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 extern int sysctl_nr_open
;
25 #define INR_OPEN 1024 /* Initial setting for nfile rlimits */
27 #define BLOCK_SIZE_BITS 10
28 #define BLOCK_SIZE (1<<BLOCK_SIZE_BITS)
30 #define SEEK_SET 0 /* seek relative to beginning of file */
31 #define SEEK_CUR 1 /* seek relative to current file position */
32 #define SEEK_END 2 /* seek relative to end of file */
33 #define SEEK_MAX SEEK_END
35 /* And dynamically-tunable limits and defaults: */
36 struct files_stat_struct
{
37 int nr_files
; /* read only */
38 int nr_free_files
; /* read only */
39 int max_files
; /* tunable */
41 extern struct files_stat_struct files_stat
;
42 extern int get_max_files(void);
44 struct inodes_stat_t
{
47 int dummy
[5]; /* padding for sysctl ABI compatibility */
49 extern struct inodes_stat_t inodes_stat
;
51 extern int leases_enable
, lease_break_time
;
54 extern int dir_notify_enable
;
57 #define NR_FILE 8192 /* this can well be larger on a larger system */
67 /* Internal kernel extensions */
70 #define FMODE_PWRITE FMODE_PREAD /* These go hand in hand */
72 /* File is being opened for execution. Primary users of this flag are
73 distributed filesystems that can use it to achieve correct ETXTBUSY
74 behavior for cross-node execution/opening_for_writing of files */
81 #define READA 2 /* read-ahead - don't block if no resources */
82 #define SWRITE 3 /* for ll_rw_block() - wait for buffer lock */
83 #define READ_SYNC (READ | (1 << BIO_RW_SYNC))
84 #define READ_META (READ | (1 << BIO_RW_META))
85 #define WRITE_SYNC (WRITE | (1 << BIO_RW_SYNC))
86 #define WRITE_BARRIER ((1 << BIO_RW) | (1 << BIO_RW_BARRIER))
92 /* public flags for file_system_type */
93 #define FS_REQUIRES_DEV 1
94 #define FS_BINARY_MOUNTDATA 2
95 #define FS_HAS_SUBTYPE 4
96 #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
97 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move()
98 * during rename() internally.
102 * These are the fs-independent mount-flags: up to 32 flags are supported
104 #define MS_RDONLY 1 /* Mount read-only */
105 #define MS_NOSUID 2 /* Ignore suid and sgid bits */
106 #define MS_NODEV 4 /* Disallow access to device special files */
107 #define MS_NOEXEC 8 /* Disallow program execution */
108 #define MS_SYNCHRONOUS 16 /* Writes are synced at once */
109 #define MS_REMOUNT 32 /* Alter flags of a mounted FS */
110 #define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
111 #define MS_DIRSYNC 128 /* Directory modifications are synchronous */
112 #define MS_NOATIME 1024 /* Do not update access times. */
113 #define MS_NODIRATIME 2048 /* Do not update directory access times */
117 #define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
118 MS_VERBOSE is deprecated. */
119 #define MS_SILENT 32768
120 #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
121 #define MS_UNBINDABLE (1<<17) /* change to unbindable */
122 #define MS_PRIVATE (1<<18) /* change to private */
123 #define MS_SLAVE (1<<19) /* change to slave */
124 #define MS_SHARED (1<<20) /* change to shared */
125 #define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */
126 #define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
127 #define MS_I_VERSION (1<<23) /* Update inode I_version field */
128 #define MS_ACTIVE (1<<30)
129 #define MS_NOUSER (1<<31)
132 * Superblock flags that can be altered by MS_REMOUNT
134 #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK)
137 * Old magic mount flag and mask
139 #define MS_MGC_VAL 0xC0ED0000
140 #define MS_MGC_MSK 0xffff0000
142 /* Inode flags - they have nothing to superblock flags now */
144 #define S_SYNC 1 /* Writes are synced at once */
145 #define S_NOATIME 2 /* Do not update access times */
146 #define S_APPEND 4 /* Append-only file */
147 #define S_IMMUTABLE 8 /* Immutable file */
148 #define S_DEAD 16 /* removed, but still open directory */
149 #define S_NOQUOTA 32 /* Inode is not counted to quota */
150 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
151 #define S_NOCMTIME 128 /* Do not update file c/mtime */
152 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
153 #define S_PRIVATE 512 /* Inode is fs-internal */
156 * Note that nosuid etc flags are inode-specific: setting some file-system
157 * flags just means all the inodes inherit those flags by default. It might be
158 * possible to override it selectively if you really wanted to with some
159 * ioctl() that is not currently implemented.
161 * Exception: MS_RDONLY is always applied to the entire file system.
163 * Unfortunately, it is possible to change a filesystems flags with it mounted
164 * with files in use. This means that all of the inodes will not have their
165 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
166 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
168 #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
170 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
171 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
172 ((inode)->i_flags & S_SYNC))
173 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
174 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
175 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
176 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
177 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
179 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
180 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
181 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
182 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
184 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
185 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
186 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
187 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
189 /* the read-only stuff doesn't really belong here, but any other place is
190 probably as bad and I don't want to create yet another include file. */
192 #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
193 #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
194 #define BLKRRPART _IO(0x12,95) /* re-read partition table */
195 #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
196 #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
197 #define BLKRASET _IO(0x12,98) /* set read ahead for block device */
198 #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
199 #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
200 #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
201 #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
202 #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
203 #define BLKSSZGET _IO(0x12,104)/* get block device sector size */
205 #define BLKPG _IO(0x12,105)/* See blkpg.h */
207 /* Some people are morons. Do not use sizeof! */
209 #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
210 #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
211 /* This was here just to show that the number is taken -
212 probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
214 /* A jump here: 108-111 have been used for various private purposes. */
215 #define BLKBSZGET _IOR(0x12,112,size_t)
216 #define BLKBSZSET _IOW(0x12,113,size_t)
217 #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
218 #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
219 #define BLKTRACESTART _IO(0x12,116)
220 #define BLKTRACESTOP _IO(0x12,117)
221 #define BLKTRACETEARDOWN _IO(0x12,118)
223 #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
224 #define FIBMAP _IO(0x00,1) /* bmap access */
225 #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
227 #define FS_IOC_GETFLAGS _IOR('f', 1, long)
228 #define FS_IOC_SETFLAGS _IOW('f', 2, long)
229 #define FS_IOC_GETVERSION _IOR('v', 1, long)
230 #define FS_IOC_SETVERSION _IOW('v', 2, long)
231 #define FS_IOC32_GETFLAGS _IOR('f', 1, int)
232 #define FS_IOC32_SETFLAGS _IOW('f', 2, int)
233 #define FS_IOC32_GETVERSION _IOR('v', 1, int)
234 #define FS_IOC32_SETVERSION _IOW('v', 2, int)
237 * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS)
239 #define FS_SECRM_FL 0x00000001 /* Secure deletion */
240 #define FS_UNRM_FL 0x00000002 /* Undelete */
241 #define FS_COMPR_FL 0x00000004 /* Compress file */
242 #define FS_SYNC_FL 0x00000008 /* Synchronous updates */
243 #define FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
244 #define FS_APPEND_FL 0x00000020 /* writes to file may only append */
245 #define FS_NODUMP_FL 0x00000040 /* do not dump file */
246 #define FS_NOATIME_FL 0x00000080 /* do not update atime */
247 /* Reserved for compression usage... */
248 #define FS_DIRTY_FL 0x00000100
249 #define FS_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
250 #define FS_NOCOMP_FL 0x00000400 /* Don't compress */
251 #define FS_ECOMPR_FL 0x00000800 /* Compression error */
252 /* End compression flags --- maybe not all used */
253 #define FS_BTREE_FL 0x00001000 /* btree format dir */
254 #define FS_INDEX_FL 0x00001000 /* hash-indexed directory */
255 #define FS_IMAGIC_FL 0x00002000 /* AFS directory */
256 #define FS_JOURNAL_DATA_FL 0x00004000 /* Reserved for ext3 */
257 #define FS_NOTAIL_FL 0x00008000 /* file tail should not be merged */
258 #define FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
259 #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
260 #define FS_EXTENT_FL 0x00080000 /* Extents */
261 #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */
262 #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */
264 #define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
265 #define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
268 #define SYNC_FILE_RANGE_WAIT_BEFORE 1
269 #define SYNC_FILE_RANGE_WRITE 2
270 #define SYNC_FILE_RANGE_WAIT_AFTER 4
274 #include <linux/linkage.h>
275 #include <linux/wait.h>
276 #include <linux/types.h>
277 #include <linux/kdev_t.h>
278 #include <linux/dcache.h>
279 #include <linux/namei.h>
280 #include <linux/stat.h>
281 #include <linux/cache.h>
282 #include <linux/kobject.h>
283 #include <linux/list.h>
284 #include <linux/radix-tree.h>
285 #include <linux/prio_tree.h>
286 #include <linux/init.h>
287 #include <linux/pid.h>
288 #include <linux/mutex.h>
289 #include <linux/capability.h>
290 #include <linux/semaphore.h>
292 #include <asm/atomic.h>
293 #include <asm/byteorder.h>
295 struct export_operations
;
300 struct pipe_inode_info
;
301 struct poll_table_struct
;
303 struct vm_area_struct
;
306 extern void __init
inode_init(void);
307 extern void __init
inode_init_early(void);
308 extern void __init
files_init(unsigned long);
311 typedef int (get_block_t
)(struct inode
*inode
, sector_t iblock
,
312 struct buffer_head
*bh_result
, int create
);
313 typedef void (dio_iodone_t
)(struct kiocb
*iocb
, loff_t offset
,
314 ssize_t bytes
, void *private);
317 * Attribute flags. These should be or-ed together to figure out what
324 #define ATTR_ATIME 16
325 #define ATTR_MTIME 32
326 #define ATTR_CTIME 64
327 #define ATTR_ATIME_SET 128
328 #define ATTR_MTIME_SET 256
329 #define ATTR_FORCE 512 /* Not a change, but a change it */
330 #define ATTR_ATTR_FLAG 1024
331 #define ATTR_KILL_SUID 2048
332 #define ATTR_KILL_SGID 4096
333 #define ATTR_FILE 8192
334 #define ATTR_KILL_PRIV 16384
335 #define ATTR_OPEN 32768 /* Truncating from open(O_TRUNC) */
338 * This is the Inode Attributes structure, used for notify_change(). It
339 * uses the above definitions as flags, to know which values have changed.
