5 * This file has definitions for some important file table
9 #include <linux/limits.h>
10 #include <linux/ioctl.h>
13 * It's silly to have NR_OPEN bigger than NR_FILE, but you can change
14 * the file limit at runtime and only root can increase the per-process
15 * nr_file rlimit, so it's safe to set up a ridiculously high absolute
16 * upper limit on files-per-process.
18 * Some programs (notably those using select()) may have to be
19 * recompiled to take full advantage of the new limits..
22 /* Fixed constants first: */
24 #define INR_OPEN 1024 /* Initial setting for nfile rlimits */
26 #define BLOCK_SIZE_BITS 10
27 #define BLOCK_SIZE (1<<BLOCK_SIZE_BITS)
29 #define SEEK_SET 0 /* seek relative to beginning of file */
30 #define SEEK_CUR 1 /* seek relative to current file position */
31 #define SEEK_END 2 /* seek relative to end of file */
32 #define SEEK_MAX SEEK_END
34 /* And dynamically-tunable limits and defaults: */
35 struct files_stat_struct {
36 int nr_files; /* read only */
37 int nr_free_files; /* read only */
38 int max_files; /* tunable */
41 struct inodes_stat_t {
44 int dummy[5]; /* padding for sysctl ABI compatibility */
48 #define NR_FILE 8192 /* this can well be larger on a larger system */
58 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
59 * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
62 /* file is open for reading */
63 #define FMODE_READ ((__force fmode_t)1)
64 /* file is open for writing */
65 #define FMODE_WRITE ((__force fmode_t)2)
66 /* file is seekable */
67 #define FMODE_LSEEK ((__force fmode_t)4)
68 /* file can be accessed using pread */
69 #define FMODE_PREAD ((__force fmode_t)8)
70 /* file can be accessed using pwrite */
71 #define FMODE_PWRITE ((__force fmode_t)16)
72 /* File is opened for execution with sys_execve / sys_uselib */
73 #define FMODE_EXEC ((__force fmode_t)32)
74 /* File is opened with O_NDELAY (only set for block devices) */
75 #define FMODE_NDELAY ((__force fmode_t)64)
76 /* File is opened with O_EXCL (only set for block devices) */
77 #define FMODE_EXCL ((__force fmode_t)128)
78 /* File is opened using open(.., 3, ..) and is writeable only for ioctls
79 (specialy hack for floppy.c) */
80 #define FMODE_WRITE_IOCTL ((__force fmode_t)256)
83 * Don't update ctime and mtime.
85 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
86 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
88 #define FMODE_NOCMTIME ((__force fmode_t)2048)
94 #define READA 2 /* read-ahead - don't block if no resources */
95 #define SWRITE 3 /* for ll_rw_block() - wait for buffer lock */
96 #define READ_SYNC (READ | (1 << BIO_RW_SYNCIO) | (1 << BIO_RW_UNPLUG))
97 #define READ_META (READ | (1 << BIO_RW_META))
98 #define WRITE_SYNC (WRITE | (1 << BIO_RW_SYNCIO) | (1 << BIO_RW_UNPLUG))
99 #define SWRITE_SYNC (SWRITE | (1 << BIO_RW_SYNCIO) | (1 << BIO_RW_UNPLUG))
100 #define WRITE_BARRIER (WRITE | (1 << BIO_RW_BARRIER))
101 #define DISCARD_NOBARRIER (1 << BIO_RW_DISCARD)
102 #define DISCARD_BARRIER ((1 << BIO_RW_DISCARD) | (1 << BIO_RW_BARRIER))
108 /* public flags for file_system_type */
109 #define FS_REQUIRES_DEV 1
110 #define FS_BINARY_MOUNTDATA 2
111 #define FS_HAS_SUBTYPE 4
112 #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
113 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move()
114 * during rename() internally.
118 * These are the fs-independent mount-flags: up to 32 flags are supported
120 #define MS_RDONLY 1 /* Mount read-only */
121 #define MS_NOSUID 2 /* Ignore suid and sgid bits */
122 #define MS_NODEV 4 /* Disallow access to device special files */
123 #define MS_NOEXEC 8 /* Disallow program execution */
124 #define MS_SYNCHRONOUS 16 /* Writes are synced at once */
125 #define MS_REMOUNT 32 /* Alter flags of a mounted FS */
126 #define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
127 #define MS_DIRSYNC 128 /* Directory modifications are synchronous */
128 #define MS_NOATIME 1024 /* Do not update access times. */
129 #define MS_NODIRATIME 2048 /* Do not update directory access times */
133 #define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
134 MS_VERBOSE is deprecated. */
135 #define MS_SILENT 32768
136 #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
137 #define MS_UNBINDABLE (1<<17) /* change to unbindable */
138 #define MS_PRIVATE (1<<18) /* change to private */
139 #define MS_SLAVE (1<<19) /* change to slave */
140 #define MS_SHARED (1<<20) /* change to shared */
141 #define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */
142 #define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
143 #define MS_I_VERSION (1<<23) /* Update inode I_version field */
144 #define MS_STRICTATIME (1<<24) /* Always perform atime updates */
145 #define MS_ACTIVE (1<<30)
146 #define MS_NOUSER (1<<31)
149 * Superblock flags that can be altered by MS_REMOUNT
151 #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK|MS_I_VERSION)
154 * Old magic mount flag and mask
156 #define MS_MGC_VAL 0xC0ED0000
157 #define MS_MGC_MSK 0xffff0000
159 /* Inode flags - they have nothing to superblock flags now */
161 #define S_SYNC 1 /* Writes are synced at once */
162 #define S_NOATIME 2 /* Do not update access times */
163 #define S_APPEND 4 /* Append-only file */
164 #define S_IMMUTABLE 8 /* Immutable file */
165 #define S_DEAD 16 /* removed, but still open directory */
166 #define S_NOQUOTA 32 /* Inode is not counted to quota */
167 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
168 #define S_NOCMTIME 128 /* Do not update file c/mtime */
169 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
170 #define S_PRIVATE 512 /* Inode is fs-internal */
173 * Note that nosuid etc flags are inode-specific: setting some file-system
174 * flags just means all the inodes inherit those flags by default. It might be
175 * possible to override it selectively if you really wanted to with some
176 * ioctl() that is not currently implemented.
178 * Exception: MS_RDONLY is always applied to the entire file system.
180 * Unfortunately, it is possible to change a filesystems flags with it mounted
181 * with files in use. This means that all of the inodes will not have their
182 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
183 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
185 #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
187 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
188 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
189 ((inode)->i_flags & S_SYNC))
190 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
191 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
192 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
193 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
194 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
196 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
197 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
198 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
199 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
201 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
202 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
203 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
204 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
206 /* the read-only stuff doesn't really belong here, but any other place is
207 probably as bad and I don't want to create yet another include file. */
209 #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
210 #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
211 #define BLKRRPART _IO(0x12,95) /* re-read partition table */
212 #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
213 #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
214 #define BLKRASET _IO(0x12,98) /* set read ahead for block device */
215 #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
216 #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
217 #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
218 #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
219 #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
220 #define BLKSSZGET _IO(0x12,104)/* get block device sector size */
222 #define BLKPG _IO(0x12,105)/* See blkpg.h */
224 /* Some people are morons. Do not use sizeof! */
226 #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
227 #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
228 /* This was here just to show that the number is taken -
229 probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
231 /* A jump here: 108-111 have been used for various private purposes. */
232 #define BLKBSZGET _IOR(0x12,112,size_t)
233 #define BLKBSZSET _IOW(0x12,113,size_t)
234 #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
235 #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
236 #define BLKTRACESTART _IO(0x12,116)
237 #define BLKTRACESTOP _IO(0x12,117)
238 #define BLKTRACETEARDOWN _IO(0x12,118)
239 #define BLKDISCARD _IO(0x12,119)
241 #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
242 #define FIBMAP _IO(0x00,1) /* bmap access */
243 #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
244 #define FIFREEZE _IOWR('X', 119, int) /* Freeze */
245 #define FITHAW _IOWR('X', 120, int) /* Thaw */
247 #define FS_IOC_GETFLAGS _IOR('f', 1, long)
248 #define FS_IOC_SETFLAGS _IOW('f', 2, long)
249 #define FS_IOC_GETVERSION _IOR('v', 1, long)
250 #define FS_IOC_SETVERSION _IOW('v', 2, long)
251 #define FS_IOC_FIEMAP _IOWR('f', 11, struct fiemap)
252 #define FS_IOC32_GETFLAGS _IOR('f', 1, int)
253 #define FS_IOC32_SETFLAGS _IOW('f', 2, int)
254 #define FS_IOC32_GETVERSION _IOR('v', 1, int)
255 #define FS_IOC32_SETVERSION _IOW('v', 2, int)
258 * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS)
260 #define FS_SECRM_FL 0x00000001 /* Secure deletion */
261 #define FS_UNRM_FL 0x00000002 /* Undelete */
262 #define FS_COMPR_FL 0x00000004 /* Compress file */
263 #define FS_SYNC_FL 0x00000008 /* Synchronous updates */
264 #define FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
265 #define FS_APPEND_FL 0x00000020 /* writes to file may only append */
266 #define FS_NODUMP_FL 0x00000040 /* do not dump file */
267 #define FS_NOATIME_FL 0x00000080 /* do not update atime */
268 /* Reserved for compression usage... */
269 #define FS_DIRTY_FL 0x00000100
270 #define FS_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
271 #define FS_NOCOMP_FL 0x00000400 /* Don't compress */
272 #define FS_ECOMPR_FL 0x00000800 /* Compression error */
