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 SWRITE_SYNC (SWRITE | (1 << BIO_RW_SYNC))
87 #define WRITE_BARRIER ((1 << BIO_RW) | (1 << BIO_RW_BARRIER))
93 /* public flags for file_system_type */
94 #define FS_REQUIRES_DEV 1
95 #define FS_BINARY_MOUNTDATA 2
96 #define FS_HAS_SUBTYPE 4
97 #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
98 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move()
99 * during rename() internally.
103 * These are the fs-independent mount-flags: up to 32 flags are supported
105 #define MS_RDONLY 1 /* Mount read-only */
106 #define MS_NOSUID 2 /* Ignore suid and sgid bits */
107 #define MS_NODEV 4 /* Disallow access to device special files */
108 #define MS_NOEXEC 8 /* Disallow program execution */
109 #define MS_SYNCHRONOUS 16 /* Writes are synced at once */
110 #define MS_REMOUNT 32 /* Alter flags of a mounted FS */
111 #define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
112 #define MS_DIRSYNC 128 /* Directory modifications are synchronous */
113 #define MS_NOATIME 1024 /* Do not update access times. */
114 #define MS_NODIRATIME 2048 /* Do not update directory access times */
118 #define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
119 MS_VERBOSE is deprecated. */
120 #define MS_SILENT 32768
121 #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
122 #define MS_UNBINDABLE (1<<17) /* change to unbindable */
123 #define MS_PRIVATE (1<<18) /* change to private */
124 #define MS_SLAVE (1<<19) /* change to slave */
125 #define MS_SHARED (1<<20) /* change to shared */
126 #define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */
127 #define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
128 #define MS_I_VERSION (1<<23) /* Update inode I_version field */
129 #define MS_ACTIVE (1<<30)
130 #define MS_NOUSER (1<<31)
133 * Superblock flags that can be altered by MS_REMOUNT
135 #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK)
138 * Old magic mount flag and mask
140 #define MS_MGC_VAL 0xC0ED0000
141 #define MS_MGC_MSK 0xffff0000
143 /* Inode flags - they have nothing to superblock flags now */
145 #define S_SYNC 1 /* Writes are synced at once */
146 #define S_NOATIME 2 /* Do not update access times */
147 #define S_APPEND 4 /* Append-only file */
148 #define S_IMMUTABLE 8 /* Immutable file */
149 #define S_DEAD 16 /* removed, but still open directory */
150 #define S_NOQUOTA 32 /* Inode is not counted to quota */
151 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
152 #define S_NOCMTIME 128 /* Do not update file c/mtime */
153 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
154 #define S_PRIVATE 512 /* Inode is fs-internal */
157 * Note that nosuid etc flags are inode-specific: setting some file-system
158 * flags just means all the inodes inherit those flags by default. It might be
159 * possible to override it selectively if you really wanted to with some
160 * ioctl() that is not currently implemented.
162 * Exception: MS_RDONLY is always applied to the entire file system.
164 * Unfortunately, it is possible to change a filesystems flags with it mounted
165 * with files in use. This means that all of the inodes will not have their
166 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
167 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
169 #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
171 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
172 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
173 ((inode)->i_flags & S_SYNC))
174 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
175 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
176 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
177 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
178 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
180 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
181 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
182 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
183 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
185 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
186 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
187 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
188 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
190 /* the read-only stuff doesn't really belong here, but any other place is
191 probably as bad and I don't want to create yet another include file. */
193 #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
194 #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
195 #define BLKRRPART _IO(0x12,95) /* re-read partition table */
196 #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
197 #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
198 #define BLKRASET _IO(0x12,98) /* set read ahead for block device */
199 #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
200 #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
201 #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