340 * Also, in this manner, a Filesystem can look at only the values it cares
341 * about. Basically, these are the attributes that the VFS layer can
342 * request to change from the FS layer.
344 * Derek Atkins <warlord@MIT.EDU> 94-10-20
347 unsigned int ia_valid
;
352 struct timespec ia_atime
;
353 struct timespec ia_mtime
;
354 struct timespec ia_ctime
;
357 * Not an attribute, but an auxilary info for filesystems wanting to
358 * implement an ftruncate() like method. NOTE: filesystem should
359 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
361 struct file
*ia_file
;
365 * Includes for diskquotas.
367 #include <linux/quota.h>
370 * enum positive_aop_returns - aop return codes with specific semantics
372 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
373 * completed, that the page is still locked, and
374 * should be considered active. The VM uses this hint
375 * to return the page to the active list -- it won't
376 * be a candidate for writeback again in the near
377 * future. Other callers must be careful to unlock
378 * the page if they get this return. Returned by
381 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
382 * unlocked it and the page might have been truncated.
383 * The caller should back up to acquiring a new page and
384 * trying again. The aop will be taking reasonable
385 * precautions not to livelock. If the caller held a page
386 * reference, it should drop it before retrying. Returned
389 * address_space_operation functions return these large constants to indicate
390 * special semantics to the caller. These are much larger than the bytes in a
391 * page to allow for functions that return the number of bytes operated on in a
395 enum positive_aop_returns
{
396 AOP_WRITEPAGE_ACTIVATE
= 0x80000,
397 AOP_TRUNCATED_PAGE
= 0x80001,
400 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
401 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
404 * oh the beauties of C type declarations.
407 struct address_space
;
408 struct writeback_control
;
411 const struct iovec
*iov
;
412 unsigned long nr_segs
;
417 size_t iov_iter_copy_from_user_atomic(struct page
*page
,
418 struct iov_iter
*i
, unsigned long offset
, size_t bytes
);
419 size_t iov_iter_copy_from_user(struct page
*page
,
420 struct iov_iter
*i
, unsigned long offset
, size_t bytes
);
421 void iov_iter_advance(struct iov_iter
*i
, size_t bytes
);
422 int iov_iter_fault_in_readable(struct iov_iter
*i
, size_t bytes
);
423 size_t iov_iter_single_seg_count(struct iov_iter
*i
);
425 static inline void iov_iter_init(struct iov_iter
*i
,
426 const struct iovec
*iov
, unsigned long nr_segs
,
427 size_t count
, size_t written
)
430 i
->nr_segs
= nr_segs
;
432 i
->count
= count
+ written
;
434 iov_iter_advance(i
, written
);
437 static inline size_t iov_iter_count(struct iov_iter
*i
)
443 struct address_space_operations
{
444 int (*writepage
)(struct page
*page
, struct writeback_control
*wbc
);
445 int (*readpage
)(struct file
*, struct page
*);
446 void (*sync_page
)(struct page
*);
448 /* Write back some dirty pages from this mapping. */
449 int (*writepages
)(struct address_space
*, struct writeback_control
*);
451 /* Set a page dirty. Return true if this dirtied it */
452 int (*set_page_dirty
)(struct page
*page
);
454 int (*readpages
)(struct file
*filp
, struct address_space
*mapping
,
455 struct list_head
*pages
, unsigned nr_pages
);
458 * ext3 requires that a successful prepare_write() call be followed
459 * by a commit_write() call - they must be balanced
461 int (*prepare_write
)(struct file
*, struct page
*, unsigned, unsigned);
462 int (*commit_write
)(struct file
*, struct page
*, unsigned, unsigned);
464 int (*write_begin
)(struct file
*, struct address_space
*mapping
,
465 loff_t pos
, unsigned len
, unsigned flags
,
466 struct page
**pagep
, void **fsdata
);
467 int (*write_end
)(struct file
*, struct address_space
*mapping
,
468 loff_t pos
, unsigned len
, unsigned copied
,
469 struct page
*page
, void *fsdata
);
471 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
472 sector_t (*bmap
)(struct address_space
*, sector_t
);
473 void (*invalidatepage
) (struct page
*, unsigned long);
474 int (*releasepage
) (struct page
*, gfp_t
);
475 ssize_t (*direct_IO
)(int, struct kiocb
*, const struct iovec
*iov
,
476 loff_t offset
, unsigned long nr_segs
);
477 int (*get_xip_mem
)(struct address_space
*, pgoff_t
, int,
478 void **, unsigned long *);
479 /* migrate the contents of a page to the specified target */
480 int (*migratepage
) (struct address_space
*,
481 struct page
*, struct page
*);
482 int (*launder_page
) (struct page
*);
486 * pagecache_write_begin/pagecache_write_end must be used by general code
487 * to write into the pagecache.
489 int pagecache_write_begin(struct file
*, struct address_space
*mapping
,
490 loff_t pos
, unsigned len
, unsigned flags
,
491 struct page
**pagep
, void **fsdata
);
493 int pagecache_write_end(struct file
*, struct address_space
*mapping
,
494 loff_t pos
, unsigned len
, unsigned copied
,
495 struct page
*page
, void *fsdata
);
497 struct backing_dev_info
;
498 struct address_space
{
499 struct inode
*host
; /* owner: inode, block_device */
500 struct radix_tree_root page_tree
; /* radix tree of all pages */
501 rwlock_t tree_lock
; /* and rwlock protecting it */
502 unsigned int i_mmap_writable
;/* count VM_SHARED mappings */
503 struct prio_tree_root i_mmap
; /* tree of private and shared mappings */
504 struct list_head i_mmap_nonlinear
;/*list VM_NONLINEAR mappings */
505 spinlock_t i_mmap_lock
; /* protect tree, count, list */
506 unsigned int truncate_count
; /* Cover race condition with truncate */
507 unsigned long nrpages
; /* number of total pages */
508 pgoff_t writeback_index
;/* writeback starts here */
509 const struct address_space_operations
*a_ops
; /* methods */
510 unsigned long flags
; /* error bits/gfp mask */
511 struct backing_dev_info
*backing_dev_info
; /* device readahead, etc */
512 spinlock_t private_lock
; /* for use by the address_space */
513 struct list_head private_list
; /* ditto */
514 struct address_space
*assoc_mapping
; /* ditto */
515 } __attribute__((aligned(sizeof(long))));
517 * On most architectures that alignment is already the case; but
518 * must be enforced here for CRIS, to let the least signficant bit
519 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
522 struct block_device
{
523 dev_t bd_dev
; /* not a kdev_t - it's a search key */
524 struct inode
* bd_inode
; /* will die */
526 struct mutex bd_mutex
; /* open/close mutex */
527 struct semaphore bd_mount_sem
;
528 struct list_head bd_inodes
;
532 struct list_head bd_holder_list
;
534 struct block_device
* bd_contains
;
535 unsigned bd_block_size
;
536 struct hd_struct
* bd_part
;
537 /* number of times partitions within this device have been opened. */
538 unsigned bd_part_count
;
540 struct gendisk
* bd_disk
;
541 struct list_head bd_list
;
542 struct backing_dev_info
*bd_inode_backing_dev_info
;
544 * Private data. You must have bd_claim'ed the block_device
545 * to use this. NOTE: bd_claim allows an owner to claim
546 * the same device multiple times, the owner must take special
547 * care to not mess up bd_private for that case.
549 unsigned long bd_private
;
553 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
556 #define PAGECACHE_TAG_DIRTY 0
557 #define PAGECACHE_TAG_WRITEBACK 1
559 int mapping_tagged(struct address_space
*mapping
, int tag
);
562 * Might pages of this file be mapped into userspace?
564 static inline int mapping_mapped(struct address_space
*mapping
)
566 return !prio_tree_empty(&mapping
->i_mmap
) ||
567 !list_empty(&mapping
->i_mmap_nonlinear
);
571 * Might pages of this file have been modified in userspace?
572 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
573 * marks vma as VM_SHARED if it is shared, and the file was opened for
574 * writing i.e. vma may be mprotected writable even if now readonly.
576 static inline int mapping_writably_mapped(struct address_space
*mapping
)
578 return mapping
->i_mmap_writable
!= 0;
582 * Use sequence counter to get consistent i_size on 32-bit processors.