273 /* End compression flags --- maybe not all used */
274 #define FS_BTREE_FL 0x00001000 /* btree format dir */
275 #define FS_INDEX_FL 0x00001000 /* hash-indexed directory */
276 #define FS_IMAGIC_FL 0x00002000 /* AFS directory */
277 #define FS_JOURNAL_DATA_FL 0x00004000 /* Reserved for ext3 */
278 #define FS_NOTAIL_FL 0x00008000 /* file tail should not be merged */
279 #define FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
280 #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
281 #define FS_EXTENT_FL 0x00080000 /* Extents */
282 #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */
283 #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */
285 #define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
286 #define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
289 #define SYNC_FILE_RANGE_WAIT_BEFORE 1
290 #define SYNC_FILE_RANGE_WRITE 2
291 #define SYNC_FILE_RANGE_WAIT_AFTER 4
295 #include <linux/linkage.h>
296 #include <linux/wait.h>
297 #include <linux/types.h>
298 #include <linux/kdev_t.h>
299 #include <linux/dcache.h>
300 #include <linux/path.h>
301 #include <linux/stat.h>
302 #include <linux/cache.h>
303 #include <linux/kobject.h>
304 #include <linux/list.h>
305 #include <linux/radix-tree.h>
306 #include <linux/prio_tree.h>
307 #include <linux/init.h>
308 #include <linux/pid.h>
309 #include <linux/mutex.h>
310 #include <linux/capability.h>
311 #include <linux/semaphore.h>
312 #include <linux/fiemap.h>
314 #include <asm/atomic.h>
315 #include <asm/byteorder.h>
317 struct export_operations;
322 struct pipe_inode_info;
323 struct poll_table_struct;
325 struct vm_area_struct;
329 extern void __init inode_init(void);
330 extern void __init inode_init_early(void);
331 extern void __init files_init(unsigned long);
333 extern struct files_stat_struct files_stat;
334 extern int get_max_files(void);
335 extern int sysctl_nr_open;
336 extern struct inodes_stat_t inodes_stat;
337 extern int leases_enable, lease_break_time;
338 #ifdef CONFIG_DNOTIFY
339 extern int dir_notify_enable;
343 typedef int (get_block_t)(struct inode *inode, sector_t iblock,
344 struct buffer_head *bh_result, int create);
345 typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
346 ssize_t bytes, void *private);
349 * Attribute flags. These should be or-ed together to figure out what
352 #define ATTR_MODE (1 << 0)
353 #define ATTR_UID (1 << 1)
354 #define ATTR_GID (1 << 2)
355 #define ATTR_SIZE (1 << 3)
356 #define ATTR_ATIME (1 << 4)
357 #define ATTR_MTIME (1 << 5)
358 #define ATTR_CTIME (1 << 6)
359 #define ATTR_ATIME_SET (1 << 7)
360 #define ATTR_MTIME_SET (1 << 8)
361 #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
362 #define ATTR_ATTR_FLAG (1 << 10)
363 #define ATTR_KILL_SUID (1 << 11)
364 #define ATTR_KILL_SGID (1 << 12)
365 #define ATTR_FILE (1 << 13)
366 #define ATTR_KILL_PRIV (1 << 14)
367 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
368 #define ATTR_TIMES_SET (1 << 16)
371 * This is the Inode Attributes structure, used for notify_change(). It
372 * uses the above definitions as flags, to know which values have changed.
373 * Also, in this manner, a Filesystem can look at only the values it cares
374 * about. Basically, these are the attributes that the VFS layer can
375 * request to change from the FS layer.
377 * Derek Atkins <warlord@MIT.EDU> 94-10-20
380 unsigned int ia_valid;
385 struct timespec ia_atime;
386 struct timespec ia_mtime;
387 struct timespec ia_ctime;
390 * Not an attribute, but an auxilary info for filesystems wanting to
391 * implement an ftruncate() like method. NOTE: filesystem should
392 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
394 struct file *ia_file;
398 * Includes for diskquotas.
400 #include <linux/quota.h>
403 * enum positive_aop_returns - aop return codes with specific semantics
405 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
406 * completed, that the page is still locked, and
407 * should be considered active. The VM uses this hint
408 * to return the page to the active list -- it won't
409 * be a candidate for writeback again in the near
410 * future. Other callers must be careful to unlock
411 * the page if they get this return. Returned by
414 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
415 * unlocked it and the page might have been truncated.
416 * The caller should back up to acquiring a new page and
417 * trying again. The aop will be taking reasonable
418 * precautions not to livelock. If the caller held a page
419 * reference, it should drop it before retrying. Returned
422 * address_space_operation functions return these large constants to indicate
423 * special semantics to the caller. These are much larger than the bytes in a
424 * page to allow for functions that return the number of bytes operated on in a
428 enum positive_aop_returns {
429 AOP_WRITEPAGE_ACTIVATE = 0x80000,
430 AOP_TRUNCATED_PAGE = 0x80001,
433 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
434 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
435 #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
436 * helper code (eg buffer layer)
437 * to clear GFP_FS from alloc */
440 * oh the beauties of C type declarations.
443 struct address_space;
444 struct writeback_control;
447 const struct iovec *iov;
448 unsigned long nr_segs;
453 size_t iov_iter_copy_from_user_atomic(struct page *page,
454 struct iov_iter *i, unsigned long offset, size_t bytes);
455 size_t iov_iter_copy_from_user(struct page *page,
456 struct iov_iter *i, unsigned long offset, size_t bytes);
457 void iov_iter_advance(struct iov_iter *i, size_t bytes);
458 int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes);
459 size_t iov_iter_single_seg_count(struct iov_iter *i);
461 static inline void iov_iter_init(struct iov_iter *i,
462 const struct iovec *iov, unsigned long nr_segs,
463 size_t count, size_t written)
466 i->nr_segs = nr_segs;
468 i->count = count + written;
470 iov_iter_advance(i, written);
473 static inline size_t iov_iter_count(struct iov_iter *i)
479 * "descriptor" for what we're up to with a read.
480 * This allows us to use the same read code yet
481 * have multiple different users of the data that
482 * we read from a file.
484 * The simplest case just copies the data to user
497 typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
498 unsigned long, unsigned long);
500 struct address_space_operations {
501 int (*writepage)(struct page *page, struct writeback_control *wbc);
502 int (*readpage)(struct file *, struct page *);
503 void (*sync_page)(struct page *);
505 /* Write back some dirty pages from this mapping. */
506 int (*writepages)(struct address_space *, struct writeback_control *);
508 /* Set a page dirty. Return true if this dirtied it */
509 int (*set_page_dirty)(struct page *page);
511 int (*readpages)(struct file *filp, struct address_space *mapping,
512 struct list_head *pages, unsigned nr_pages);
514 int (*write_begin)(struct file *, struct address_space *mapping,
515 loff_t pos, unsigned len, unsigned flags,
516 struct page **pagep, void **fsdata);
517 int (*write_end)(struct file *, struct address_space *mapping,
518 loff_t pos, unsigned len, unsigned copied,
519 struct page *page, void *fsdata);
521 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
522 sector_t (*bmap)(struct address_space *, sector_t);
523 void (*invalidatepage) (struct page *, unsigned long);
524 int (*releasepage) (struct page *, gfp_t);
525 ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
526 loff_t offset, unsigned long nr_segs);
527 int (*get_xip_mem)(struct address_space *, pgoff_t, int,
528 void **, unsigned long *);
529 /* migrate the contents of a page to the specified target */
530 int (*migratepage) (struct address_space *,
531 struct page *, struct page *);
532 int (*launder_page) (struct page *);
533 int (*is_partially_uptodate) (struct page *, read_descriptor_t *,
538 * pagecache_write_begin/pagecache_write_end must be used by general code
539 * to write into the pagecache.
541 int pagecache_write_begin(struct file *, struct address_space *mapping,
542 loff_t pos, unsigned len, unsigned flags,
543 struct page **pagep, void **fsdata);
545 int pagecache_write_end(struct file *, struct address_space *mapping,
546 loff_t pos, unsigned len, unsigned copied,
547 struct page *page, void *fsdata);
549 struct backing_dev_info;
550 struct address_space {
551 struct inode *host; /* owner: inode, block_device */
552 struct radix_tree_root page_tree; /* radix tree of all pages */
553 spinlock_t tree_lock; /* and lock protecting it */
554 unsigned int i_mmap_writable;/* count VM_SHARED mappings */
555 struct prio_tree_root i_mmap; /* tree of private and shared mappings */
556 struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
557 spinlock_t i_mmap_lock; /* protect tree, count, list */
558 unsigned int truncate_count; /* Cover race condition with truncate */
559 unsigned long nrpages; /* number of total pages */
560 pgoff_t writeback_index;/* writeback starts here */
561 const struct address_space_operations *a_ops; /* methods */
562 unsigned long flags; /* error bits/gfp mask */
563 struct backing_dev_info *backing_dev_info; /* device readahead, etc */
564 spinlock_t private_lock; /* for use by the address_space */
565 struct list_head private_list; /* ditto */
566 struct address_space *assoc_mapping; /* ditto */
567 } __attribute__((aligned(sizeof(long))));
569 * On most architectures that alignment is already the case; but
570 * must be enforced here for CRIS, to let the least signficant bit
571 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
574 struct block_device {
575 dev_t bd_dev; /* not a kdev_t - it's a search key */
576 struct inode * bd_inode; /* will die */
577 struct super_block * bd_super;
579 struct mutex bd_mutex; /* open/close mutex */
580 struct semaphore bd_mount_sem;
581 struct list_head bd_inodes;
585 struct list_head bd_holder_list;
587 struct block_device * bd_contains;
588 unsigned bd_block_size;
589 struct hd_struct * bd_part;
590 /* number of times partitions within this device have been opened. */
591 unsigned bd_part_count;
593 struct gendisk * bd_disk;
594 struct list_head bd_list;
595 struct backing_dev_info *bd_inode_backing_dev_info;
597 * Private data. You must have bd_claim'ed the block_device
598 * to use this. NOTE: bd_claim allows an owner to claim
599 * the same device multiple times, the owner must take special
600 * care to not mess up bd_private for that case.