202 #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
203 #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
204 #define BLKSSZGET _IO(0x12,104)/* get block device sector size */
206 #define BLKPG _IO(0x12,105)/* See blkpg.h */
208 /* Some people are morons. Do not use sizeof! */
210 #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
211 #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
212 /* This was here just to show that the number is taken -
213 probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
215 /* A jump here: 108-111 have been used for various private purposes. */
216 #define BLKBSZGET _IOR(0x12,112,size_t)
217 #define BLKBSZSET _IOW(0x12,113,size_t)
218 #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
219 #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
220 #define BLKTRACESTART _IO(0x12,116)
221 #define BLKTRACESTOP _IO(0x12,117)
222 #define BLKTRACETEARDOWN _IO(0x12,118)
224 #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
225 #define FIBMAP _IO(0x00,1) /* bmap access */
226 #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
228 #define FS_IOC_GETFLAGS _IOR('f', 1, long)
229 #define FS_IOC_SETFLAGS _IOW('f', 2, long)
230 #define FS_IOC_GETVERSION _IOR('v', 1, long)
231 #define FS_IOC_SETVERSION _IOW('v', 2, long)
232 #define FS_IOC32_GETFLAGS _IOR('f', 1, int)
233 #define FS_IOC32_SETFLAGS _IOW('f', 2, int)
234 #define FS_IOC32_GETVERSION _IOR('v', 1, int)
235 #define FS_IOC32_SETVERSION _IOW('v', 2, int)
238 * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS)
240 #define FS_SECRM_FL 0x00000001 /* Secure deletion */
241 #define FS_UNRM_FL 0x00000002 /* Undelete */
242 #define FS_COMPR_FL 0x00000004 /* Compress file */
243 #define FS_SYNC_FL 0x00000008 /* Synchronous updates */
244 #define FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
245 #define FS_APPEND_FL 0x00000020 /* writes to file may only append */
246 #define FS_NODUMP_FL 0x00000040 /* do not dump file */
247 #define FS_NOATIME_FL 0x00000080 /* do not update atime */
248 /* Reserved for compression usage... */
249 #define FS_DIRTY_FL 0x00000100
250 #define FS_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
251 #define FS_NOCOMP_FL 0x00000400 /* Don't compress */
252 #define FS_ECOMPR_FL 0x00000800 /* Compression error */
253 /* End compression flags --- maybe not all used */
254 #define FS_BTREE_FL 0x00001000 /* btree format dir */
255 #define FS_INDEX_FL 0x00001000 /* hash-indexed directory */
256 #define FS_IMAGIC_FL 0x00002000 /* AFS directory */
257 #define FS_JOURNAL_DATA_FL 0x00004000 /* Reserved for ext3 */
258 #define FS_NOTAIL_FL 0x00008000 /* file tail should not be merged */
259 #define FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
260 #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
261 #define FS_EXTENT_FL 0x00080000 /* Extents */
262 #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */
263 #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */
265 #define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
266 #define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
269 #define SYNC_FILE_RANGE_WAIT_BEFORE 1
270 #define SYNC_FILE_RANGE_WRITE 2
271 #define SYNC_FILE_RANGE_WAIT_AFTER 4
275 #include <linux/linkage.h>
276 #include <linux/wait.h>
277 #include <linux/types.h>
278 #include <linux/kdev_t.h>
279 #include <linux/dcache.h>
280 #include <linux/namei.h>
281 #include <linux/stat.h>
282 #include <linux/cache.h>
283 #include <linux/kobject.h>
284 #include <linux/list.h>
285 #include <linux/radix-tree.h>
286 #include <linux/prio_tree.h>
287 #include <linux/init.h>
288 #include <linux/pid.h>
289 #include <linux/mutex.h>
290 #include <linux/capability.h>
291 #include <linux/semaphore.h>
293 #include <asm/atomic.h>
294 #include <asm/byteorder.h>
296 struct export_operations;
301 struct pipe_inode_info;
302 struct poll_table_struct;
304 struct vm_area_struct;
307 extern void __init inode_init(void);
308 extern void __init inode_init_early(void);
309 extern void __init files_init(unsigned long);
312 typedef int (get_block_t)(struct inode *inode, sector_t iblock,
313 struct buffer_head *bh_result, int create);
314 typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
315 ssize_t bytes, void *private);
318 * Attribute flags. These should be or-ed together to figure out what
325 #define ATTR_ATIME 16
326 #define ATTR_MTIME 32
327 #define ATTR_CTIME 64
328 #define ATTR_ATIME_SET 128
329 #define ATTR_MTIME_SET 256
330 #define ATTR_FORCE 512 /* Not a change, but a change it */
331 #define ATTR_ATTR_FLAG 1024
332 #define ATTR_KILL_SUID 2048
333 #define ATTR_KILL_SGID 4096
334 #define ATTR_FILE 8192
335 #define ATTR_KILL_PRIV 16384
336 #define ATTR_OPEN 32768 /* Truncating from open(O_TRUNC) */
339 * This is the Inode Attributes structure, used for notify_change(). It
340 * uses the above definitions as flags, to know which values have changed.