584 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
585 #include <linux/seqlock.h>
586 #define __NEED_I_SIZE_ORDERED
587 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
589 #define i_size_ordered_init(inode) do { } while (0)
593 struct hlist_node i_hash
;
594 struct list_head i_list
;
595 struct list_head i_sb_list
;
596 struct list_head i_dentry
;
599 unsigned int i_nlink
;
605 #ifdef __NEED_I_SIZE_ORDERED
606 seqcount_t i_size_seqcount
;
608 struct timespec i_atime
;
609 struct timespec i_mtime
;
610 struct timespec i_ctime
;
611 unsigned int i_blkbits
;
613 unsigned short i_bytes
;
615 spinlock_t i_lock
; /* i_blocks, i_bytes, maybe i_size */
616 struct mutex i_mutex
;
617 struct rw_semaphore i_alloc_sem
;
618 const struct inode_operations
*i_op
;
619 const struct file_operations
*i_fop
; /* former ->i_op->default_file_ops */
620 struct super_block
*i_sb
;
621 struct file_lock
*i_flock
;
622 struct address_space
*i_mapping
;
623 struct address_space i_data
;
625 struct dquot
*i_dquot
[MAXQUOTAS
];
627 struct list_head i_devices
;
629 struct pipe_inode_info
*i_pipe
;
630 struct block_device
*i_bdev
;
637 #ifdef CONFIG_DNOTIFY
638 unsigned long i_dnotify_mask
; /* Directory notify events */
639 struct dnotify_struct
*i_dnotify
; /* for directory notifications */
642 #ifdef CONFIG_INOTIFY
643 struct list_head inotify_watches
; /* watches on this inode */
644 struct mutex inotify_mutex
; /* protects the watches list */
647 unsigned long i_state
;
648 unsigned long dirtied_when
; /* jiffies of first dirtying */
650 unsigned int i_flags
;
652 atomic_t i_writecount
;
653 #ifdef CONFIG_SECURITY
656 void *i_private
; /* fs or device private pointer */
660 * inode->i_mutex nesting subclasses for the lock validator:
662 * 0: the object of the current VFS operation
667 * The locking order between these classes is
668 * parent -> child -> normal -> xattr -> quota
670 enum inode_i_mutex_lock_class
679 extern void inode_double_lock(struct inode
*inode1
, struct inode
*inode2
);
680 extern void inode_double_unlock(struct inode
*inode1
, struct inode
*inode2
);
683 * NOTE: in a 32bit arch with a preemptable kernel and
684 * an UP compile the i_size_read/write must be atomic
685 * with respect to the local cpu (unlike with preempt disabled),
686 * but they don't need to be atomic with respect to other cpus like in
687 * true SMP (so they need either to either locally disable irq around
688 * the read or for example on x86 they can be still implemented as a
689 * cmpxchg8b without the need of the lock prefix). For SMP compiles
690 * and 64bit archs it makes no difference if preempt is enabled or not.
692 static inline loff_t
i_size_read(const struct inode
*inode
)
694 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
699 seq
= read_seqcount_begin(&inode
->i_size_seqcount
);
700 i_size
= inode
->i_size
;
701 } while (read_seqcount_retry(&inode
->i_size_seqcount
, seq
));
703 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
707 i_size
= inode
->i_size
;
711 return inode
->i_size
;
716 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
717 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
718 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
720 static inline void i_size_write(struct inode
*inode
, loff_t i_size
)
722 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
723 write_seqcount_begin(&inode
->i_size_seqcount
);
724 inode
->i_size
= i_size
;
725 write_seqcount_end(&inode
->i_size_seqcount
);
726 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
728 inode
->i_size
= i_size
;
731 inode
->i_size
= i_size
;
735 static inline unsigned iminor(const struct inode
*inode
)
737 return MINOR(inode
->i_rdev
);
740 static inline unsigned imajor(const struct inode
*inode
)
742 return MAJOR(inode
->i_rdev
);
745 extern struct block_device
*I_BDEV(struct inode
*inode
);
748 rwlock_t lock
; /* protects pid, uid, euid fields */
749 struct pid
*pid
; /* pid or -pgrp where SIGIO should be sent */
750 enum pid_type pid_type
; /* Kind of process group SIGIO should be sent to */
751 uid_t uid
, euid
; /* uid/euid of process setting the owner */
752 int signum
; /* posix.1b rt signal to be delivered on IO */
756 * Track a single file's readahead state
758 struct file_ra_state
{
759 pgoff_t start
; /* where readahead started */
760 unsigned int size
; /* # of readahead pages */
761 unsigned int async_size
; /* do asynchronous readahead when
762 there are only # of pages ahead */
764 unsigned int ra_pages
; /* Maximum readahead window */
765 int mmap_miss
; /* Cache miss stat for mmap accesses */
766 loff_t prev_pos
; /* Cache last read() position */
770 * Check if @index falls in the readahead windows.
772 static inline int ra_has_index(struct file_ra_state
*ra
, pgoff_t index
)
774 return (index
>= ra
->start
&&
775 index
< ra
->start
+ ra
->size
);
778 #define FILE_MNT_WRITE_TAKEN 1
779 #define FILE_MNT_WRITE_RELEASED 2
783 * fu_list becomes invalid after file_free is called and queued via
784 * fu_rcuhead for RCU freeing
787 struct list_head fu_list
;
788 struct rcu_head fu_rcuhead
;
791 #define f_dentry f_path.dentry
792 #define f_vfsmnt f_path.mnt
793 const struct file_operations
*f_op
;
795 unsigned int f_flags
;
798 struct fown_struct f_owner
;
799 unsigned int f_uid
, f_gid
;
800 struct file_ra_state f_ra
;
803 #ifdef CONFIG_SECURITY
806 /* needed for tty driver, and maybe others */
810 /* Used by fs/eventpoll.c to link all the hooks to this file */
811 struct list_head f_ep_links
;
812 spinlock_t f_ep_lock
;
813 #endif /* #ifdef CONFIG_EPOLL */
814 struct address_space
*f_mapping
;
815 #ifdef CONFIG_DEBUG_WRITECOUNT
816 unsigned long f_mnt_write_state
;
819 extern spinlock_t files_lock
;
820 #define file_list_lock() spin_lock(&files_lock);
821 #define file_list_unlock() spin_unlock(&files_lock);
823 #define get_file(x) atomic_inc(&(x)->f_count)
824 #define file_count(x) atomic_read(&(x)->f_count)
826 #ifdef CONFIG_DEBUG_WRITECOUNT
827 static inline void file_take_write(struct file
*f
)
829 WARN_ON(f
->f_mnt_write_state
!= 0);
830 f
->f_mnt_write_state
= FILE_MNT_WRITE_TAKEN
;
832 static inline void file_release_write(struct file
*f
)
834 f
->f_mnt_write_state
|= FILE_MNT_WRITE_RELEASED
;
836 static inline void file_reset_write(struct file
*f
)
838 f
->f_mnt_write_state
= 0;
840 static inline void file_check_state(struct file
*f
)
843 * At this point, either both or neither of these bits
846 WARN_ON(f
->f_mnt_write_state
== FILE_MNT_WRITE_TAKEN
);
847 WARN_ON(f
->f_mnt_write_state
== FILE_MNT_WRITE_RELEASED
);
849 static inline int file_check_writeable(struct file
*f
)
851 if (f
->f_mnt_write_state
== FILE_MNT_WRITE_TAKEN
)
853 printk(KERN_WARNING
"writeable file with no "
854 "mnt_want_write()\n");
858 #else /* !CONFIG_DEBUG_WRITECOUNT */
859 static inline void file_take_write(struct file
*filp
) {}
860 static inline void file_release_write(struct file
*filp
) {}
861 static inline void file_reset_write(struct file
*filp
) {}
862 static inline void file_check_state(struct file
*filp
) {}
863 static inline int file_check_writeable(struct file
*filp
)
867 #endif /* CONFIG_DEBUG_WRITECOUNT */
869 #define MAX_NON_LFS ((1UL<<31) - 1)
871 /* Page cache limit. The filesystems should put that into their s_maxbytes
872 limits, otherwise bad things can happen in VM. */
873 #if BITS_PER_LONG==32
874 #define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
875 #elif BITS_PER_LONG==64
876 #define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
881 #define FL_ACCESS 8 /* not trying to lock, just looking */
882 #define FL_EXISTS 16 /* when unlocking, test for existence */
883 #define FL_LEASE 32 /* lease held on this file */
884 #define FL_CLOSE 64 /* unlock on close */
885 #define FL_SLEEP 128 /* A blocking lock */
888 * The POSIX file lock owner is determined by
889 * the "struct files_struct" in the thread group
890 * (or NULL for no owner - BSD locks).
892 * Lockd stuffs a "host" pointer into this.