602 unsigned long bd_private;
604 /* The counter of freeze processes */
605 int bd_fsfreeze_count;
606 /* Mutex for freeze */
607 struct mutex bd_fsfreeze_mutex;
611 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
614 #define PAGECACHE_TAG_DIRTY 0
615 #define PAGECACHE_TAG_WRITEBACK 1
617 int mapping_tagged(struct address_space *mapping, int tag);
620 * Might pages of this file be mapped into userspace?
622 static inline int mapping_mapped(struct address_space *mapping)
624 return !prio_tree_empty(&mapping->i_mmap) ||
625 !list_empty(&mapping->i_mmap_nonlinear);
629 * Might pages of this file have been modified in userspace?
630 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
631 * marks vma as VM_SHARED if it is shared, and the file was opened for
632 * writing i.e. vma may be mprotected writable even if now readonly.
634 static inline int mapping_writably_mapped(struct address_space *mapping)
636 return mapping->i_mmap_writable != 0;
640 * Use sequence counter to get consistent i_size on 32-bit processors.
642 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
643 #include <linux/seqlock.h>
644 #define __NEED_I_SIZE_ORDERED
645 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
647 #define i_size_ordered_init(inode) do { } while (0)
651 struct hlist_node i_hash;
652 struct list_head i_list;
653 struct list_head i_sb_list;
654 struct list_head i_dentry;
657 unsigned int i_nlink;
663 #ifdef __NEED_I_SIZE_ORDERED
664 seqcount_t i_size_seqcount;
666 struct timespec i_atime;
667 struct timespec i_mtime;
668 struct timespec i_ctime;
669 unsigned int i_blkbits;
671 unsigned short i_bytes;
673 spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
674 struct mutex i_mutex;
675 struct rw_semaphore i_alloc_sem;
676 const struct inode_operations *i_op;
677 const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
678 struct super_block *i_sb;
679 struct file_lock *i_flock;
680 struct address_space *i_mapping;
681 struct address_space i_data;
683 struct dquot *i_dquot[MAXQUOTAS];
685 struct list_head i_devices;
687 struct pipe_inode_info *i_pipe;
688 struct block_device *i_bdev;
695 #ifdef CONFIG_DNOTIFY
696 unsigned long i_dnotify_mask; /* Directory notify events */
697 struct dnotify_struct *i_dnotify; /* for directory notifications */
700 #ifdef CONFIG_INOTIFY
701 struct list_head inotify_watches; /* watches on this inode */
702 struct mutex inotify_mutex; /* protects the watches list */
705 unsigned long i_state;
706 unsigned long dirtied_when; /* jiffies of first dirtying */
708 unsigned int i_flags;
710 atomic_t i_writecount;
711 #ifdef CONFIG_SECURITY
714 void *i_private; /* fs or device private pointer */
718 * inode->i_mutex nesting subclasses for the lock validator:
720 * 0: the object of the current VFS operation
725 * The locking order between these classes is
726 * parent -> child -> normal -> xattr -> quota
728 enum inode_i_mutex_lock_class
737 extern void inode_double_lock(struct inode *inode1, struct inode *inode2);
738 extern void inode_double_unlock(struct inode *inode1, struct inode *inode2);
741 * NOTE: in a 32bit arch with a preemptable kernel and
742 * an UP compile the i_size_read/write must be atomic
743 * with respect to the local cpu (unlike with preempt disabled),
744 * but they don't need to be atomic with respect to other cpus like in
745 * true SMP (so they need either to either locally disable irq around
746 * the read or for example on x86 they can be still implemented as a
747 * cmpxchg8b without the need of the lock prefix). For SMP compiles
748 * and 64bit archs it makes no difference if preempt is enabled or not.
750 static inline loff_t i_size_read(const struct inode *inode)
752 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
757 seq = read_seqcount_begin(&inode->i_size_seqcount);
758 i_size = inode->i_size;
759 } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
761 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
765 i_size = inode->i_size;
769 return inode->i_size;
774 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
775 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
776 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
778 static inline void i_size_write(struct inode *inode, loff_t i_size)
780 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
781 write_seqcount_begin(&inode->i_size_seqcount);
782 inode->i_size = i_size;
783 write_seqcount_end(&inode->i_size_seqcount);
784 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
786 inode->i_size = i_size;
789 inode->i_size = i_size;
793 static inline unsigned iminor(const struct inode *inode)
795 return MINOR(inode->i_rdev);
798 static inline unsigned imajor(const struct inode *inode)
800 return MAJOR(inode->i_rdev);
803 extern struct block_device *I_BDEV(struct inode *inode);
806 rwlock_t lock; /* protects pid, uid, euid fields */
807 struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
808 enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
809 uid_t uid, euid; /* uid/euid of process setting the owner */
810 int signum; /* posix.1b rt signal to be delivered on IO */
814 * Track a single file's readahead state
816 struct file_ra_state {
817 pgoff_t start; /* where readahead started */
818 unsigned int size; /* # of readahead pages */
819 unsigned int async_size; /* do asynchronous readahead when
820 there are only # of pages ahead */
822 unsigned int ra_pages; /* Maximum readahead window */
823 int mmap_miss; /* Cache miss stat for mmap accesses */
824 loff_t prev_pos; /* Cache last read() position */
828 * Check if @index falls in the readahead windows.
830 static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
832 return (index >= ra->start &&
833 index < ra->start + ra->size);
836 #define FILE_MNT_WRITE_TAKEN 1
837 #define FILE_MNT_WRITE_RELEASED 2
841 * fu_list becomes invalid after file_free is called and queued via
842 * fu_rcuhead for RCU freeing
845 struct list_head fu_list;
846 struct rcu_head fu_rcuhead;
849 #define f_dentry f_path.dentry
850 #define f_vfsmnt f_path.mnt
851 const struct file_operations *f_op;
852 spinlock_t f_lock; /* f_ep_links, f_flags, no IRQ */
853 atomic_long_t f_count;
854 unsigned int f_flags;
857 struct fown_struct f_owner;
858 const struct cred *f_cred;
859 struct file_ra_state f_ra;
862 #ifdef CONFIG_SECURITY
865 /* needed for tty driver, and maybe others */
869 /* Used by fs/eventpoll.c to link all the hooks to this file */
870 struct list_head f_ep_links;
871 #endif /* #ifdef CONFIG_EPOLL */
872 struct address_space *f_mapping;
873 #ifdef CONFIG_DEBUG_WRITECOUNT
874 unsigned long f_mnt_write_state;
877 extern spinlock_t files_lock;
878 #define file_list_lock() spin_lock(&files_lock);
879 #define file_list_unlock() spin_unlock(&files_lock);
881 #define get_file(x) atomic_long_inc(&(x)->f_count)
882 #define file_count(x) atomic_long_read(&(x)->f_count)
884 #ifdef CONFIG_DEBUG_WRITECOUNT
885 static inline void file_take_write(struct file *f)
887 WARN_ON(f->f_mnt_write_state != 0);
888 f->f_mnt_write_state = FILE_MNT_WRITE_TAKEN;
890 static inline void file_release_write(struct file *f)
892 f->f_mnt_write_state |= FILE_MNT_WRITE_RELEASED;
894 static inline void file_reset_write(struct file *f)
896 f->f_mnt_write_state = 0;
898 static inline void file_check_state(struct file *f)
901 * At this point, either both or neither of these bits
904 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN);
905 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_RELEASED);
907 static inline int file_check_writeable(struct file *f)
909 if (f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN)
911 printk(KERN_WARNING "writeable file with no "
912 "mnt_want_write()\n");
916 #else /* !CONFIG_DEBUG_WRITECOUNT */
917 static inline void file_take_write(struct file *filp) {}
918 static inline void file_release_write(struct file *filp) {}
919 static inline void file_reset_write(struct file *filp) {}
920 static inline void file_check_state(struct file *filp) {}
921 static inline int file_check_writeable(struct file *filp)
925 #endif /* CONFIG_DEBUG_WRITECOUNT */
927 #define MAX_NON_LFS ((1UL<<31) - 1)
929 /* Page cache limit. The filesystems should put that into their s_maxbytes
930 limits, otherwise bad things can happen in VM. */
931 #if BITS_PER_LONG==32
932 #define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
933 #elif BITS_PER_LONG==64
934 #define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
939 #define FL_ACCESS 8 /* not trying to lock, just looking */
940 #define FL_EXISTS 16 /* when unlocking, test for existence */
941 #define FL_LEASE 32 /* lease held on this file */
942 #define FL_CLOSE 64 /* unlock on close */
943 #define FL_SLEEP 128 /* A blocking lock */
946 * Special return value from posix_lock_file() and vfs_lock_file() for
947 * asynchronous locking.