341 * Also, in this manner, a Filesystem can look at only the values it cares
342 * about. Basically, these are the attributes that the VFS layer can
343 * request to change from the FS layer.
345 * Derek Atkins <warlord@MIT.EDU> 94-10-20
348 unsigned int ia_valid;
353 struct timespec ia_atime;
354 struct timespec ia_mtime;
355 struct timespec ia_ctime;
358 * Not an attribute, but an auxilary info for filesystems wanting to
359 * implement an ftruncate() like method. NOTE: filesystem should
360 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
362 struct file *ia_file;
366 * Includes for diskquotas.
368 #include <linux/quota.h>
371 * enum positive_aop_returns - aop return codes with specific semantics
373 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
374 * completed, that the page is still locked, and
375 * should be considered active. The VM uses this hint
376 * to return the page to the active list -- it won't
377 * be a candidate for writeback again in the near
378 * future. Other callers must be careful to unlock
379 * the page if they get this return. Returned by
382 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
383 * unlocked it and the page might have been truncated.
384 * The caller should back up to acquiring a new page and
385 * trying again. The aop will be taking reasonable
386 * precautions not to livelock. If the caller held a page
387 * reference, it should drop it before retrying. Returned
390 * address_space_operation functions return these large constants to indicate
391 * special semantics to the caller. These are much larger than the bytes in a
392 * page to allow for functions that return the number of bytes operated on in a
396 enum positive_aop_returns {
397 AOP_WRITEPAGE_ACTIVATE = 0x80000,
398 AOP_TRUNCATED_PAGE = 0x80001,
401 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
402 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
405 * oh the beauties of C type declarations.
408 struct address_space;
409 struct writeback_control;
412 const struct iovec *iov;
413 unsigned long nr_segs;
418 size_t iov_iter_copy_from_user_atomic(struct page *page,
419 struct iov_iter *i, unsigned long offset, size_t bytes);
420 size_t iov_iter_copy_from_user(struct page *page,
421 struct iov_iter *i, unsigned long offset, size_t bytes);
422 void iov_iter_advance(struct iov_iter *i, size_t bytes);
423 int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes);
424 size_t iov_iter_single_seg_count(struct iov_iter *i);
426 static inline void iov_iter_init(struct iov_iter *i,
427 const struct iovec *iov, unsigned long nr_segs,
428 size_t count, size_t written)
431 i->nr_segs = nr_segs;
433 i->count = count + written;
435 iov_iter_advance(i, written);
438 static inline size_t iov_iter_count(struct iov_iter *i)
444 struct address_space_operations {
445 int (*writepage)(struct page *page, struct writeback_control *wbc);
446 int (*readpage)(struct file *, struct page *);
447 void (*sync_page)(struct page *);
449 /* Write back some dirty pages from this mapping. */
450 int (*writepages)(struct address_space *, struct writeback_control *);
452 /* Set a page dirty. Return true if this dirtied it */
453 int (*set_page_dirty)(struct page *page);
455 int (*readpages)(struct file *filp, struct address_space *mapping,
456 struct list_head *pages, unsigned nr_pages);
459 * ext3 requires that a successful prepare_write() call be followed
460 * by a commit_write() call - they must be balanced
462 int (*prepare_write)(struct file *, struct page *, unsigned, unsigned);
463 int (*commit_write)(struct file *, struct page *, unsigned, unsigned);
465 int (*write_begin)(struct file *, struct address_space *mapping,
466 loff_t pos, unsigned len, unsigned flags,
467 struct page **pagep, void **fsdata);
468 int (*write_end)(struct file *, struct address_space *mapping,
469 loff_t pos, unsigned len, unsigned copied,
470 struct page *page, void *fsdata);
472 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
473 sector_t (*bmap)(struct address_space *, sector_t);
474 void (*invalidatepage) (struct page *, unsigned long);
475 int (*releasepage) (struct page *, gfp_t);
476 ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
477 loff_t offset, unsigned long nr_segs);
478 int (*get_xip_mem)(struct address_space *, pgoff_t, int,
479 void **, unsigned long *);
480 /* migrate the contents of a page to the specified target */
481 int (*migratepage) (struct address_space *,
482 struct page *, struct page *);
483 int (*launder_page) (struct page *);
487 * pagecache_write_begin/pagecache_write_end must be used by general code
488 * to write into the pagecache.
