops = { \
2587 .owner = THIS_MODULE, \
2588 .open = __fops ## _open, \
2589 .release = simple_attr_release, \
2590 .read = simple_attr_read, \
2591 .write = simple_attr_write, \
2592 .llseek = generic_file_llseek, \
2595 static inline void __attribute__((format(printf, 1, 2)))
2596 __simple_attr_check_format(const char *fmt, ...)
2598 /* don't do anything, just let the compiler check the arguments; */
2601 int simple_attr_open(struct inode *inode, struct file *file,
2602 int (*get)(void *, u64 *), int (*set)(void *, u64),
2604 int simple_attr_release(struct inode *inode, struct file *file);
2605 ssize_t simple_attr_read(struct file *file, char __user *buf,
2606 size_t len, loff_t *ppos);
2607 ssize_t simple_attr_write(struct file *file, const char __user *buf,
2608 size_t len, loff_t *ppos);
2611 int proc_nr_files(struct ctl_table *table, int write,
2612 void __user *buffer, size_t *lenp, loff_t *ppos);
2613 int proc_nr_dentry(struct ctl_table *table, int write,
2614 void __user *buffer, size_t *lenp, loff_t *ppos);
2615 int proc_nr_inodes(struct ctl_table *table, int write,
2616 void __user *buffer, size_t *lenp, loff_t *ppos);
2617 int __init get_filesystem_list(char *buf);
2619 #define __FMODE_EXEC ((__force int) FMODE_EXEC)
2620 #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
2622 #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
2623 #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
2624 (flag & __FMODE_NONOTIFY)))
2626 static inline int is_sxid(mode_t mode)
2628 return (mode & S_ISUID) || ((mode & S_ISGID) && (mode & S_IXGRP));
2631 static inline void inode_has_no_xattr(struct inode *inode)
2633 if (!is_sxid(inode->i_mode) && (inode->i_sb->s_flags & MS_NOSEC))
2634 inode->i_flags |= S_NOSEC;
2637 #endif /* __KERNEL__ */
2638 #endif /* _LINUX_FS_H */