949 #define FILE_LOCK_DEFERRED 1
952 * The POSIX file lock owner is determined by
953 * the "struct files_struct" in the thread group
954 * (or NULL for no owner - BSD locks).
956 * Lockd stuffs a "host" pointer into this.
958 typedef struct files_struct *fl_owner_t;
960 struct file_lock_operations {
961 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
962 void (*fl_release_private)(struct file_lock *);
965 struct lock_manager_operations {
966 int (*fl_compare_owner)(struct file_lock *, struct file_lock *);
967 void (*fl_notify)(struct file_lock *); /* unblock callback */
968 int (*fl_grant)(struct file_lock *, struct file_lock *, int);
969 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
970 void (*fl_release_private)(struct file_lock *);
971 void (*fl_break)(struct file_lock *);
972 int (*fl_mylease)(struct file_lock *, struct file_lock *);
973 int (*fl_change)(struct file_lock **, int);
976 struct lock_manager {
977 struct list_head list;
980 void locks_start_grace(struct lock_manager *);
981 void locks_end_grace(struct lock_manager *);
982 int locks_in_grace(void);
984 /* that will die - we need it for nfs_lock_info */
985 #include <linux/nfs_fs_i.h>
988 struct file_lock *fl_next; /* singly linked list for this inode */
989 struct list_head fl_link; /* doubly linked list of all locks */
990 struct list_head fl_block; /* circular list of blocked processes */
992 unsigned char fl_flags;
993 unsigned char fl_type;
995 struct pid *fl_nspid;
996 wait_queue_head_t fl_wait;
997 struct file *fl_file;
1001 struct fasync_struct * fl_fasync; /* for lease break notifications */
1002 unsigned long fl_break_time; /* for nonblocking lease breaks */
1004 struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
1005 struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
1007 struct nfs_lock_info nfs_fl;
1008 struct nfs4_lock_info nfs4_fl;
1010 struct list_head link; /* link in AFS vnode's pending_locks list */
1011 int state; /* state of grant or error if -ve */
1016 /* The following constant reflects the upper bound of the file/locking space */
1018 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
1019 #define OFFSET_MAX INT_LIMIT(loff_t)
1020 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
1023 #include <linux/fcntl.h>
1025 extern void send_sigio(struct fown_struct *fown, int fd, int band);
1028 extern int do_sync_mapping_range(struct address_space *mapping, loff_t offset,
1029 loff_t endbyte, unsigned int flags);
1031 #ifdef CONFIG_FILE_LOCKING
1032 extern int fcntl_getlk(struct file *, struct flock __user *);
1033 extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
1034 struct flock __user *);
1036 #if BITS_PER_LONG == 32
1037 extern int fcntl_getlk64(struct file *, struct flock64 __user *);
1038 extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
1039 struct flock64 __user *);
1042 extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
1043 extern int fcntl_getlease(struct file *filp);
1046 extern void locks_init_lock(struct file_lock *);
1047 extern void locks_copy_lock(struct file_lock *, struct file_lock *);
1048 extern void __locks_copy_lock(struct file_lock *, const struct file_lock *);
1049 extern void locks_remove_posix(struct file *, fl_owner_t);
1050 extern void locks_remove_flock(struct file *);
1051 extern void posix_test_lock(struct file *, struct file_lock *);
1052 extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
1053 extern int posix_lock_file_wait(struct file *, struct file_lock *);
1054 extern int posix_unblock_lock(struct file *, struct file_lock *);
1055 extern int vfs_test_lock(struct file *, struct file_lock *);
1056 extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
1057 extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
1058 extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
1059 extern int __break_lease(struct inode *inode, unsigned int flags);
1060 extern void lease_get_mtime(struct inode *, struct timespec *time);
1061 extern int generic_setlease(struct file *, long, struct file_lock **);
1062 extern int vfs_setlease(struct file *, long, struct file_lock **);
1063 extern int lease_modify(struct file_lock **, int);
1064 extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
1065 extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
1066 #else /* !CONFIG_FILE_LOCKING */
1067 static inline int fcntl_getlk(struct file *file, struct flock __user *user)
1072 static inline int fcntl_setlk(unsigned int fd, struct file *file,
1073 unsigned int cmd, struct flock __user *user)
1078 #if BITS_PER_LONG == 32
1079 static inline int fcntl_getlk64(struct file *file, struct flock64 __user *user)
1084 static inline int fcntl_setlk64(unsigned int fd, struct file *file,
1085 unsigned int cmd, struct flock64 __user *user)
1090 static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
1095 static inline int fcntl_getlease(struct file *filp)
1100 static inline void locks_init_lock(struct file_lock *fl)
1105 static inline void __locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1110 static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1115 static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
1120 static inline void locks_remove_flock(struct file *filp)
1125 static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
1130 static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
1131 struct file_lock *conflock)
1136 static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
1141 static inline int posix_unblock_lock(struct file *filp,
1142 struct file_lock *waiter)
1147 static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
1152 static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
1153 struct file_lock *fl, struct file_lock *conf)
1158 static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
1163 static inline int flock_lock_file_wait(struct file *filp,
1164 struct file_lock *request)
1169 static inline int __break_lease(struct inode *inode, unsigned int mode)
1174 static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
1179 static inline int generic_setlease(struct file *filp, long arg,
1180 struct file_lock **flp)
1185 static inline int vfs_setlease(struct file *filp, long arg,
1186 struct file_lock **lease)
1191 static inline int lease_modify(struct file_lock **before, int arg)
1196 static inline int lock_may_read(struct inode *inode, loff_t start,
1202 static inline int lock_may_write(struct inode *inode, loff_t start,
1208 #endif /* !CONFIG_FILE_LOCKING */
1211 struct fasync_struct {
1214 struct fasync_struct *fa_next; /* singly linked list */
1215 struct file *fa_file;
1218 #define FASYNC_MAGIC 0x4601
1220 /* SMP safe fasync helpers: */
1221 extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
1222 /* can be called from interrupts */
1223 extern void kill_fasync(struct fasync_struct **, int, int);
1224 /* only for net: no internal synchronization */
1225 extern void __kill_fasync(struct fasync_struct *, int, int);
1227 extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
1228 extern int f_setown(struct file *filp, unsigned long arg, int force);
1229 extern void f_delown(struct file *filp);
1230 extern pid_t f_getown(struct file *filp);
1231 extern int send_sigurg(struct fown_struct *fown);
1237 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1238 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1239 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1241 extern struct list_head super_blocks;
1242 extern spinlock_t sb_lock;
1244 #define sb_entry(list) list_entry((list), struct super_block, s_list)
1245 #define S_BIAS (1<<30)
1246 struct super_block {
1247 struct list_head s_list; /* Keep this first */
1248 dev_t s_dev; /* search index; _not_ kdev_t */
1249 unsigned long s_blocksize;
1250 unsigned char s_blocksize_bits;
1251 unsigned char s_dirt;
1252 unsigned long long s_maxbytes; /* Max file size */
1253 struct file_system_type *s_type;
1254 const struct super_operations *s_op;
1255 struct dquot_operations *dq_op;
1256 struct quotactl_ops *s_qcop;
1257 const struct export_operations *s_export_op;
1258 unsigned long s_flags;
1259 unsigned long s_magic;
1260 struct dentry *s_root;
1261 struct rw_semaphore s_umount;
1262 struct mutex s_lock;
1266 #ifdef CONFIG_SECURITY
1269 struct xattr_handler **s_xattr;
1271 struct list_head s_inodes; /* all inodes */
1272 struct list_head s_dirty; /* dirty inodes */
1273 struct list_head s_io; /* parked for writeback */
1274 struct list_head s_more_io; /* parked for more writeback */
1275 struct hlist_head s_anon; /* anonymous dentries for (nfs) exporting */
1276 struct list_head s_files;
1277 /* s_dentry_lru and s_nr_dentry_unused are protected by dcache_lock */
1278 struct list_head s_dentry_lru; /* unused dentry lru */
1279 int s_nr_dentry_unused; /* # of dentry on lru */
1281 struct block_device *s_bdev;
1282 struct mtd_info *s_mtd;
1283 struct list_head s_instances;
1284 struct quota_info s_dquot; /* Diskquota specific options */
1287 wait_queue_head_t s_wait_unfrozen;
1289 char s_id[32]; /* Informational name */
1291 void *s_fs_info; /* Filesystem private info */
1295 * The next field is for VFS *only*. No filesystems have any business
1296 * even looking at it. You had been warned.
1298 struct mutex s_vfs_rename_mutex; /* Kludge */
1300 /* Granularity of c/m/atime in ns.
1301 Cannot be worse than a second */
1305 * Filesystem subtype. If non-empty the filesystem type field
1306 * in /proc/mounts will be "type.subtype"
1311 * Saved mount options for lazy filesystems using
1312 * generic_show_options()
1317 * storage for asynchronous operations
1319 struct list_head s_async_list;
1322 extern struct timespec current_fs_time(struct super_block *sb);
1325 * Snapshotting support.