490 int pagecache_write_begin(struct file *, struct address_space *mapping,
491 loff_t pos, unsigned len, unsigned flags,
492 struct page **pagep, void **fsdata);
494 int pagecache_write_end(struct file *, struct address_space *mapping,
495 loff_t pos, unsigned len, unsigned copied,
496 struct page *page, void *fsdata);
498 struct backing_dev_info;
499 struct address_space {
500 struct inode *host; /* owner: inode, block_device */
501 struct radix_tree_root page_tree; /* radix tree of all pages */
502 rwlock_t tree_lock; /* and rwlock protecting it */
503 unsigned int i_mmap_writable;/* count VM_SHARED mappings */
504 struct prio_tree_root i_mmap; /* tree of private and shared mappings */
505 struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
506 spinlock_t i_mmap_lock; /* protect tree, count, list */
507 unsigned int truncate_count; /* Cover race condition with truncate */
508 unsigned long nrpages; /* number of total pages */
509 pgoff_t writeback_index;/* writeback starts here */
510 const struct address_space_operations *a_ops; /* methods */
511 unsigned long flags; /* error bits/gfp mask */
512 struct backing_dev_info *backing_dev_info; /* device readahead, etc */
513 spinlock_t private_lock; /* for use by the address_space */
514 struct list_head private_list; /* ditto */
515 struct address_space *assoc_mapping; /* ditto */
516 } __attribute__((aligned(sizeof(long))));
518 * On most architectures that alignment is already the case; but
519 * must be enforced here for CRIS, to let the least signficant bit
520 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
523 struct block_device {
524 dev_t bd_dev; /* not a kdev_t - it's a search key */
525 struct inode * bd_inode; /* will die */
527 struct mutex bd_mutex; /* open/close mutex */
528 struct semaphore bd_mount_sem;
529 struct list_head bd_inodes;
533 struct list_head bd_holder_list;
535 struct block_device * bd_contains;
536 unsigned bd_block_size;
537 struct hd_struct * bd_part;
538 /* number of times partitions within this device have been opened. */
539 unsigned bd_part_count;
541 struct gendisk * bd_disk;
542 struct list_head bd_list;
543 struct backing_dev_info *bd_inode_backing_dev_info;
545 * Private data. You must have bd_claim'ed the block_device
546 * to use this. NOTE: bd_claim allows an owner to claim
547 * the same device multiple times, the owner must take special
548 * care to not mess up bd_private for that case.
550 unsigned long bd_private;
554 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
557 #define PAGECACHE_TAG_DIRTY 0
558 #define PAGECACHE_TAG_WRITEBACK 1
560 int mapping_tagged(struct address_space *mapping, int tag);
563 * Might pages of this file be mapped into userspace?
565 static inline int mapping_mapped(struct address_space *mapping)
567 return !prio_tree_empty(&mapping->i_mmap) ||
568 !list_empty(&mapping->i_mmap_nonlinear);
572 * Might pages of this file have been modified in userspace?
573 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
574 * marks vma as VM_SHARED if it is shared, and the file was opened for
575 * writing i.e. vma may be mprotected writable even if now readonly.
577 static inline int mapping_writably_mapped(struct address_space *mapping)
579 return mapping->i_mmap_writable != 0;
583 * Use sequence counter to get consistent i_size on 32-bit processors.