894 typedef struct files_struct
*fl_owner_t
;
896 struct file_lock_operations
{
897 void (*fl_insert
)(struct file_lock
*); /* lock insertion callback */
898 void (*fl_remove
)(struct file_lock
*); /* lock removal callback */
899 void (*fl_copy_lock
)(struct file_lock
*, struct file_lock
*);
900 void (*fl_release_private
)(struct file_lock
*);
903 struct lock_manager_operations
{
904 int (*fl_compare_owner
)(struct file_lock
*, struct file_lock
*);
905 void (*fl_notify
)(struct file_lock
*); /* unblock callback */
906 int (*fl_grant
)(struct file_lock
*, struct file_lock
*, int);
907 void (*fl_copy_lock
)(struct file_lock
*, struct file_lock
*);
908 void (*fl_release_private
)(struct file_lock
*);
909 void (*fl_break
)(struct file_lock
*);
910 int (*fl_mylease
)(struct file_lock
*, struct file_lock
*);
911 int (*fl_change
)(struct file_lock
**, int);
914 /* that will die - we need it for nfs_lock_info */
915 #include <linux/nfs_fs_i.h>
918 struct file_lock
*fl_next
; /* singly linked list for this inode */
919 struct list_head fl_link
; /* doubly linked list of all locks */
920 struct list_head fl_block
; /* circular list of blocked processes */
923 struct pid
*fl_nspid
;
924 wait_queue_head_t fl_wait
;
925 struct file
*fl_file
;
926 unsigned char fl_flags
;
927 unsigned char fl_type
;
931 struct fasync_struct
* fl_fasync
; /* for lease break notifications */
932 unsigned long fl_break_time
; /* for nonblocking lease breaks */
934 struct file_lock_operations
*fl_ops
; /* Callbacks for filesystems */
935 struct lock_manager_operations
*fl_lmops
; /* Callbacks for lockmanagers */
937 struct nfs_lock_info nfs_fl
;
938 struct nfs4_lock_info nfs4_fl
;
940 struct list_head link
; /* link in AFS vnode's pending_locks list */
941 int state
; /* state of grant or error if -ve */
946 /* The following constant reflects the upper bound of the file/locking space */
948 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
949 #define OFFSET_MAX INT_LIMIT(loff_t)
950 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
953 #include <linux/fcntl.h>
955 extern int fcntl_getlk(struct file
*, struct flock __user
*);
956 extern int fcntl_setlk(unsigned int, struct file
*, unsigned int,
957 struct flock __user
*);
959 #if BITS_PER_LONG == 32
960 extern int fcntl_getlk64(struct file
*, struct flock64 __user
*);
961 extern int fcntl_setlk64(unsigned int, struct file
*, unsigned int,
962 struct flock64 __user
*);
965 extern void send_sigio(struct fown_struct
*fown
, int fd
, int band
);
966 extern int fcntl_setlease(unsigned int fd
, struct file
*filp
, long arg
);
967 extern int fcntl_getlease(struct file
*filp
);
970 extern int do_sync_mapping_range(struct address_space
*mapping
, loff_t offset
,
971 loff_t endbyte
, unsigned int flags
);
974 extern void locks_init_lock(struct file_lock
*);
975 extern void locks_copy_lock(struct file_lock
*, struct file_lock
*);
976 extern void __locks_copy_lock(struct file_lock
*, const struct file_lock
*);
977 extern void locks_remove_posix(struct file
*, fl_owner_t
);
978 extern void locks_remove_flock(struct file
*);
979 extern void posix_test_lock(struct file
*, struct file_lock
*);
980 extern int posix_lock_file(struct file
*, struct file_lock
*, struct file_lock
*);
981 extern int posix_lock_file_wait(struct file
*, struct file_lock
*);
982 extern int posix_unblock_lock(struct file
*, struct file_lock
*);
983 extern int vfs_test_lock(struct file
*, struct file_lock
*);
984 extern int vfs_lock_file(struct file
*, unsigned int, struct file_lock
*, struct file_lock
*);
985 extern int vfs_cancel_lock(struct file
*filp
, struct file_lock
*fl
);
986 extern int flock_lock_file_wait(struct file
*filp
, struct file_lock
*fl
);
987 extern int __break_lease(struct inode
*inode
, unsigned int flags
);
988 extern void lease_get_mtime(struct inode
*, struct timespec
*time
);
989 extern int generic_setlease(struct file
*, long, struct file_lock
**);
990 extern int vfs_setlease(struct file
*, long, struct file_lock
**);
991 extern int lease_modify(struct file_lock
**, int);
992 extern int lock_may_read(struct inode
*, loff_t start
, unsigned long count
);
993 extern int lock_may_write(struct inode
*, loff_t start
, unsigned long count
);
994 extern struct seq_operations locks_seq_operations
;
996 struct fasync_struct
{
999 struct fasync_struct
*fa_next
; /* singly linked list */
1000 struct file
*fa_file
;
1003 #define FASYNC_MAGIC 0x4601
1005 /* SMP safe fasync helpers: */
1006 extern int fasync_helper(int, struct file
*, int, struct fasync_struct
**);
1007 /* can be called from interrupts */
1008 extern void kill_fasync(struct fasync_struct
**, int, int);
1009 /* only for net: no internal synchronization */
1010 extern void __kill_fasync(struct fasync_struct
*, int, int);
1012 extern int __f_setown(struct file
*filp
, struct pid
*, enum pid_type
, int force
);
1013 extern int f_setown(struct file
*filp
, unsigned long arg
, int force
);
1014 extern void f_delown(struct file
*filp
);
1015 extern pid_t
f_getown(struct file
*filp
);
1016 extern int send_sigurg(struct fown_struct
*fown
);
1022 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1023 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1024 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1026 extern struct list_head super_blocks
;
1027 extern spinlock_t sb_lock
;
1029 #define S_BIAS (1<<30)
1030 struct super_block
{
1031 struct list_head s_list
; /* Keep this first */
1032 dev_t s_dev
; /* search index; _not_ kdev_t */
1033 unsigned long s_blocksize
;
1034 unsigned char s_blocksize_bits
;
1035 unsigned char s_dirt
;
1036 unsigned long long s_maxbytes
; /* Max file size */
1037 struct file_system_type
*s_type
;
1038 const struct super_operations
*s_op
;
1039 struct dquot_operations
*dq_op
;
1040 struct quotactl_ops
*s_qcop
;
1041 const struct export_operations
*s_export_op
;
1042 unsigned long s_flags
;
1043 unsigned long s_magic
;
1044 struct dentry
*s_root
;
1045 struct rw_semaphore s_umount
;
1046 struct mutex s_lock
;
1051 #ifdef CONFIG_SECURITY
1054 struct xattr_handler
**s_xattr
;
1056 struct list_head s_inodes
; /* all inodes */
1057 struct list_head s_dirty
; /* dirty inodes */
1058 struct list_head s_io
; /* parked for writeback */
1059 struct list_head s_more_io
; /* parked for more writeback */
1060 struct hlist_head s_anon
; /* anonymous dentries for (nfs) exporting */
1061 struct list_head s_files
;
1063 struct block_device
*s_bdev
;
1064 struct mtd_info
*s_mtd
;
1065 struct list_head s_instances
;
1066 struct quota_info s_dquot
; /* Diskquota specific options */
1069 wait_queue_head_t s_wait_unfrozen
;
1071 char s_id
[32]; /* Informational name */
1073 void *s_fs_info
; /* Filesystem private info */
1076 * The next field is for VFS *only*. No filesystems have any business
1077 * even looking at it. You had been warned.
1079 struct mutex s_vfs_rename_mutex
; /* Kludge */
1081 /* Granularity of c/m/atime in ns.
1082 Cannot be worse than a second */
1086 * Filesystem subtype. If non-empty the filesystem type field
1087 * in /proc/mounts will be "type.subtype"
1092 * Saved mount options for lazy filesystems using
1093 * generic_show_options()
1098 extern struct timespec
current_fs_time(struct super_block
*sb
);
1101 * Snapshotting support.
1105 SB_FREEZE_WRITE
= 1,
1106 SB_FREEZE_TRANS
= 2,
1109 #define vfs_check_frozen(sb, level) \
1110 wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
1112 #define get_fs_excl() atomic_inc(¤t->fs_excl)
1113 #define put_fs_excl() atomic_dec(¤t->fs_excl)
1114 #define has_fs_excl() atomic_read(¤t->fs_excl)
1116 #define is_owner_or_cap(inode) \
1117 ((current->fsuid == (inode)->i_uid) || capable(CAP_FOWNER))
1119 /* not quite ready to be deprecated, but... */
1120 extern void lock_super(struct super_block
*);
1121 extern void unlock_super(struct super_block
*);
1124 * VFS helper functions..
1126 extern int vfs_permission(struct nameidata
*, int);
1127 extern int vfs_create(struct inode
*, struct dentry
*, int, struct nameidata
*);
1128 extern int vfs_mkdir(struct inode
*, struct dentry
*, int);
1129 extern int vfs_mknod(struct inode
*, struct dentry
*, int, dev_t
);
1130 extern int vfs_symlink(struct inode
*, struct dentry
*, const char *, int);
1131 extern int vfs_link(struct dentry
*, struct inode
*, struct dentry
*);
1132 extern int vfs_rmdir(struct inode
*, struct dentry
*);
1133 extern int vfs_unlink(struct inode
*, struct dentry
*);
1134 extern int vfs_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
1137 * VFS dentry helper functions.
1139 extern void dentry_unhash(struct dentry
*dentry
);
1142 * VFS file helper functions.
1144 extern int file_permission(struct file
*, int);
1149 * NOTE! These match bits 12..15 of stat.st_mode
1150 * (ie "(i_mode >> 12) & 15").
1152 #define DT_UNKNOWN 0
1162 #define OSYNC_METADATA (1<<0)
1163 #define OSYNC_DATA (1<<1)
1164 #define OSYNC_INODE (1<<2)
1165 int generic_osync_inode(struct inode
*, struct address_space
*, int);
1168 * This is the "filldir" function type, used by readdir() to let
1169 * the kernel specify what kind of dirent layout it wants to have.
1170 * This allows the kernel to read directories into kernel space or
1171 * to have different dirent layouts depending on the binary type.
1173 typedef int (*filldir_t
)(void *, const char *, int, loff_t
, u64
, unsigned);
1175 struct block_device_operations
{
1176 int (*open
) (struct inode
*, struct file
*);
1177 int (*release
) (struct inode
*, struct file
*);
1178 int (*ioctl
) (struct inode
*, struct file
*, unsigned, unsigned long);
1179 long (*unlocked_ioctl
) (struct file
*, unsigned, unsigned long);
1180 long (*compat_ioctl
) (struct file
*, unsigned, unsigned long);
1181 int (*direct_access
) (struct block_device
*, sector_t
,
1182 void **, unsigned long *);
1183 int (*media_changed
) (struct gendisk
*);
1184 int (*revalidate_disk
) (struct gendisk
*);
1185 int (*getgeo
)(struct block_device
*, struct hd_geometry
*);
1186 struct module
*owner
;
1190 * "descriptor" for what we're up to with a read.
1191 * This allows us to use the same read code yet
1192 * have multiple different users of the data that
1193 * we read from a file.
1195 * The simplest case just copies the data to user
1206 } read_descriptor_t
;
1208 typedef int (*read_actor_t
)(read_descriptor_t
*, struct page
*, unsigned long, unsigned long);
1210 /* These macros are for out of kernel modules to test that
1211 * the kernel supports the unlocked_ioctl and compat_ioctl
1212 * fields in struct file_operations. */
1213 #define HAVE_COMPAT_IOCTL 1
1214 #define HAVE_UNLOCKED_IOCTL 1
1218 * read, write, poll, fsync, readv, writev, unlocked_ioctl and compat_ioctl
1219 * can be called without the big kernel lock held in all filesystems.