1329 SB_FREEZE_WRITE = 1,
1330 SB_FREEZE_TRANS = 2,
1333 #define vfs_check_frozen(sb, level) \
1334 wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
1336 #define get_fs_excl() atomic_inc(¤t->fs_excl)
1337 #define put_fs_excl() atomic_dec(¤t->fs_excl)
1338 #define has_fs_excl() atomic_read(¤t->fs_excl)
1340 #define is_owner_or_cap(inode) \
1341 ((current_fsuid() == (inode)->i_uid) || capable(CAP_FOWNER))
1343 /* not quite ready to be deprecated, but... */
1344 extern void lock_super(struct super_block *);
1345 extern void unlock_super(struct super_block *);
1348 * VFS helper functions..
1350 extern int vfs_create(struct inode *, struct dentry *, int, struct nameidata *);
1351 extern int vfs_mkdir(struct inode *, struct dentry *, int);
1352 extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t);
1353 extern int vfs_symlink(struct inode *, struct dentry *, const char *);
1354 extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
1355 extern int vfs_rmdir(struct inode *, struct dentry *);
1356 extern int vfs_unlink(struct inode *, struct dentry *);
1357 extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
1360 * VFS dentry helper functions.
1362 extern void dentry_unhash(struct dentry *dentry);
1365 * VFS file helper functions.
1367 extern int file_permission(struct file *, int);
1370 * VFS FS_IOC_FIEMAP helper definitions.
1372 struct fiemap_extent_info {
1373 unsigned int fi_flags; /* Flags as passed from user */
1374 unsigned int fi_extents_mapped; /* Number of mapped extents */
1375 unsigned int fi_extents_max; /* Size of fiemap_extent array */
1376 struct fiemap_extent *fi_extents_start; /* Start of fiemap_extent
1379 int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
1380 u64 phys, u64 len, u32 flags);
1381 int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
1386 * NOTE! These match bits 12..15 of stat.st_mode
1387 * (ie "(i_mode >> 12) & 15").
1389 #define DT_UNKNOWN 0
1399 #define OSYNC_METADATA (1<<0)
1400 #define OSYNC_DATA (1<<1)
1401 #define OSYNC_INODE (1<<2)
1402 int generic_osync_inode(struct inode *, struct address_space *, int);
1405 * This is the "filldir" function type, used by readdir() to let
1406 * the kernel specify what kind of dirent layout it wants to have.
1407 * This allows the kernel to read directories into kernel space or
1408 * to have different dirent layouts depending on the binary type.
1410 typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned);
1411 struct block_device_operations;
1413 /* These macros are for out of kernel modules to test that
1414 * the kernel supports the unlocked_ioctl and compat_ioctl
1415 * fields in struct file_operations. */
1416 #define HAVE_COMPAT_IOCTL 1
1417 #define HAVE_UNLOCKED_IOCTL 1
1421 * read, write, poll, fsync, readv, writev, unlocked_ioctl and compat_ioctl
1422 * can be called without the big kernel lock held in all filesystems.
1424 struct file_operations {
1425 struct module *owner;
1426 loff_t (*llseek) (struct file *, loff_t, int);
1427 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1428 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
1429 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1430 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1431 int (*readdir) (struct file *, void *, filldir_t);
1432 unsigned int (*poll) (struct file *, struct poll_table_struct *);
1433 int (*ioctl) (struct inode *, struct file *, unsigned int, unsigned long);
1434 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1435 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1436 int (*mmap) (struct file *, struct vm_area_struct *);
1437 int (*open) (struct inode *, struct file *);
1438 int (*flush) (struct file *, fl_owner_t id);
1439 int (*release) (struct inode *, struct file *);
1440 int (*fsync) (struct file *, struct dentry *, int datasync);
1441 int (*aio_fsync) (struct kiocb *, int datasync);
1442 int (*fasync) (int, struct file *, int);
1443 int (*lock) (struct file *, int, struct file_lock *);
1444 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1445 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1446 int (*check_flags)(int);
1447 int (*flock) (struct file *, int, struct file_lock *);
1448 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1449 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
1450 int (*setlease)(struct file *, long, struct file_lock **);
1453 struct inode_operations {
1454 int (*create) (struct inode *,struct dentry *,int, struct nameidata *);
1455 struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
1456 int (*link) (struct dentry *,struct inode *,struct dentry *);
1457 int (*unlink) (struct inode *,struct dentry *);
1458 int (*symlink) (struct inode *,struct dentry *,const char *);
1459 int (*mkdir) (struct inode *,struct dentry *,int);
1460 int (*rmdir) (struct inode *,struct dentry *);
1461 int (*mknod) (struct inode *,struct dentry *,int,dev_t);
1462 int (*rename) (struct inode *, struct dentry *,
1463 struct inode *, struct dentry *);
1464 int (*readlink) (struct dentry *, char __user *,int);
1465 void * (*follow_link) (struct dentry *, struct nameidata *);
1466 void (*put_link) (struct dentry *, struct nameidata *, void *);
1467 void (*truncate) (struct inode *);
1468 int (*permission) (struct inode *, int);
1469 int (*setattr) (struct dentry *, struct iattr *);
1470 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1471 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1472 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1473 ssize_t (*listxattr) (struct dentry *, char *, size_t);
1474 int (*removexattr) (struct dentry *, const char *);
1475 void (*truncate_range)(struct inode *, loff_t, loff_t);
1476 long (*fallocate)(struct inode *inode, int mode, loff_t offset,
1478 int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
1484 ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
1485 unsigned long nr_segs, unsigned long fast_segs,
1486 struct iovec *fast_pointer,
1487 struct iovec **ret_pointer);
1489 extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1490 extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1491 extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1492 unsigned long, loff_t *);
1493 extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1494 unsigned long, loff_t *);
1496 struct super_operations {
1497 struct inode *(*alloc_inode)(struct super_block *sb);
1498 void (*destroy_inode)(struct inode *);
1500 void (*dirty_inode) (struct inode *);
1501 int (*write_inode) (struct inode *, int);
1502 void (*drop_inode) (struct inode *);
1503 void (*delete_inode) (struct inode *);
1504 void (*put_super) (struct super_block *);
1505 void (*write_super) (struct super_block *);
1506 int (*sync_fs)(struct super_block *sb, int wait);
1507 int (*freeze_fs) (struct super_block *);
1508 int (*unfreeze_fs) (struct super_block *);
1509 int (*statfs) (struct dentry *, struct kstatfs *);
1510 int (*remount_fs) (struct super_block *, int *, char *);
1511 void (*clear_inode) (struct inode *);
1512 void (*umount_begin) (struct super_block *);
1514 int (*show_options)(struct seq_file *, struct vfsmount *);
1515 int (*show_stats)(struct seq_file *, struct vfsmount *);
1517 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1518 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
1520 int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
1524 * Inode state bits. Protected by inode_lock.
1526 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1527 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1529 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1530 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1531 * various stages of removing an inode.
1533 * Two bits are used for locking and completion notification, I_LOCK and I_SYNC.
1535 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1536 * fdatasync(). i_atime is the usual cause.
1537 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1538 * these changes separately from I_DIRTY_SYNC so that we
1539 * don't have to write inode on fdatasync() when only
1540 * mtime has changed in it.
1541 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1542 * I_NEW get_new_inode() sets i_state to I_LOCK|I_NEW. Both
1543 * are cleared by unlock_new_inode(), called from iget().
1544 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1545 * is zero. I_FREEING must be set when I_WILL_FREE is
1547 * I_FREEING Set when inode is about to be freed but still has dirty
1548 * pages or buffers attached or the inode itself is still
1550 * I_CLEAR Set by clear_inode(). In this state the inode is clean
1551 * and can be destroyed.
1553 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1554 * prohibited for many purposes. iget() must wait for
1555 * the inode to be completely released, then create it
1556 * anew. Other functions will just ignore such inodes,
1557 * if appropriate. I_LOCK is used for waiting.
1559 * I_LOCK Serves as both a mutex and completion notification.
1560 * New inodes set I_LOCK. If two processes both create
1561 * the same inode, one of them will release its inode and
1562 * wait for I_LOCK to be released before returning.
1563 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1564 * also cause waiting on I_LOCK, without I_LOCK actually
1565 * being set. find_inode() uses this to prevent returning
1566 * nearly-dead inodes.
1567 * I_SYNC Similar to I_LOCK, but limited in scope to writeback
1568 * of inode dirty data. Having a separate lock for this
1569 * purpose reduces latency and prevents some filesystem-
1570 * specific deadlocks.
1572 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1573 * Q: igrab() only checks on (I_FREEING|I_WILL_FREE). Should it also check on
1574 * I_CLEAR? If not, why?
1576 #define I_DIRTY_SYNC 1
1577 #define I_DIRTY_DATASYNC 2
1578 #define I_DIRTY_PAGES 4
1580 #define I_WILL_FREE 16
1581 #define I_FREEING 32
1584 #define I_LOCK (1 << __I_LOCK)
1586 #define I_SYNC (1 << __I_SYNC)
1588 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1590 extern void __mark_inode_dirty(struct inode *, int);
1591 static inline void mark_inode_dirty(struct inode *inode)
1593 __mark_inode_dirty(inode, I_DIRTY);
1596 static inline void mark_inode_dirty_sync(struct inode *inode)
1598 __mark_inode_dirty(inode, I_DIRTY_SYNC);
1602 * inc_nlink - directly increment an inode's link count
1605 * This is a low-level filesystem helper to replace any
1606 * direct filesystem manipulation of i_nlink. Currently,
1607 * it is only here for parity with dec_nlink().