585 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
586 #include <linux/seqlock.h>
587 #define __NEED_I_SIZE_ORDERED
588 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
590 #define i_size_ordered_init(inode) do { } while (0)
594 struct hlist_node i_hash;
595 struct list_head i_list;
596 struct list_head i_sb_list;
597 struct list_head i_dentry;
600 unsigned int i_nlink;
606 #ifdef __NEED_I_SIZE_ORDERED
607 seqcount_t i_size_seqcount;
609 struct timespec i_atime;
610 struct timespec i_mtime;
611 struct timespec i_ctime;
612 unsigned int i_blkbits;
614 unsigned short i_bytes;
616 spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
617 struct mutex i_mutex;
618 struct rw_semaphore i_alloc_sem;
619 const struct inode_operations *i_op;
620 const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
621 struct super_block *i_sb;
622 struct file_lock *i_flock;
623 struct address_space *i_mapping;
624 struct address_space i_data;
626 struct dquot *i_dquot[MAXQUOTAS];
628 struct list_head i_devices;
630 struct pipe_inode_info *i_pipe;
631 struct block_device *i_bdev;
638 #ifdef CONFIG_DNOTIFY
639 unsigned long i_dnotify_mask; /* Directory notify events */
640 struct dnotify_struct *i_dnotify; /* for directory notifications */
643 #ifdef CONFIG_INOTIFY
644 struct list_head inotify_watches; /* watches on this inode */
645 struct mutex inotify_mutex; /* protects the watches list */
648 unsigned long i_state;
649 unsigned long dirtied_when; /* jiffies of first dirtying */
651 unsigned int i_flags;
653 atomic_t i_writecount;
654 #ifdef CONFIG_SECURITY
657 void *i_private; /* fs or device private pointer */
661 * inode->i_mutex nesting subclasses for the lock validator:
663 * 0: the object of the current VFS operation
668 * The locking order between these classes is
669 * parent -> child -> normal -> xattr -> quota
671 enum inode_i_mutex_lock_class
680 extern void inode_double_lock(struct inode *inode1, struct inode *inode2);
681 extern void inode_double_unlock(struct inode *inode1, struct inode *inode2);
684 * NOTE: in a 32bit arch with a preemptable kernel and
685 * an UP compile the i_size_read/write must be atomic
686 * with respect to the local cpu (unlike with preempt disabled),
687 * but they don't need to be atomic with respect to other cpus like in
688 * true SMP (so they need either to either locally disable irq around
689 * the read or for example on x86 they can be still implemented as a
690 * cmpxchg8b without the need of the lock prefix). For SMP compiles
691 * and 64bit archs it makes no difference if preempt is enabled or not.
693 static inline loff_t i_size_read(const struct inode *inode)
695 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
700 seq = read_seqcount_begin(&inode->i_size_seqcount);
701 i_size = inode->i_size;
702 } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
704 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
708 i_size = inode->i_size;
712 return inode->i_size;
717 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
718 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
719 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
721 static inline void i_size_write(struct inode *inode, loff_t i_size)
723 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
724 write_seqcount_begin(&inode->i_size_seqcount);
725 inode->i_size = i_size;
726 write_seqcount_end(&inode->i_size_seqcount);
727 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
729 inode->i_size = i_size;
732 inode->i_size = i_size;
736 static inline unsigned iminor(const struct inode *inode)
738 return MINOR(inode->i_rdev);
741 static inline unsigned imajor(const struct inode *inode)
743 return MAJOR(inode->i_rdev);
746 extern struct block_device *I_BDEV(struct inode *inode);
749 rwlock_t lock; /* protects pid, uid, euid fields */
750 struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
751 enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
752 uid_t uid, euid; /* uid/euid of process setting the owner */
753 int signum; /* posix.1b rt signal to be delivered on IO */
757 * Track a single file's readahead state
759 struct file_ra_state {
760 pgoff_t start; /* where readahead started */
761 unsigned int size; /* # of readahead pages */
762 unsigned int async_size; /* do asynchronous readahead when
763 there are only # of pages ahead */
765 unsigned int ra_pages; /* Maximum readahead window */
766 int mmap_miss; /* Cache miss stat for mmap accesses */
767 loff_t prev_pos; /* Cache last read() position */
771 * Check if @index falls in the readahead windows.