1221 struct file_operations
{
1222 struct module
*owner
;
1223 loff_t (*llseek
) (struct file
*, loff_t
, int);
1224 ssize_t (*read
) (struct file
*, char __user
*, size_t, loff_t
*);
1225 ssize_t (*write
) (struct file
*, const char __user
*, size_t, loff_t
*);
1226 ssize_t (*aio_read
) (struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1227 ssize_t (*aio_write
) (struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1228 int (*readdir
) (struct file
*, void *, filldir_t
);
1229 unsigned int (*poll
) (struct file
*, struct poll_table_struct
*);
1230 int (*ioctl
) (struct inode
*, struct file
*, unsigned int, unsigned long);
1231 long (*unlocked_ioctl
) (struct file
*, unsigned int, unsigned long);
1232 long (*compat_ioctl
) (struct file
*, unsigned int, unsigned long);
1233 int (*mmap
) (struct file
*, struct vm_area_struct
*);
1234 int (*open
) (struct inode
*, struct file
*);
1235 int (*flush
) (struct file
*, fl_owner_t id
);
1236 int (*release
) (struct inode
*, struct file
*);
1237 int (*fsync
) (struct file
*, struct dentry
*, int datasync
);
1238 int (*aio_fsync
) (struct kiocb
*, int datasync
);
1239 int (*fasync
) (int, struct file
*, int);
1240 int (*lock
) (struct file
*, int, struct file_lock
*);
1241 ssize_t (*sendpage
) (struct file
*, struct page
*, int, size_t, loff_t
*, int);
1242 unsigned long (*get_unmapped_area
)(struct file
*, unsigned long, unsigned long, unsigned long, unsigned long);
1243 int (*check_flags
)(int);
1244 int (*dir_notify
)(struct file
*filp
, unsigned long arg
);
1245 int (*flock
) (struct file
*, int, struct file_lock
*);
1246 ssize_t (*splice_write
)(struct pipe_inode_info
*, struct file
*, loff_t
*, size_t, unsigned int);
1247 ssize_t (*splice_read
)(struct file
*, loff_t
*, struct pipe_inode_info
*, size_t, unsigned int);
1248 int (*setlease
)(struct file
*, long, struct file_lock
**);
1251 struct inode_operations
{
1252 int (*create
) (struct inode
*,struct dentry
*,int, struct nameidata
*);
1253 struct dentry
* (*lookup
) (struct inode
*,struct dentry
*, struct nameidata
*);
1254 int (*link
) (struct dentry
*,struct inode
*,struct dentry
*);
1255 int (*unlink
) (struct inode
*,struct dentry
*);
1256 int (*symlink
) (struct inode
*,struct dentry
*,const char *);
1257 int (*mkdir
) (struct inode
*,struct dentry
*,int);
1258 int (*rmdir
) (struct inode
*,struct dentry
*);
1259 int (*mknod
) (struct inode
*,struct dentry
*,int,dev_t
);
1260 int (*rename
) (struct inode
*, struct dentry
*,
1261 struct inode
*, struct dentry
*);
1262 int (*readlink
) (struct dentry
*, char __user
*,int);
1263 void * (*follow_link
) (struct dentry
*, struct nameidata
*);
1264 void (*put_link
) (struct dentry
*, struct nameidata
*, void *);
1265 void (*truncate
) (struct inode
*);
1266 int (*permission
) (struct inode
*, int, struct nameidata
*);
1267 int (*setattr
) (struct dentry
*, struct iattr
*);
1268 int (*getattr
) (struct vfsmount
*mnt
, struct dentry
*, struct kstat
*);
1269 int (*setxattr
) (struct dentry
*, const char *,const void *,size_t,int);
1270 ssize_t (*getxattr
) (struct dentry
*, const char *, void *, size_t);
1271 ssize_t (*listxattr
) (struct dentry
*, char *, size_t);
1272 int (*removexattr
) (struct dentry
*, const char *);
1273 void (*truncate_range
)(struct inode
*, loff_t
, loff_t
);
1274 long (*fallocate
)(struct inode
*inode
, int mode
, loff_t offset
,
1280 ssize_t
rw_copy_check_uvector(int type
, const struct iovec __user
* uvector
,
1281 unsigned long nr_segs
, unsigned long fast_segs
,
1282 struct iovec
*fast_pointer
,
1283 struct iovec
**ret_pointer
);
1285 extern ssize_t
vfs_read(struct file
*, char __user
*, size_t, loff_t
*);
1286 extern ssize_t
vfs_write(struct file
*, const char __user
*, size_t, loff_t
*);
1287 extern ssize_t
vfs_readv(struct file
*, const struct iovec __user
*,
1288 unsigned long, loff_t
*);
1289 extern ssize_t
vfs_writev(struct file
*, const struct iovec __user
*,
1290 unsigned long, loff_t
*);
1292 struct super_operations
{
1293 struct inode
*(*alloc_inode
)(struct super_block
*sb
);
1294 void (*destroy_inode
)(struct inode
*);
1296 void (*dirty_inode
) (struct inode
*);
1297 int (*write_inode
) (struct inode
*, int);
1298 void (*drop_inode
) (struct inode
*);
1299 void (*delete_inode
) (struct inode
*);
1300 void (*put_super
) (struct super_block
*);
1301 void (*write_super
) (struct super_block
*);
1302 int (*sync_fs
)(struct super_block
*sb
, int wait
);
1303 void (*write_super_lockfs
) (struct super_block
*);
1304 void (*unlockfs
) (struct super_block
*);
1305 int (*statfs
) (struct dentry
*, struct kstatfs
*);
1306 int (*remount_fs
) (struct super_block
*, int *, char *);
1307 void (*clear_inode
) (struct inode
*);
1308 void (*umount_begin
) (struct super_block
*);
1310 int (*show_options
)(struct seq_file
*, struct vfsmount
*);
1311 int (*show_stats
)(struct seq_file
*, struct vfsmount
*);
1313 ssize_t (*quota_read
)(struct super_block
*, int, char *, size_t, loff_t
);
1314 ssize_t (*quota_write
)(struct super_block
*, int, const char *, size_t, loff_t
);
1319 * Inode state bits. Protected by inode_lock.
1321 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1322 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1324 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1325 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1326 * various stages of removing an inode.
1328 * Two bits are used for locking and completion notification, I_LOCK and I_SYNC.
1330 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1331 * fdatasync(). i_atime is the usual cause.
1332 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1333 * these changes separately from I_DIRTY_SYNC so that we
1334 * don't have to write inode on fdatasync() when only
1335 * mtime has changed in it.
1336 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1337 * I_NEW get_new_inode() sets i_state to I_LOCK|I_NEW. Both
1338 * are cleared by unlock_new_inode(), called from iget().
1339 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1340 * is zero. I_FREEING must be set when I_WILL_FREE is
1342 * I_FREEING Set when inode is about to be freed but still has dirty
1343 * pages or buffers attached or the inode itself is still
1345 * I_CLEAR Set by clear_inode(). In this state the inode is clean
1346 * and can be destroyed.
1348 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1349 * prohibited for many purposes. iget() must wait for
1350 * the inode to be completely released, then create it
1351 * anew. Other functions will just ignore such inodes,
1352 * if appropriate. I_LOCK is used for waiting.
1354 * I_LOCK Serves as both a mutex and completion notification.
1355 * New inodes set I_LOCK. If two processes both create
1356 * the same inode, one of them will release its inode and
1357 * wait for I_LOCK to be released before returning.
1358 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1359 * also cause waiting on I_LOCK, without I_LOCK actually
1360 * being set. find_inode() uses this to prevent returning
1361 * nearly-dead inodes.
1362 * I_SYNC Similar to I_LOCK, but limited in scope to writeback
1363 * of inode dirty data. Having a separate lock for this
1364 * purpose reduces latency and prevents some filesystem-
1365 * specific deadlocks.
1367 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1368 * Q: igrab() only checks on (I_FREEING|I_WILL_FREE). Should it also check on
1369 * I_CLEAR? If not, why?
1371 #define I_DIRTY_SYNC 1
1372 #define I_DIRTY_DATASYNC 2
1373 #define I_DIRTY_PAGES 4
1375 #define I_WILL_FREE 16
1376 #define I_FREEING 32
1379 #define I_LOCK (1 << __I_LOCK)
1381 #define I_SYNC (1 << __I_SYNC)
1383 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1385 extern void __mark_inode_dirty(struct inode
*, int);
1386 static inline void mark_inode_dirty(struct inode
*inode
)
1388 __mark_inode_dirty(inode
, I_DIRTY
);
1391 static inline void mark_inode_dirty_sync(struct inode
*inode
)
1393 __mark_inode_dirty(inode
, I_DIRTY_SYNC
);
1397 * inc_nlink - directly increment an inode's link count
1400 * This is a low-level filesystem helper to replace any
1401 * direct filesystem manipulation of i_nlink. Currently,
1402 * it is only here for parity with dec_nlink().
1404 static inline void inc_nlink(struct inode
*inode
)
1409 static inline void inode_inc_link_count(struct inode
*inode
)
1412 mark_inode_dirty(inode
);
1416 * drop_nlink - directly drop an inode's link count
1419 * This is a low-level filesystem helper to replace any
1420 * direct filesystem manipulation of i_nlink. In cases
1421 * where we are attempting to track writes to the
1422 * filesystem, a decrement to zero means an imminent
1423 * write when the file is truncated and actually unlinked
1424 * on the filesystem.
1426 static inline void drop_nlink(struct inode
*inode
)
1432 * clear_nlink - directly zero an inode's link count
1435 * This is a low-level filesystem helper to replace any
1436 * direct filesystem manipulation of i_nlink. See
1437 * drop_nlink() for why we care about i_nlink hitting zero.
1439 static inline void clear_nlink(struct inode
*inode
)
1444 static inline void inode_dec_link_count(struct inode
*inode
)
1447 mark_inode_dirty(inode
);
1451 * inode_inc_iversion - increments i_version
1452 * @inode: inode that need to be updated
1454 * Every time the inode is modified, the i_version field will be incremented.