1609 static inline void inc_nlink(struct inode *inode)
1614 static inline void inode_inc_link_count(struct inode *inode)
1617 mark_inode_dirty(inode);
1621 * drop_nlink - directly drop an inode's link count
1624 * This is a low-level filesystem helper to replace any
1625 * direct filesystem manipulation of i_nlink. In cases
1626 * where we are attempting to track writes to the
1627 * filesystem, a decrement to zero means an imminent
1628 * write when the file is truncated and actually unlinked
1629 * on the filesystem.
1631 static inline void drop_nlink(struct inode *inode)
1637 * clear_nlink - directly zero an inode's link count
1640 * This is a low-level filesystem helper to replace any
1641 * direct filesystem manipulation of i_nlink. See
1642 * drop_nlink() for why we care about i_nlink hitting zero.
1644 static inline void clear_nlink(struct inode *inode)
1649 static inline void inode_dec_link_count(struct inode *inode)
1652 mark_inode_dirty(inode);
1656 * inode_inc_iversion - increments i_version
1657 * @inode: inode that need to be updated
1659 * Every time the inode is modified, the i_version field will be incremented.
1660 * The filesystem has to be mounted with i_version flag
1663 static inline void inode_inc_iversion(struct inode *inode)
1665 spin_lock(&inode->i_lock);
1667 spin_unlock(&inode->i_lock);
1670 extern void touch_atime(struct vfsmount *mnt, struct dentry *dentry);
1671 static inline void file_accessed(struct file *file)
1673 if (!(file->f_flags & O_NOATIME))
1674 touch_atime(file->f_path.mnt, file->f_path.dentry);
1677 int sync_inode(struct inode *inode, struct writeback_control *wbc);
1679 struct file_system_type {
1682 int (*get_sb) (struct file_system_type *, int,
1683 const char *, void *, struct vfsmount *);
1684 void (*kill_sb) (struct super_block *);
1685 struct module *owner;
1686 struct file_system_type * next;
1687 struct list_head fs_supers;
1689 struct lock_class_key s_lock_key;
1690 struct lock_class_key s_umount_key;
1692 struct lock_class_key i_lock_key;
1693 struct lock_class_key i_mutex_key;
1694 struct lock_class_key i_mutex_dir_key;
1695 struct lock_class_key i_alloc_sem_key;
1698 extern int get_sb_bdev(struct file_system_type *fs_type,
1699 int flags, const char *dev_name, void *data,
1700 int (*fill_super)(struct super_block *, void *, int),
1701 struct vfsmount *mnt);
1702 extern int get_sb_single(struct file_system_type *fs_type,
1703 int flags, void *data,
1704 int (*fill_super)(struct super_block *, void *, int),
1705 struct vfsmount *mnt);
1706 extern int get_sb_nodev(struct file_system_type *fs_type,
1707 int flags, void *data,
1708 int (*fill_super)(struct super_block *, void *, int),
1709 struct vfsmount *mnt);
1710 void generic_shutdown_super(struct super_block *sb);
1711 void kill_block_super(struct super_block *sb);
1712 void kill_anon_super(struct super_block *sb);
1713 void kill_litter_super(struct super_block *sb);
1714 void deactivate_super(struct super_block *sb);
1715 int set_anon_super(struct super_block *s, void *data);
1716 struct super_block *sget(struct file_system_type *type,
1717 int (*test)(struct super_block *,void *),
1718 int (*set)(struct super_block *,void *),
1720 extern int get_sb_pseudo(struct file_system_type *, char *,
1721 const struct super_operations *ops, unsigned long,
1722 struct vfsmount *mnt);
1723 extern void simple_set_mnt(struct vfsmount *mnt, struct super_block *sb);
1724 int __put_super_and_need_restart(struct super_block *sb);
1726 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1727 #define fops_get(fops) \
1728 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1729 #define fops_put(fops) \
1730 do { if (fops) module_put((fops)->owner); } while(0)
1732 extern int register_filesystem(struct file_system_type *);
1733 extern int unregister_filesystem(struct file_system_type *);
1734 extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
1735 #define kern_mount(type) kern_mount_data(type, NULL)
1736 extern int may_umount_tree(struct vfsmount *);
1737 extern int may_umount(struct vfsmount *);
1738 extern long do_mount(char *, char *, char *, unsigned long, void *);
1739 extern struct vfsmount *collect_mounts(struct vfsmount *, struct dentry *);
1740 extern void drop_collected_mounts(struct vfsmount *);
1742 extern int vfs_statfs(struct dentry *, struct kstatfs *);
1745 extern struct kobject *fs_kobj;
1747 extern int rw_verify_area(int, struct file *, loff_t *, size_t);
1749 #define FLOCK_VERIFY_READ 1
1750 #define FLOCK_VERIFY_WRITE 2
1752 #ifdef CONFIG_FILE_LOCKING
1753 extern int locks_mandatory_locked(struct inode *);
1754 extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
1757 * Candidates for mandatory locking have the setgid bit set
1758 * but no group execute bit - an otherwise meaningless combination.
1761 static inline int __mandatory_lock(struct inode *ino)
1763 return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
1767 * ... and these candidates should be on MS_MANDLOCK mounted fs,
1768 * otherwise these will be advisory locks
1771 static inline int mandatory_lock(struct inode *ino)
1773 return IS_MANDLOCK(ino) && __mandatory_lock(ino);
1776 static inline int locks_verify_locked(struct inode *inode)
1778 if (mandatory_lock(inode))
1779 return locks_mandatory_locked(inode);
1783 static inline int locks_verify_truncate(struct inode *inode,
1787 if (inode->i_flock && mandatory_lock(inode))
1788 return locks_mandatory_area(
1789 FLOCK_VERIFY_WRITE, inode, filp,
1790 size < inode->i_size ? size : inode->i_size,
1791 (size < inode->i_size ? inode->i_size - size
1792 : size - inode->i_size)
1797 static inline int break_lease(struct inode *inode, unsigned int mode)
1800 return __break_lease(inode, mode);
1803 #else /* !CONFIG_FILE_LOCKING */
1804 static inline int locks_mandatory_locked(struct inode *inode)
1809 static inline int locks_mandatory_area(int rw, struct inode *inode,
1810 struct file *filp, loff_t offset,
1816 static inline int __mandatory_lock(struct inode *inode)
1821 static inline int mandatory_lock(struct inode *inode)
1826 static inline int locks_verify_locked(struct inode *inode)
1831 static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
1837 static inline int break_lease(struct inode *inode, unsigned int mode)
1842 #endif /* CONFIG_FILE_LOCKING */
1846 extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
1848 extern long do_sys_open(int dfd, const char __user *filename, int flags,
1850 extern struct file *filp_open(const char *, int, int);
1851 extern struct file * dentry_open(struct dentry *, struct vfsmount *, int,
1852 const struct cred *);
1853 extern int filp_close(struct file *, fl_owner_t id);
1854 extern char * getname(const char __user *);
1857 extern void __init vfs_caches_init_early(void);
1858 extern void __init vfs_caches_init(unsigned long);
1860 extern struct kmem_cache *names_cachep;
1862 #define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL)
1863 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
1864 #ifndef CONFIG_AUDITSYSCALL
1865 #define putname(name) __putname(name)
1867 extern void putname(const char *name);
1871 extern int register_blkdev(unsigned int, const char *);
1872 extern void unregister_blkdev(unsigned int, const char *);
1873 extern struct block_device *bdget(dev_t);
1874 extern void bd_set_size(struct block_device *, loff_t size);
1875 extern void bd_forget(struct inode *inode);
1876 extern void bdput(struct block_device *);
1877 extern struct block_device *open_by_devnum(dev_t, fmode_t);
1878 extern void invalidate_bdev(struct block_device *);
1879 extern int sync_blockdev(struct block_device *bdev);
1880 extern struct super_block *freeze_bdev(struct block_device *);
1881 extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
1882 extern int fsync_bdev(struct block_device *);
1883 extern int fsync_super(struct super_block *);
1884 extern int fsync_no_super(struct block_device *);
1886 static inline void bd_forget(struct inode *inode) {}
1888 extern const struct file_operations def_blk_fops;
1889 extern const struct file_operations def_chr_fops;
1890 extern const struct file_operations bad_sock_fops;
1891 extern const struct file_operations def_fifo_fops;
1893 extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
1894 extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
1895 extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
1896 extern int blkdev_get(struct block_device *, fmode_t);
1897 extern int blkdev_put(struct block_device *, fmode_t);
1898 extern int bd_claim(struct block_device *, void *);
1899 extern void bd_release(struct block_device *);
1901 extern int bd_claim_by_disk(struct block_device *, void *, struct gendisk *);
1902 extern void bd_release_from_disk(struct block_device *, struct gendisk *);
1904 #define bd_claim_by_disk(bdev, holder, disk) bd_claim(bdev, holder)
1905 #define bd_release_from_disk(bdev, disk) bd_release(bdev)
1910 #define CHRDEV_MAJOR_HASH_SIZE 255
1911 extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
1912 extern int register_chrdev_region(dev_t, unsigned, const char *);
1913 extern int register_chrdev(unsigned int, const char *,
1914 const struct file_operations *);
1915 extern void unregister_chrdev(unsigned int, const char *);
1916 extern void unregister_chrdev_region(dev_t, unsigned);
1917 extern void chrdev_show(struct seq_file *,off_t);
1919 /* fs/block_dev.c */
1920 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
1921 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
1924 #define BLKDEV_MAJOR_HASH_SIZE 255
1925 extern const char *__bdevname(dev_t, char *buffer);
1926 extern const char *bdevname(struct block_device *bdev, char *buffer);