773 static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
775 return (index >= ra->start &&
776 index < ra->start + ra->size);
779 #define FILE_MNT_WRITE_TAKEN 1
780 #define FILE_MNT_WRITE_RELEASED 2
784 * fu_list becomes invalid after file_free is called and queued via
785 * fu_rcuhead for RCU freeing
788 struct list_head fu_list;
789 struct rcu_head fu_rcuhead;
792 #define f_dentry f_path.dentry
793 #define f_vfsmnt f_path.mnt
794 const struct file_operations *f_op;
796 unsigned int f_flags;
799 struct fown_struct f_owner;
800 unsigned int f_uid, f_gid;
801 struct file_ra_state f_ra;
804 #ifdef CONFIG_SECURITY
807 /* needed for tty driver, and maybe others */
811 /* Used by fs/eventpoll.c to link all the hooks to this file */
812 struct list_head f_ep_links;
813 spinlock_t f_ep_lock;
814 #endif /* #ifdef CONFIG_EPOLL */
815 struct address_space *f_mapping;
816 #ifdef CONFIG_DEBUG_WRITECOUNT
817 unsigned long f_mnt_write_state;
820 extern spinlock_t files_lock;
821 #define file_list_lock() spin_lock(&files_lock);
822 #define file_list_unlock() spin_unlock(&files_lock);
824 #define get_file(x) atomic_inc(&(x)->f_count)
825 #define file_count(x) atomic_read(&(x)->f_count)
827 #ifdef CONFIG_DEBUG_WRITECOUNT
828 static inline void file_take_write(struct file *f)
830 WARN_ON(f->f_mnt_write_state != 0);
831 f->f_mnt_write_state = FILE_MNT_WRITE_TAKEN;
833 static inline void file_release_write(struct file *f)
835 f->f_mnt_write_state |= FILE_MNT_WRITE_RELEASED;
837 static inline void file_reset_write(struct file *f)
839 f->f_mnt_write_state = 0;
841 static inline void file_check_state(struct file *f)
844 * At this point, either both or neither of these bits
847 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN);
848 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_RELEASED);
850 static inline int file_check_writeable(struct file *f)
852 if (f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN)
854 printk(KERN_WARNING "writeable file with no "
855 "mnt_want_write()\n");
859 #else /* !CONFIG_DEBUG_WRITECOUNT */
860 static inline void file_take_write(struct file *filp) {}
861 static inline void file_release_write(struct file *filp) {}
862 static inline void file_reset_write(struct file *filp) {}
863 static inline void file_check_state(struct file *filp) {}
864 static inline int file_check_writeable(struct file *filp)
868 #endif /* CONFIG_DEBUG_WRITECOUNT */
870 #define MAX_NON_LFS ((1UL<<31) - 1)
872 /* Page cache limit. The filesystems should put that into their s_maxbytes
873 limits, otherwise bad things can happen in VM. */
874 #if BITS_PER_LONG==32
875 #define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
876 #elif BITS_PER_LONG==64
877 #define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
882 #define FL_ACCESS 8 /* not trying to lock, just looking */
883 #define FL_EXISTS 16 /* when unlocking, test for existence */
884 #define FL_LEASE 32 /* lease held on this file */
885 #define FL_CLOSE 64 /* unlock on close */
886 #define FL_SLEEP 128 /* A blocking lock */
889 * The POSIX file lock owner is determined by
890 * the "struct files_struct" in the thread group
891 * (or NULL for no owner - BSD locks).
893 * Lockd stuffs a "host" pointer into this.
895 typedef struct files_struct *fl_owner_t;
897 struct file_lock_operations {
898 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
899 void (*fl_release_private)(struct file_lock *);
902 struct lock_manager_operations {
903 int (*fl_compare_owner)(struct file_lock *, struct file_lock *);
904 void (*fl_notify)(struct file_lock *); /* unblock callback */
905 int (*fl_grant)(struct file_lock *, struct file_lock *, int);
906 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
907 void (*fl_release_private)(struct file_lock *);
908 void (*fl_break)(struct file_lock *);
909 int (*fl_mylease)(struct file_lock *, struct file_lock *);
910 int (*fl_change)(struct file_lock **, int);
913 /* that will die - we need it for nfs_lock_info */
914 #include <linux/nfs_fs_i.h>
917 struct file_lock *fl_next; /* singly linked list for this inode */
918 struct list_head fl_link; /* doubly linked list of all locks */
919 struct list_head fl_block; /* circular list of blocked processes */
921 unsigned char fl_flags;
922 unsigned char fl_type;
924 struct pid *fl_nspid;
925 wait_queue_head_t fl_wait;
926 struct file *fl_file;
930 struct fasync_struct * fl_fasync; /* for lease break notifications */
931 unsigned long fl_break_time; /* for nonblocking lease breaks */
933 struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
934 struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
936 struct nfs_lock_info nfs_fl;
937 struct nfs4_lock_info nfs4_fl;
939 struct list_head link; /* link in AFS vnode's pending_locks list */
940 int state; /* state of grant or error if -ve */