1455 * The filesystem has to be mounted with i_version flag
1458 static inline void inode_inc_iversion(struct inode
*inode
)
1460 spin_lock(&inode
->i_lock
);
1462 spin_unlock(&inode
->i_lock
);
1465 extern void touch_atime(struct vfsmount
*mnt
, struct dentry
*dentry
);
1466 static inline void file_accessed(struct file
*file
)
1468 if (!(file
->f_flags
& O_NOATIME
))
1469 touch_atime(file
->f_path
.mnt
, file
->f_path
.dentry
);
1472 int sync_inode(struct inode
*inode
, struct writeback_control
*wbc
);
1474 struct file_system_type
{
1477 int (*get_sb
) (struct file_system_type
*, int,
1478 const char *, void *, struct vfsmount
*);
1479 void (*kill_sb
) (struct super_block
*);
1480 struct module
*owner
;
1481 struct file_system_type
* next
;
1482 struct list_head fs_supers
;
1484 struct lock_class_key s_lock_key
;
1485 struct lock_class_key s_umount_key
;
1487 struct lock_class_key i_lock_key
;
1488 struct lock_class_key i_mutex_key
;
1489 struct lock_class_key i_mutex_dir_key
;
1490 struct lock_class_key i_alloc_sem_key
;
1493 extern int get_sb_bdev(struct file_system_type
*fs_type
,
1494 int flags
, const char *dev_name
, void *data
,
1495 int (*fill_super
)(struct super_block
*, void *, int),
1496 struct vfsmount
*mnt
);
1497 extern int get_sb_single(struct file_system_type
*fs_type
,
1498 int flags
, void *data
,
1499 int (*fill_super
)(struct super_block
*, void *, int),
1500 struct vfsmount
*mnt
);
1501 extern int get_sb_nodev(struct file_system_type
*fs_type
,
1502 int flags
, void *data
,
1503 int (*fill_super
)(struct super_block
*, void *, int),
1504 struct vfsmount
*mnt
);
1505 void generic_shutdown_super(struct super_block
*sb
);
1506 void kill_block_super(struct super_block
*sb
);
1507 void kill_anon_super(struct super_block
*sb
);
1508 void kill_litter_super(struct super_block
*sb
);
1509 void deactivate_super(struct super_block
*sb
);
1510 int set_anon_super(struct super_block
*s
, void *data
);
1511 struct super_block
*sget(struct file_system_type
*type
,
1512 int (*test
)(struct super_block
*,void *),
1513 int (*set
)(struct super_block
*,void *),
1515 extern int get_sb_pseudo(struct file_system_type
*, char *,
1516 const struct super_operations
*ops
, unsigned long,
1517 struct vfsmount
*mnt
);
1518 extern int simple_set_mnt(struct vfsmount
*mnt
, struct super_block
*sb
);
1519 int __put_super_and_need_restart(struct super_block
*sb
);
1520 void unnamed_dev_init(void);
1522 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1523 #define fops_get(fops) \
1524 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1525 #define fops_put(fops) \
1526 do { if (fops) module_put((fops)->owner); } while(0)
1528 extern int register_filesystem(struct file_system_type
*);
1529 extern int unregister_filesystem(struct file_system_type
*);
1530 extern struct vfsmount
*kern_mount_data(struct file_system_type
*, void *data
);
1531 #define kern_mount(type) kern_mount_data(type, NULL)
1532 extern int may_umount_tree(struct vfsmount
*);
1533 extern int may_umount(struct vfsmount
*);
1534 extern long do_mount(char *, char *, char *, unsigned long, void *);
1535 extern struct vfsmount
*collect_mounts(struct vfsmount
*, struct dentry
*);
1536 extern void drop_collected_mounts(struct vfsmount
*);
1538 extern int vfs_statfs(struct dentry
*, struct kstatfs
*);
1541 extern struct kobject
*fs_kobj
;
1543 #define FLOCK_VERIFY_READ 1
1544 #define FLOCK_VERIFY_WRITE 2
1546 extern int locks_mandatory_locked(struct inode
*);
1547 extern int locks_mandatory_area(int, struct inode
*, struct file
*, loff_t
, size_t);
1550 * Candidates for mandatory locking have the setgid bit set
1551 * but no group execute bit - an otherwise meaningless combination.
1554 static inline int __mandatory_lock(struct inode
*ino
)
1556 return (ino
->i_mode
& (S_ISGID
| S_IXGRP
)) == S_ISGID
;
1560 * ... and these candidates should be on MS_MANDLOCK mounted fs,
1561 * otherwise these will be advisory locks
1564 static inline int mandatory_lock(struct inode
*ino
)
1566 return IS_MANDLOCK(ino
) && __mandatory_lock(ino
);
1569 static inline int locks_verify_locked(struct inode
*inode
)
1571 if (mandatory_lock(inode
))
1572 return locks_mandatory_locked(inode
);
1576 extern int rw_verify_area(int, struct file
*, loff_t
*, size_t);
1578 static inline int locks_verify_truncate(struct inode
*inode
,
1582 if (inode
->i_flock
&& mandatory_lock(inode
))
1583 return locks_mandatory_area(
1584 FLOCK_VERIFY_WRITE
, inode
, filp
,
1585 size
< inode
->i_size
? size
: inode
->i_size
,
1586 (size
< inode
->i_size
? inode
->i_size
- size
1587 : size
- inode
->i_size
)
1592 static inline int break_lease(struct inode
*inode
, unsigned int mode
)
1595 return __break_lease(inode
, mode
);
1601 extern int do_truncate(struct dentry
*, loff_t start
, unsigned int time_attrs
,
1603 extern long do_sys_open(int dfd
, const char __user
*filename
, int flags
,
1605 extern struct file
*filp_open(const char *, int, int);
1606 extern struct file
* dentry_open(struct dentry
*, struct vfsmount
*, int);
1607 extern int filp_close(struct file
*, fl_owner_t id
);
1608 extern char * getname(const char __user
*);
1611 extern void __init
vfs_caches_init_early(void);
1612 extern void __init
vfs_caches_init(unsigned long);
1614 extern struct kmem_cache
*names_cachep
;
1616 #define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL)
1617 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
1618 #ifndef CONFIG_AUDITSYSCALL
1619 #define putname(name) __putname(name)
1621 extern void putname(const char *name
);
1625 extern int register_blkdev(unsigned int, const char *);
1626 extern void unregister_blkdev(unsigned int, const char *);
1627 extern struct block_device
*bdget(dev_t
);
1628 extern void bd_set_size(struct block_device
*, loff_t size
);
1629 extern void bd_forget(struct inode
*inode
);
1630 extern void bdput(struct block_device
*);
1631 extern struct block_device
*open_by_devnum(dev_t
, unsigned);
1633 static inline void bd_forget(struct inode
*inode
) {}
1635 extern const struct file_operations def_blk_fops
;
1636 extern const struct file_operations def_chr_fops
;
1637 extern const struct file_operations bad_sock_fops
;
1638 extern const struct file_operations def_fifo_fops
;
1640 extern int ioctl_by_bdev(struct block_device
*, unsigned, unsigned long);
1641 extern int blkdev_ioctl(struct inode
*, struct file
*, unsigned, unsigned long);
1642 extern int blkdev_driver_ioctl(struct inode
*inode
, struct file
*file
,
1643 struct gendisk
*disk
, unsigned cmd
,
1645 extern long compat_blkdev_ioctl(struct file
*, unsigned, unsigned long);
1646 extern int blkdev_get(struct block_device
*, mode_t
, unsigned);
1647 extern int blkdev_put(struct block_device
*);
1648 extern int bd_claim(struct block_device
*, void *);
1649 extern void bd_release(struct block_device
*);
1651 extern int bd_claim_by_disk(struct block_device
*, void *, struct gendisk
*);
1652 extern void bd_release_from_disk(struct block_device
*, struct gendisk
*);
1654 #define bd_claim_by_disk(bdev, holder, disk) bd_claim(bdev, holder)
1655 #define bd_release_from_disk(bdev, disk) bd_release(bdev)
1660 #define CHRDEV_MAJOR_HASH_SIZE 255
1661 extern int alloc_chrdev_region(dev_t
*, unsigned, unsigned, const char *);
1662 extern int register_chrdev_region(dev_t
, unsigned, const char *);
1663 extern int register_chrdev(unsigned int, const char *,
1664 const struct file_operations
*);
1665 extern void unregister_chrdev(unsigned int, const char *);
1666 extern void unregister_chrdev_region(dev_t
, unsigned);
1667 extern void chrdev_show(struct seq_file
*,off_t
);
1669 /* fs/block_dev.c */
1670 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
1673 #define BLKDEV_MAJOR_HASH_SIZE 255
1674 extern const char *__bdevname(dev_t
, char *buffer
);
1675 extern const char *bdevname(struct block_device
*bdev
, char *buffer
);
1676 extern struct block_device
*lookup_bdev(const char *);
1677 extern struct block_device
*open_bdev_excl(const char *, int, void *);
1678 extern void close_bdev_excl(struct block_device
*);
1679 extern void blkdev_show(struct seq_file
*,off_t
);
1681 #define BLKDEV_MAJOR_HASH_SIZE 0
1684 extern void init_special_inode(struct inode
*, umode_t
, dev_t
);
1686 /* Invalid inode operations -- fs/bad_inode.c */
1687 extern void make_bad_inode(struct inode
*);
1688 extern int is_bad_inode(struct inode
*);
1690 extern const struct file_operations read_fifo_fops
;
1691 extern const struct file_operations write_fifo_fops
;
1692 extern const struct file_operations rdwr_fifo_fops
;
1694 extern int fs_may_remount_ro(struct super_block
*);
1698 * return READ, READA, or WRITE
1700 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
1703 * return data direction, READ or WRITE
1705 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
1707 extern int check_disk_change(struct block_device
*);
1708 extern int __invalidate_device(struct block_device
*);
1709 extern int invalidate_partition(struct gendisk
*, int);
1711 extern int invalidate_inodes(struct super_block
*);
1712 unsigned long __invalidate_mapping_pages(struct address_space