1927 extern struct block_device *lookup_bdev(const char *);
1928 extern struct block_device *open_bdev_exclusive(const char *, fmode_t, void *);
1929 extern void close_bdev_exclusive(struct block_device *, fmode_t);
1930 extern void blkdev_show(struct seq_file *,off_t);
1933 #define BLKDEV_MAJOR_HASH_SIZE 0
1936 extern void init_special_inode(struct inode *, umode_t, dev_t);
1938 /* Invalid inode operations -- fs/bad_inode.c */
1939 extern void make_bad_inode(struct inode *);
1940 extern int is_bad_inode(struct inode *);
1942 extern const struct file_operations read_pipefifo_fops;
1943 extern const struct file_operations write_pipefifo_fops;
1944 extern const struct file_operations rdwr_pipefifo_fops;
1946 extern int fs_may_remount_ro(struct super_block *);
1950 * return READ, READA, or WRITE
1952 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
1955 * return data direction, READ or WRITE
1957 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
1959 extern void check_disk_size_change(struct gendisk *disk,
1960 struct block_device *bdev);
1961 extern int revalidate_disk(struct gendisk *);
1962 extern int check_disk_change(struct block_device *);
1963 extern int __invalidate_device(struct block_device *);
1964 extern int invalidate_partition(struct gendisk *, int);
1966 extern int invalidate_inodes(struct super_block *);
1967 unsigned long __invalidate_mapping_pages(struct address_space *mapping,
1968 pgoff_t start, pgoff_t end,
1970 unsigned long invalidate_mapping_pages(struct address_space *mapping,
1971 pgoff_t start, pgoff_t end);
1973 static inline unsigned long __deprecated
1974 invalidate_inode_pages(struct address_space *mapping)
1976 return invalidate_mapping_pages(mapping, 0, ~0UL);
1979 static inline void invalidate_remote_inode(struct inode *inode)
1981 if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
1982 S_ISLNK(inode->i_mode))
1983 invalidate_mapping_pages(inode->i_mapping, 0, -1);
1985 extern int invalidate_inode_pages2(struct address_space *mapping);
1986 extern int invalidate_inode_pages2_range(struct address_space *mapping,
1987 pgoff_t start, pgoff_t end);
1988 extern void generic_sync_sb_inodes(struct super_block *sb,
1989 struct writeback_control *wbc);
1990 extern int write_inode_now(struct inode *, int);
1991 extern int filemap_fdatawrite(struct address_space *);
1992 extern int filemap_flush(struct address_space *);
1993 extern int filemap_fdatawait(struct address_space *);
1994 extern int filemap_write_and_wait(struct address_space *mapping);
1995 extern int filemap_write_and_wait_range(struct address_space *mapping,
1996 loff_t lstart, loff_t lend);
1997 extern int wait_on_page_writeback_range(struct address_space *mapping,
1998 pgoff_t start, pgoff_t end);
1999 extern int __filemap_fdatawrite_range(struct address_space *mapping,
2000 loff_t start, loff_t end, int sync_mode);
2001 extern int filemap_fdatawrite_range(struct address_space *mapping,
2002 loff_t start, loff_t end);
2004 extern int vfs_fsync(struct file *file, struct dentry *dentry, int datasync);
2005 extern void sync_supers(void);
2006 extern void sync_filesystems(int wait);
2007 extern void __fsync_super(struct super_block *sb);
2008 extern void emergency_sync(void);
2009 extern void emergency_remount(void);
2010 extern int do_remount_sb(struct super_block *sb, int flags,
2011 void *data, int force);
2013 extern sector_t bmap(struct inode *, sector_t);
2015 extern int notify_change(struct dentry *, struct iattr *);
2016 extern int inode_permission(struct inode *, int);
2017 extern int generic_permission(struct inode *, int,
2018 int (*check_acl)(struct inode *, int));
2020 static inline bool execute_ok(struct inode *inode)
2022 return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
2025 extern int get_write_access(struct inode *);
2026 extern int deny_write_access(struct file *);
2027 static inline void put_write_access(struct inode * inode)
2029 atomic_dec(&inode->i_writecount);
2031 static inline void allow_write_access(struct file *file)
2034 atomic_inc(&file->f_path.dentry->d_inode->i_writecount);
2036 extern int do_pipe_flags(int *, int);
2037 extern struct file *create_read_pipe(struct file *f, int flags);
2038 extern struct file *create_write_pipe(int flags);
2039 extern void free_write_pipe(struct file *);
2041 extern struct file *do_filp_open(int dfd, const char *pathname,
2042 int open_flag, int mode);
2043 extern int may_open(struct path *, int, int);
2045 extern int kernel_read(struct file *, unsigned long, char *, unsigned long);
2046 extern struct file * open_exec(const char *);
2048 /* fs/dcache.c -- generic fs support functions */
2049 extern int is_subdir(struct dentry *, struct dentry *);
2050 extern ino_t find_inode_number(struct dentry *, struct qstr *);
2052 #include <linux/err.h>
2054 /* needed for stackable file system support */
2055 extern loff_t default_llseek(struct file *file, loff_t offset, int origin);
2057 extern loff_t vfs_llseek(struct file *file, loff_t offset, int origin);
2059 extern struct inode * inode_init_always(struct super_block *, struct inode *);
2060 extern void inode_init_once(struct inode *);
2061 extern void inode_add_to_lists(struct super_block *, struct inode *);
2062 extern void iput(struct inode *);
2063 extern struct inode * igrab(struct inode *);
2064 extern ino_t iunique(struct super_block *, ino_t);
2065 extern int inode_needs_sync(struct inode *inode);
2066 extern void generic_delete_inode(struct inode *inode);
2067 extern void generic_drop_inode(struct inode *inode);
2069 extern struct inode *ilookup5_nowait(struct super_block *sb,
2070 unsigned long hashval, int (*test)(struct inode *, void *),
2072 extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
2073 int (*test)(struct inode *, void *), void *data);
2074 extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
2076 extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
2077 extern struct inode * iget_locked(struct super_block *, unsigned long);
2078 extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
2079 extern int insert_inode_locked(struct inode *);
2080 extern void unlock_new_inode(struct inode *);
2082 extern void __iget(struct inode * inode);
2083 extern void iget_failed(struct inode *);
2084 extern void clear_inode(struct inode *);
2085 extern void destroy_inode(struct inode *);
2086 extern struct inode *new_inode(struct super_block *);
2087 extern int should_remove_suid(struct dentry *);
2088 extern int file_remove_suid(struct file *);
2090 extern void __insert_inode_hash(struct inode *, unsigned long hashval);
2091 extern void remove_inode_hash(struct inode *);
2092 static inline void insert_inode_hash(struct inode *inode) {
2093 __insert_inode_hash(inode, inode->i_ino);
2096 extern struct file * get_empty_filp(void);
2097 extern void file_move(struct file *f, struct list_head *list);
2098 extern void file_kill(struct file *f);
2101 extern void submit_bio(int, struct bio *);
2102 extern int bdev_read_only(struct block_device *);
2104 extern int set_blocksize(struct block_device *, int);
2105 extern int sb_set_blocksize(struct super_block *, int);
2106 extern int sb_min_blocksize(struct super_block *, int);
2107 extern int sb_has_dirty_inodes(struct super_block *);
2109 extern int generic_file_mmap(struct file *, struct vm_area_struct *);
2110 extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
2111 extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
2112 int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
2113 extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2114 extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2115 extern ssize_t generic_file_aio_write_nolock(struct kiocb *, const struct iovec *,
2116 unsigned long, loff_t);
2117 extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
2118 unsigned long *, loff_t, loff_t *, size_t, size_t);
2119 extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
2120 unsigned long, loff_t, loff_t *, size_t, ssize_t);
2121 extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
2122 extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
2123 extern int generic_segment_checks(const struct iovec *iov,
2124 unsigned long *nr_segs, size_t *count, int access_flags);
2127 extern ssize_t generic_file_splice_read(struct file *, loff_t *,
2128 struct pipe_inode_info *, size_t, unsigned int);
2129 extern ssize_t generic_file_splice_write(struct pipe_inode_info *,
2130 struct file *, loff_t *, size_t, unsigned int);
2131 extern ssize_t generic_file_splice_write_nolock(struct pipe_inode_info *,
2132 struct file *, loff_t *, size_t, unsigned int);
2133 extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
2134 struct file *out, loff_t *, size_t len, unsigned int flags);
2135 extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
2136 size_t len, unsigned int flags);
2139 file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
2140 extern loff_t no_llseek(struct file *file, loff_t offset, int origin);
2141 extern loff_t generic_file_llseek(struct file *file, loff_t offset, int origin);
2142 extern loff_t generic_file_llseek_unlocked(struct file *file, loff_t offset,
2144 extern int generic_file_open(struct inode * inode, struct file * filp);
2145 extern int nonseekable_open(struct inode * inode, struct file * filp);
2147 #ifdef CONFIG_FS_XIP
2148 extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
2150 extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
2151 extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
2152 size_t len, loff_t *ppos);
2153 extern int xip_truncate_page(struct address_space *mapping, loff_t from);