945 /* The following constant reflects the upper bound of the file/locking space */
947 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
948 #define OFFSET_MAX INT_LIMIT(loff_t)
949 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
952 #include <linux/fcntl.h>
954 extern int fcntl_getlk(struct file *, struct flock __user *);
955 extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
956 struct flock __user *);
958 #if BITS_PER_LONG == 32
959 extern int fcntl_getlk64(struct file *, struct flock64 __user *);
960 extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
961 struct flock64 __user *);
964 extern void send_sigio(struct fown_struct *fown, int fd, int band);
965 extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
966 extern int fcntl_getlease(struct file *filp);
969 extern int do_sync_mapping_range(struct address_space *mapping, loff_t offset,
970 loff_t endbyte, unsigned int flags);
973 extern void locks_init_lock(struct file_lock *);
974 extern void locks_copy_lock(struct file_lock *, struct file_lock *);
975 extern void __locks_copy_lock(struct file_lock *, const struct file_lock *);
976 extern void locks_remove_posix(struct file *, fl_owner_t);
977 extern void locks_remove_flock(struct file *);
978 extern void posix_test_lock(struct file *, struct file_lock *);
979 extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
980 extern int posix_lock_file_wait(struct file *, struct file_lock *);
981 extern int posix_unblock_lock(struct file *, struct file_lock *);
982 extern int vfs_test_lock(struct file *, struct file_lock *);
983 extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
984 extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
985 extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
986 extern int __break_lease(struct inode *inode, unsigned int flags);
987 extern void lease_get_mtime(struct inode *, struct timespec *time);
988 extern int generic_setlease(struct file *, long, struct file_lock **);
989 extern int vfs_setlease(struct file *, long, struct file_lock **);
990 extern int lease_modify(struct file_lock **, int);
991 extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
992 extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
993 extern struct seq_operations locks_seq_operations;
995 struct fasync_struct {
998 struct fasync_struct *fa_next; /* singly linked list */
999 struct file *fa_file;
1002 #define FASYNC_MAGIC 0x4601
1004 /* SMP safe fasync helpers: */
1005 extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
1006 /* can be called from interrupts */
1007 extern void kill_fasync(struct fasync_struct **, int, int);
1008 /* only for net: no internal synchronization */
1009 extern void __kill_fasync(struct fasync_struct *, int, int);
1011 extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
1012 extern int f_setown(struct file *filp, unsigned long arg, int force);
1013 extern void f_delown(struct file *filp);
1014 extern pid_t f_getown(struct file *filp);
1015 extern int send_sigurg(struct fown_struct *fown);
1021 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1022 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1023 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1025 extern struct list_head super_blocks;
1026 extern spinlock_t sb_lock;
1028 #define S_BIAS (1<<30)
1029 struct super_block {
1030 struct list_head s_list; /* Keep this first */
1031 dev_t s_dev; /* search index; _not_ kdev_t */
1032 unsigned long s_blocksize;
1033 unsigned char s_blocksize_bits;
1034 unsigned char s_dirt;
1035 unsigned long long s_maxbytes; /* Max file size */
1036 struct file_system_type *s_type;
1037 const struct super_operations *s_op;
1038 struct dquot_operations *dq_op;
1039 struct quotactl_ops *s_qcop;
1040 const struct export_operations *s_export_op;
1041 unsigned long s_flags;
1042 unsigned long s_magic;
1043 struct dentry *s_root;
1044 struct rw_semaphore s_umount;
1045 struct mutex s_lock;
1050 #ifdef CONFIG_SECURITY
1053 struct xattr_handler **s_xattr;
1055 struct list_head s_inodes; /* all inodes */
1056 struct list_head s_dirty; /* dirty inodes */
1057 struct list_head s_io; /* parked for writeback */
1058 struct list_head s_more_io; /* parked for more writeback */
1059 struct hlist_head s_anon; /* anonymous dentries for (nfs) exporting */
1060 struct list_head s_files;
1062 struct block_device *s_bdev;
1063 struct mtd_info *s_mtd;
1064 struct list_head s_instances;
1065 struct quota_info s_dquot; /* Diskquota specific options */
1068 wait_queue_head_t s_wait_unfrozen;
1070 char s_id[32]; /* Informational name */
1072 void *s_fs_info; /* Filesystem private info */
1075 * The next field is for VFS *only*. No filesystems have any business
1076 * even looking at it. You had been warned.
1078 struct mutex s_vfs_rename_mutex; /* Kludge */
1080 /* Granularity of c/m/atime in ns.