*mapping
,
1713 pgoff_t start
, pgoff_t end
,
1715 unsigned long invalidate_mapping_pages(struct address_space
*mapping
,
1716 pgoff_t start
, pgoff_t end
);
1718 static inline unsigned long __deprecated
1719 invalidate_inode_pages(struct address_space
*mapping
)
1721 return invalidate_mapping_pages(mapping
, 0, ~0UL);
1724 static inline void invalidate_remote_inode(struct inode
*inode
)
1726 if (S_ISREG(inode
->i_mode
) || S_ISDIR(inode
->i_mode
) ||
1727 S_ISLNK(inode
->i_mode
))
1728 invalidate_mapping_pages(inode
->i_mapping
, 0, -1);
1730 extern int invalidate_inode_pages2(struct address_space
*mapping
);
1731 extern int invalidate_inode_pages2_range(struct address_space
*mapping
,
1732 pgoff_t start
, pgoff_t end
);
1733 extern int write_inode_now(struct inode
*, int);
1734 extern int filemap_fdatawrite(struct address_space
*);
1735 extern int filemap_flush(struct address_space
*);
1736 extern int filemap_fdatawait(struct address_space
*);
1737 extern int filemap_write_and_wait(struct address_space
*mapping
);
1738 extern int filemap_write_and_wait_range(struct address_space
*mapping
,
1739 loff_t lstart
, loff_t lend
);
1740 extern int wait_on_page_writeback_range(struct address_space
*mapping
,
1741 pgoff_t start
, pgoff_t end
);
1742 extern int __filemap_fdatawrite_range(struct address_space
*mapping
,
1743 loff_t start
, loff_t end
, int sync_mode
);
1745 extern long do_fsync(struct file
*file
, int datasync
);
1746 extern void sync_supers(void);
1747 extern void sync_filesystems(int wait
);
1748 extern void __fsync_super(struct super_block
*sb
);
1749 extern void emergency_sync(void);
1750 extern void emergency_remount(void);
1751 extern int do_remount_sb(struct super_block
*sb
, int flags
,
1752 void *data
, int force
);
1754 extern sector_t
bmap(struct inode
*, sector_t
);
1756 extern int notify_change(struct dentry
*, struct iattr
*);
1757 extern int permission(struct inode
*, int, struct nameidata
*);
1758 extern int generic_permission(struct inode
*, int,
1759 int (*check_acl
)(struct inode
*, int));
1761 extern int get_write_access(struct inode
*);
1762 extern int deny_write_access(struct file
*);
1763 static inline void put_write_access(struct inode
* inode
)
1765 atomic_dec(&inode
->i_writecount
);
1767 static inline void allow_write_access(struct file
*file
)
1770 atomic_inc(&file
->f_path
.dentry
->d_inode
->i_writecount
);
1772 extern int do_pipe(int *);
1773 extern struct file
*create_read_pipe(struct file
*f
);
1774 extern struct file
*create_write_pipe(void);
1775 extern void free_write_pipe(struct file
*);
1777 extern struct file
*do_filp_open(int dfd
, const char *pathname
,
1778 int open_flag
, int mode
);
1779 extern int may_open(struct nameidata
*, int, int);
1781 extern int kernel_read(struct file
*, unsigned long, char *, unsigned long);
1782 extern struct file
* open_exec(const char *);
1784 /* fs/dcache.c -- generic fs support functions */
1785 extern int is_subdir(struct dentry
*, struct dentry
*);
1786 extern ino_t
find_inode_number(struct dentry
*, struct qstr
*);
1788 #include <linux/err.h>
1790 /* needed for stackable file system support */
1791 extern loff_t
default_llseek(struct file
*file
, loff_t offset
, int origin
);
1793 extern loff_t
vfs_llseek(struct file
*file
, loff_t offset
, int origin
);
1795 extern void inode_init_once(struct inode
*);
1796 extern void iput(struct inode
*);
1797 extern struct inode
* igrab(struct inode
*);
1798 extern ino_t
iunique(struct super_block
*, ino_t
);
1799 extern int inode_needs_sync(struct inode
*inode
);
1800 extern void generic_delete_inode(struct inode
*inode
);
1801 extern void generic_drop_inode(struct inode
*inode
);
1803 extern struct inode
*ilookup5_nowait(struct super_block
*sb
,
1804 unsigned long hashval
, int (*test
)(struct inode
*, void *),
1806 extern struct inode
*ilookup5(struct super_block
*sb
, unsigned long hashval
,
1807 int (*test
)(struct inode
*, void *), void *data
);
1808 extern struct inode
*ilookup(struct super_block
*sb
, unsigned long ino
);
1810 extern struct inode
* iget5_locked(struct super_block
*, unsigned long, int (*test
)(struct inode
*, void *), int (*set
)(struct inode
*, void *), void *);
1811 extern struct inode
* iget_locked(struct super_block
*, unsigned long);
1812 extern void unlock_new_inode(struct inode
*);
1814 extern void __iget(struct inode
* inode
);
1815 extern void iget_failed(struct inode
*);
1816 extern void clear_inode(struct inode
*);
1817 extern void destroy_inode(struct inode
*);
1818 extern struct inode
*new_inode(struct super_block
*);
1819 extern int should_remove_suid(struct dentry
*);
1820 extern int remove_suid(struct dentry
*);
1822 extern void __insert_inode_hash(struct inode
*, unsigned long hashval
);
1823 extern void remove_inode_hash(struct inode
*);
1824 static inline void insert_inode_hash(struct inode
*inode
) {
1825 __insert_inode_hash(inode
, inode
->i_ino
);
1828 extern struct file
* get_empty_filp(void);
1829 extern void file_move(struct file
*f
, struct list_head
*list
);
1830 extern void file_kill(struct file
*f
);
1833 extern void submit_bio(int, struct bio
*);
1834 extern int bdev_read_only(struct block_device
*);
1836 extern int set_blocksize(struct block_device
*, int);
1837 extern int sb_set_blocksize(struct super_block
*, int);
1838 extern int sb_min_blocksize(struct super_block
*, int);
1839 extern int sb_has_dirty_inodes(struct super_block
*);
1841 extern int generic_file_mmap(struct file
*, struct vm_area_struct
*);
1842 extern int generic_file_readonly_mmap(struct file
*, struct vm_area_struct
*);
1843 extern int file_read_actor(read_descriptor_t
* desc
, struct page
*page
, unsigned long offset
, unsigned long size
);
1844 int generic_write_checks(struct file
*file
, loff_t
*pos
, size_t *count
, int isblk
);
1845 extern ssize_t
generic_file_aio_read(struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1846 extern ssize_t
generic_file_aio_write(struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1847 extern ssize_t
generic_file_aio_write_nolock(struct kiocb
*, const struct iovec
*,
1848 unsigned long, loff_t
);
1849 extern ssize_t
generic_file_direct_write(struct kiocb
*, const struct iovec
*,
1850 unsigned long *, loff_t
, loff_t
*, size_t, size_t);
1851 extern ssize_t
generic_file_buffered_write(struct kiocb
*, const struct iovec
*,
1852 unsigned long, loff_t
, loff_t
*, size_t, ssize_t
);
1853 extern ssize_t
do_sync_read(struct file
*filp
, char __user
*buf
, size_t len
, loff_t
*ppos
);
1854 extern ssize_t
do_sync_write(struct file
*filp
, const char __user
*buf
, size_t len
, loff_t
*ppos
);
1855 extern int generic_segment_checks(const struct iovec
*iov
,
1856 unsigned long *nr_segs
, size_t *count
, int access_flags
);
1859 extern ssize_t
generic_file_splice_read(struct file
*, loff_t
*,
1860 struct pipe_inode_info
*, size_t, unsigned int);
1861 extern ssize_t
generic_file_splice_write(struct pipe_inode_info
*,
1862 struct file
*, loff_t
*, size_t, unsigned int);
1863 extern ssize_t
generic_file_splice_write_nolock(struct pipe_inode_info
*,
1864 struct file
*, loff_t
*, size_t, unsigned int);
1865 extern ssize_t
generic_splice_sendpage(struct pipe_inode_info
*pipe
,
1866 struct file
*out
, loff_t
*, size_t len
, unsigned int flags
);
1867 extern long do_splice_direct(struct file
*in
, loff_t
*ppos
, struct file
*out
,
1868 size_t len
, unsigned int flags
);
1871 file_ra_state_init(struct file_ra_state
*ra
, struct address_space
*mapping
);
1872 extern loff_t
no_llseek(struct file
*file
, loff_t offset
, int origin
);
1873 extern loff_t
generic_file_llseek(struct file
*file
, loff_t offset
, int origin
);
1874 extern loff_t
remote_llseek(struct file
*file
, loff_t offset
, int origin
);
1875 extern int generic_file_open(struct inode
* inode
, struct file
* filp
);
1876 extern int nonseekable_open(struct inode
* inode
, struct file
* filp
);
1878 #ifdef CONFIG_FS_XIP
1879 extern ssize_t
xip_file_read(struct file
*filp
, char __user
*buf
, size_t len
,
1881 extern int xip_file_mmap(struct file
* file
, struct vm_area_struct
* vma
);
1882 extern ssize_t
xip_file_write(struct file
*filp
, const char __user
*buf
,
1883 size_t len
, loff_t
*ppos
);
1884 extern int xip_truncate_page(struct address_space
*mapping
, loff_t from
);
1886 static inline int xip_truncate_page(struct address_space
*mapping
, loff_t from
)
1893 ssize_t
__blockdev_direct_IO(int rw
, struct kiocb
*iocb
, struct inode
*inode
,
1894 struct block_device
*bdev
, const struct iovec
*iov
, loff_t offset
,
1895 unsigned long nr_segs
, get_block_t get_block
, dio_iodone_t end_io
,
1899 DIO_LOCKING
= 1, /* need locking between buffered and direct access */
1900 DIO_NO_LOCKING
, /* bdev; no locking at all between buffered/direct */
1901 DIO_OWN_LOCKING
, /* filesystem locks buffered and direct internally */
1904 static inline ssize_t
blockdev_direct_IO(int rw
, struct kiocb
*iocb
,
1905 struct inode
*inode
, struct block_device
*bdev
, const struct iovec
*iov
,
1906 loff_t offset
, unsigned long nr_segs
, get_block_t get_block
,