2155 static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
2162 ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
2163 struct block_device *bdev, const struct iovec *iov, loff_t offset,
2164 unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
2168 DIO_LOCKING = 1, /* need locking between buffered and direct access */
2169 DIO_NO_LOCKING, /* bdev; no locking at all between buffered/direct */
2170 DIO_OWN_LOCKING, /* filesystem locks buffered and direct internally */
2173 static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
2174 struct inode *inode, struct block_device *bdev, const struct iovec *iov,
2175 loff_t offset, unsigned long nr_segs, get_block_t get_block,
2176 dio_iodone_t end_io)
2178 return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset,
2179 nr_segs, get_block, end_io, DIO_LOCKING);
2182 static inline ssize_t blockdev_direct_IO_no_locking(int rw, struct kiocb *iocb,
2183 struct inode *inode, struct block_device *bdev, const struct iovec *iov,
2184 loff_t offset, unsigned long nr_segs, get_block_t get_block,
2185 dio_iodone_t end_io)
2187 return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset,
2188 nr_segs, get_block, end_io, DIO_NO_LOCKING);
2191 static inline ssize_t blockdev_direct_IO_own_locking(int rw, struct kiocb *iocb,
2192 struct inode *inode, struct block_device *bdev, const struct iovec *iov,
2193 loff_t offset, unsigned long nr_segs, get_block_t get_block,
2194 dio_iodone_t end_io)
2196 return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset,
2197 nr_segs, get_block, end_io, DIO_OWN_LOCKING);
2201 extern const struct file_operations generic_ro_fops;
2203 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2205 extern int vfs_readlink(struct dentry *, char __user *, int, const char *);
2206 extern int vfs_follow_link(struct nameidata *, const char *);
2207 extern int page_readlink(struct dentry *, char __user *, int);
2208 extern void *page_follow_link_light(struct dentry *, struct nameidata *);
2209 extern void page_put_link(struct dentry *, struct nameidata *, void *);
2210 extern int __page_symlink(struct inode *inode, const char *symname, int len,
2212 extern int page_symlink(struct inode *inode, const char *symname, int len);
2213 extern const struct inode_operations page_symlink_inode_operations;
2214 extern int generic_readlink(struct dentry *, char __user *, int);
2215 extern void generic_fillattr(struct inode *, struct kstat *);
2216 extern int vfs_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2217 void inode_add_bytes(struct inode *inode, loff_t bytes);
2218 void inode_sub_bytes(struct inode *inode, loff_t bytes);
2219 loff_t inode_get_bytes(struct inode *inode);
2220 void inode_set_bytes(struct inode *inode, loff_t bytes);
2222 extern int vfs_readdir(struct file *, filldir_t, void *);
2224 extern int vfs_stat(char __user *, struct kstat *);
2225 extern int vfs_lstat(char __user *, struct kstat *);
2226 extern int vfs_stat_fd(int dfd, char __user *, struct kstat *);
2227 extern int vfs_lstat_fd(int dfd, char __user *, struct kstat *);
2228 extern int vfs_fstat(unsigned int, struct kstat *);
2230 extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
2232 extern int __generic_block_fiemap(struct inode *inode,
2233 struct fiemap_extent_info *fieinfo, u64 start,
2234 u64 len, get_block_t *get_block);
2235 extern int generic_block_fiemap(struct inode *inode,
2236 struct fiemap_extent_info *fieinfo, u64 start,
2237 u64 len, get_block_t *get_block);
2239 extern void get_filesystem(struct file_system_type *fs);
2240 extern void put_filesystem(struct file_system_type *fs);
2241 extern struct file_system_type *get_fs_type(const char *name);
2242 extern struct super_block *get_super(struct block_device *);
2243 extern struct super_block *user_get_super(dev_t);
2244 extern void drop_super(struct super_block *sb);
2246 extern int dcache_dir_open(struct inode *, struct file *);
2247 extern int dcache_dir_close(struct inode *, struct file *);
2248 extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
2249 extern int dcache_readdir(struct file *, void *, filldir_t);
2250 extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2251 extern int simple_statfs(struct dentry *, struct kstatfs *);
2252 extern int simple_link(struct dentry *, struct inode *, struct dentry *);
2253 extern int simple_unlink(struct inode *, struct dentry *);
2254 extern int simple_rmdir(struct inode *, struct dentry *);
2255 extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
2256 extern int simple_sync_file(struct file *, struct dentry *, int);
2257 extern int simple_empty(struct dentry *);
2258 extern int simple_readpage(struct file *file, struct page *page);
2259 extern int simple_prepare_write(struct file *file, struct page *page,
2260 unsigned offset, unsigned to);
2261 extern int simple_write_begin(struct file *file, struct address_space *mapping,
2262 loff_t pos, unsigned len, unsigned flags,
2263 struct page **pagep, void **fsdata);
2264 extern int simple_write_end(struct file *file, struct address_space *mapping,
2265 loff_t pos, unsigned len, unsigned copied,
2266 struct page *page, void *fsdata);
2268 extern struct dentry *simple_lookup(struct inode *, struct dentry *, struct nameidata *);
2269 extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
2270 extern const struct file_operations simple_dir_operations;
2271 extern const struct inode_operations simple_dir_inode_operations;
2272 struct tree_descr { char *name; const struct file_operations *ops; int mode; };
2273 struct dentry *d_alloc_name(struct dentry *, const char *);
2274 extern int simple_fill_super(struct super_block *, int, struct tree_descr *);
2275 extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
2276 extern void simple_release_fs(struct vfsmount **mount, int *count);
2278 extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
2279 loff_t *ppos, const void *from, size_t available);
2281 #ifdef CONFIG_MIGRATION
2282 extern int buffer_migrate_page(struct address_space *,
2283 struct page *, struct page *);
2285 #define buffer_migrate_page NULL
2288 extern int inode_change_ok(struct inode *, struct iattr *);
2289 extern int __must_check inode_setattr(struct inode *, struct iattr *);
2291 extern void file_update_time(struct file *file);
2293 extern int generic_show_options(struct seq_file *m, struct vfsmount *mnt);
2294 extern void save_mount_options(struct super_block *sb, char *options);
2296 static inline ino_t parent_ino(struct dentry *dentry)
2300 spin_lock(&dentry->d_lock);
2301 res = dentry->d_parent->d_inode->i_ino;
2302 spin_unlock(&dentry->d_lock);
2306 /* Transaction based IO helpers */
2309 * An argresp is stored in an allocated page and holds the
2310 * size of the argument or response, along with its content
2312 struct simple_transaction_argresp {
2317 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2319 char *simple_transaction_get(struct file *file, const char __user *buf,
2321 ssize_t simple_transaction_read(struct file *file, char __user *buf,
2322 size_t size, loff_t *pos);
2323 int simple_transaction_release(struct inode *inode, struct file *file);
2325 static inline void simple_transaction_set(struct file *file, size_t n)
2327 struct simple_transaction_argresp *ar = file->private_data;
2329 BUG_ON(n > SIMPLE_TRANSACTION_LIMIT);
2332 * The barrier ensures that ar->size will really remain zero until
2333 * ar->data is ready for reading.
2340 * simple attribute files
2342 * These attributes behave similar to those in sysfs:
2344 * Writing to an attribute immediately sets a value, an open file can be
2345 * written to multiple times.
2347 * Reading from an attribute creates a buffer from the value that might get
2348 * read with multiple read calls. When the attribute has been read
2349 * completely, no further read calls are possible until the file is opened
2352 * All attributes contain a text representation of a numeric value
2353 * that are accessed with the get() and set() functions.
2355 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2356 static int __fops ## _open(struct inode *inode, struct file *file) \
2358 __simple_attr_check_format(__fmt, 0ull); \
2359 return simple_attr_open(inode, file, __get, __set, __fmt); \
2361 static struct file_operations __fops = { \
2362 .owner = THIS_MODULE, \
2363 .open = __fops ## _open, \
2364 .release = simple_attr_release, \
2365 .read = simple_attr_read, \
2366 .write = simple_attr_write, \
2369 static inline void __attribute__((format(printf, 1, 2)))
2370 __simple_attr_check_format(const char *fmt, ...)
2372 /* don't do anything, just let the compiler check the arguments; */
2375 int simple_attr_open(struct inode *inode, struct file *file,
2376 int (*get)(void *, u64 *), int (*set)(void *, u64),
2378 int simple_attr_release(struct inode *inode, struct file *file);
2379 ssize_t simple_attr_read(struct file *file, char __user *buf,
2380 size_t len, loff_t *ppos);
2381 ssize_t simple_attr_write(struct file *file, const char __user *buf,
2382 size_t len, loff_t *ppos);
2385 #ifdef CONFIG_SECURITY
2386 static inline char *alloc_secdata(void)
2388 return (char *)get_zeroed_page(GFP_KERNEL);
2391 static inline void free_secdata(void *secdata)
2393 free_page((unsigned long)secdata);
2396 static inline char *alloc_secdata(void)
2401 static inline void free_secdata(void *secdata)
2403 #endif /* CONFIG_SECURITY */
2406 int proc_nr_files(struct ctl_table *table, int write, struct file *filp,
2407 void __user *buffer, size_t *lenp, loff_t *ppos);
2409 int get_filesystem_list(char * buf);
2411 #endif /* __KERNEL__ */
2412 #endif /* _LINUX_FS_H */