1907 dio_iodone_t end_io
)
1909 return __blockdev_direct_IO(rw
, iocb
, inode
, bdev
, iov
, offset
,
1910 nr_segs
, get_block
, end_io
, DIO_LOCKING
);
1913 static inline ssize_t
blockdev_direct_IO_no_locking(int rw
, struct kiocb
*iocb
,
1914 struct inode
*inode
, struct block_device
*bdev
, const struct iovec
*iov
,
1915 loff_t offset
, unsigned long nr_segs
, get_block_t get_block
,
1916 dio_iodone_t end_io
)
1918 return __blockdev_direct_IO(rw
, iocb
, inode
, bdev
, iov
, offset
,
1919 nr_segs
, get_block
, end_io
, DIO_NO_LOCKING
);
1922 static inline ssize_t
blockdev_direct_IO_own_locking(int rw
, struct kiocb
*iocb
,
1923 struct inode
*inode
, struct block_device
*bdev
, const struct iovec
*iov
,
1924 loff_t offset
, unsigned long nr_segs
, get_block_t get_block
,
1925 dio_iodone_t end_io
)
1927 return __blockdev_direct_IO(rw
, iocb
, inode
, bdev
, iov
, offset
,
1928 nr_segs
, get_block
, end_io
, DIO_OWN_LOCKING
);
1932 extern const struct file_operations generic_ro_fops
;
1934 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
1936 extern int vfs_readlink(struct dentry
*, char __user
*, int, const char *);
1937 extern int vfs_follow_link(struct nameidata
*, const char *);
1938 extern int page_readlink(struct dentry
*, char __user
*, int);
1939 extern void *page_follow_link_light(struct dentry
*, struct nameidata
*);
1940 extern void page_put_link(struct dentry
*, struct nameidata
*, void *);
1941 extern int __page_symlink(struct inode
*inode
, const char *symname
, int len
,
1943 extern int page_symlink(struct inode
*inode
, const char *symname
, int len
);
1944 extern const struct inode_operations page_symlink_inode_operations
;
1945 extern int generic_readlink(struct dentry
*, char __user
*, int);
1946 extern void generic_fillattr(struct inode
*, struct kstat
*);
1947 extern int vfs_getattr(struct vfsmount
*, struct dentry
*, struct kstat
*);
1948 void inode_add_bytes(struct inode
*inode
, loff_t bytes
);
1949 void inode_sub_bytes(struct inode
*inode
, loff_t bytes
);
1950 loff_t
inode_get_bytes(struct inode
*inode
);
1951 void inode_set_bytes(struct inode
*inode
, loff_t bytes
);
1953 extern int vfs_readdir(struct file
*, filldir_t
, void *);
1955 extern int vfs_stat(char __user
*, struct kstat
*);
1956 extern int vfs_lstat(char __user
*, struct kstat
*);
1957 extern int vfs_stat_fd(int dfd
, char __user
*, struct kstat
*);
1958 extern int vfs_lstat_fd(int dfd
, char __user
*, struct kstat
*);
1959 extern int vfs_fstat(unsigned int, struct kstat
*);
1961 extern int do_vfs_ioctl(struct file
*filp
, unsigned int fd
, unsigned int cmd
,
1964 extern void get_filesystem(struct file_system_type
*fs
);
1965 extern void put_filesystem(struct file_system_type
*fs
);
1966 extern struct file_system_type
*get_fs_type(const char *name
);
1967 extern struct super_block
*get_super(struct block_device
*);
1968 extern struct super_block
*user_get_super(dev_t
);
1969 extern void drop_super(struct super_block
*sb
);
1971 extern int dcache_dir_open(struct inode
*, struct file
*);
1972 extern int dcache_dir_close(struct inode
*, struct file
*);
1973 extern loff_t
dcache_dir_lseek(struct file
*, loff_t
, int);
1974 extern int dcache_readdir(struct file
*, void *, filldir_t
);
1975 extern int simple_getattr(struct vfsmount
*, struct dentry
*, struct kstat
*);
1976 extern int simple_statfs(struct dentry
*, struct kstatfs
*);
1977 extern int simple_link(struct dentry
*, struct inode
*, struct dentry
*);
1978 extern int simple_unlink(struct inode
*, struct dentry
*);
1979 extern int simple_rmdir(struct inode
*, struct dentry
*);
1980 extern int simple_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
1981 extern int simple_sync_file(struct file
*, struct dentry
*, int);
1982 extern int simple_empty(struct dentry
*);
1983 extern int simple_readpage(struct file
*file
, struct page
*page
);
1984 extern int simple_prepare_write(struct file
*file
, struct page
*page
,
1985 unsigned offset
, unsigned to
);
1986 extern int simple_write_begin(struct file
*file
, struct address_space
*mapping
,
1987 loff_t pos
, unsigned len
, unsigned flags
,
1988 struct page
**pagep
, void **fsdata
);
1989 extern int simple_write_end(struct file
*file
, struct address_space
*mapping
,
1990 loff_t pos
, unsigned len
, unsigned copied
,
1991 struct page
*page
, void *fsdata
);
1993 extern struct dentry
*simple_lookup(struct inode
*, struct dentry
*, struct nameidata
*);
1994 extern ssize_t
generic_read_dir(struct file
*, char __user
*, size_t, loff_t
*);
1995 extern const struct file_operations simple_dir_operations
;
1996 extern const struct inode_operations simple_dir_inode_operations
;
1997 struct tree_descr
{ char *name
; const struct file_operations
*ops
; int mode
; };
1998 struct dentry
*d_alloc_name(struct dentry
*, const char *);
1999 extern int simple_fill_super(struct super_block
*, int, struct tree_descr
*);
2000 extern int simple_pin_fs(struct file_system_type
*, struct vfsmount
**mount
, int *count
);
2001 extern void simple_release_fs(struct vfsmount
**mount
, int *count
);
2003 extern ssize_t
simple_read_from_buffer(void __user
*to
, size_t count
,
2004 loff_t
*ppos
, const void *from
, size_t available
);
2005 extern ssize_t
memory_read_from_buffer(void *to
, size_t count
, loff_t
*ppos
,
2006 const void *from
, size_t available
);
2008 #ifdef CONFIG_MIGRATION
2009 extern int buffer_migrate_page(struct address_space
*,
2010 struct page
*, struct page
*);
2012 #define buffer_migrate_page NULL
2015 extern int inode_change_ok(struct inode
*, struct iattr
*);
2016 extern int __must_check
inode_setattr(struct inode
*, struct iattr
*);
2018 extern void file_update_time(struct file
*file
);
2020 extern int generic_show_options(struct seq_file
*m
, struct vfsmount
*mnt
);
2021 extern void save_mount_options(struct super_block
*sb
, char *options
);
2023 static inline ino_t
parent_ino(struct dentry
*dentry
)
2027 spin_lock(&dentry
->d_lock
);
2028 res
= dentry
->d_parent
->d_inode
->i_ino
;
2029 spin_unlock(&dentry
->d_lock
);
2033 /* Transaction based IO helpers */
2036 * An argresp is stored in an allocated page and holds the
2037 * size of the argument or response, along with its content
2039 struct simple_transaction_argresp
{
2044 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2046 char *simple_transaction_get(struct file
*file
, const char __user
*buf
,
2048 ssize_t
simple_transaction_read(struct file
*file
, char __user
*buf
,
2049 size_t size
, loff_t
*pos
);
2050 int simple_transaction_release(struct inode
*inode
, struct file
*file
);
2052 static inline void simple_transaction_set(struct file
*file
, size_t n
)
2054 struct simple_transaction_argresp
*ar
= file
->private_data
;
2056 BUG_ON(n
> SIMPLE_TRANSACTION_LIMIT
);
2059 * The barrier ensures that ar->size will really remain zero until
2060 * ar->data is ready for reading.
2067 * simple attribute files
2069 * These attributes behave similar to those in sysfs:
2071 * Writing to an attribute immediately sets a value, an open file can be
2072 * written to multiple times.
2074 * Reading from an attribute creates a buffer from the value that might get
2075 * read with multiple read calls. When the attribute has been read
2076 * completely, no further read calls are possible until the file is opened
2079 * All attributes contain a text representation of a numeric value
2080 * that are accessed with the get() and set() functions.
2082 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2083 static int __fops ## _open(struct inode *inode, struct file *file) \
2085 __simple_attr_check_format(__fmt, 0ull); \
2086 return simple_attr_open(inode, file, __get, __set, __fmt); \
2088 static struct file_operations __fops = { \
2089 .owner = THIS_MODULE, \
2090 .open = __fops ## _open, \
2091 .release = simple_attr_release, \
2092 .read = simple_attr_read, \
2093 .write = simple_attr_write, \
2096 static inline void __attribute__((format(printf
, 1, 2)))
2097 __simple_attr_check_format(const char *fmt
, ...)
2099 /* don't do anything, just let the compiler check the arguments; */
2102 int simple_attr_open(struct inode
*inode
, struct file
*file
,
2103 int (*get
)(void *, u64
*), int (*set
)(void *, u64
),
2105 int simple_attr_release(struct inode
*inode
, struct file
*file
);
2106 ssize_t
simple_attr_read(struct file
*file
, char __user
*buf
,
2107 size_t len
, loff_t
*ppos
);
2108 ssize_t
simple_attr_write(struct file
*file
, const char __user
*buf
,
2109 size_t len
, loff_t
*ppos
);
2112 #ifdef CONFIG_SECURITY
2113 static inline char *alloc_secdata(void)
2115 return (char *)get_zeroed_page(GFP_KERNEL
);
2118 static inline void free_secdata(void *secdata
)
2120 free_page((unsigned long)secdata
);
2123 static inline char *alloc_secdata(void)
2128 static inline void free_secdata(void *secdata
)
2130 #endif /* CONFIG_SECURITY */
2133 int proc_nr_files(struct ctl_table
*table
, int write
, struct file
*filp
,
2134 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2136 int get_filesystem_list(char * buf
);
2138 #endif /* __KERNEL__ */
2139 #endif /* _LINUX_FS_H */