1081 Cannot be worse than a second */
1085 * Filesystem subtype. If non-empty the filesystem type field
1086 * in /proc/mounts will be "type.subtype"
1091 * Saved mount options for lazy filesystems using
1092 * generic_show_options()
1097 extern struct timespec current_fs_time(struct super_block *sb);
1100 * Snapshotting support.
1104 SB_FREEZE_WRITE = 1,
1105 SB_FREEZE_TRANS = 2,
1108 #define vfs_check_frozen(sb, level) \
1109 wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
1111 #define get_fs_excl() atomic_inc(¤t->fs_excl)
1112 #define put_fs_excl() atomic_dec(¤t->fs_excl)
1113 #define has_fs_excl() atomic_read(¤t->fs_excl)
1115 #define is_owner_or_cap(inode) \
1116 ((current->fsuid == (inode)->i_uid) || capable(CAP_FOWNER))
1118 /* not quite ready to be deprecated, but... */
1119 extern void lock_super(struct super_block *);
1120 extern void unlock_super(struct super_block *);
1123 * VFS helper functions..
1125 extern int vfs_permission(struct nameidata *, int);
1126 extern int vfs_create(struct inode *, struct dentry *, int, struct nameidata *);
1127 extern int vfs_mkdir(struct inode *, struct dentry *, int);
1128 extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t);
1129 extern int vfs_symlink(struct inode *, struct dentry *, const char *, int);
1130 extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
1131 extern int vfs_rmdir(struct inode *, struct dentry *);
1132 extern int vfs_unlink(struct inode *, struct dentry *);
1133 extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
1136 * VFS dentry helper functions.
1138 extern void dentry_unhash(struct dentry *dentry);
1141 * VFS file helper functions.
1143 extern int file_permission(struct file *, int);
1148 * NOTE! These match bits 12..15 of stat.st_mode
1149 * (ie "(i_mode >> 12) & 15").
1151 #define DT_UNKNOWN 0
1161 #define OSYNC_METADATA (1<<0)
1162 #define OSYNC_DATA (1<<1)
1163 #define OSYNC_INODE (1<<2)
1164 int generic_osync_inode(struct inode *, struct address_space *, int);
1167 * This is the "filldir" function type, used by readdir() to let
1168 * the kernel specify what kind of dirent layout it wants to have.
1169 * This allows the kernel to read directories into kernel space or
1170 * to have different dirent layouts depending on the binary type.
1172 typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned);
1174 struct block_device_operations {
1175 int (*open) (struct inode *, struct file *);
1176 int (*release) (struct inode *, struct file *);
1177 int (*ioctl) (struct inode *, struct file *, unsigned, unsigned long);
1178 long (*unlocked_ioctl) (struct file *, unsigned, unsigned long);
1179 long (*compat_ioctl) (struct file *, unsigned, unsigned long);
1180 int (*direct_access) (struct block_device *, sector_t,
1181 void **, unsigned long *);
1182 int (*media_changed) (struct gendisk *);
1183 int (*revalidate_disk) (struct gendisk *);
1184 int (*getgeo)(struct block_device *, struct hd_geometry *);
1185 struct module *owner;
1189 * "descriptor" for what we're up to with a read.
1190 * This allows us to use the same read code yet
1191 * have multiple different users of the data that
1192 * we read from a file.
1194 * The simplest case just copies the data to user
1205 } read_descriptor_t;
1207 typedef int (*read_actor_t)(read_descriptor_t *, struct page *, unsigned long, unsigned long);
1209 /* These macros are for out of kernel modules to test that
1210 * the kernel supports the unlocked_ioctl and compat_ioctl
1211 * fields in struct file_operations. */
1212 #define HAVE_COMPAT_IOCTL 1
1213 #define HAVE_UNLOCKED_IOCTL 1
1217 * read, write, poll, fsync, readv, writev, unlocked_ioctl and compat_ioctl
1218 * can be called without the big kernel lock held in all filesystems.
1220 struct file_operations {
1221 struct module *owner;
1222 loff_t (*llseek) (struct file *, loff_t, int);
1223 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1224 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
1225 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1226 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1227 int (*readdir) (struct file *, void *, filldir_t);
1228 unsigned int (*poll) (struct file *, struct poll_table_struct *);
1229 int (*ioctl) (struct inode *, struct file *, unsigned int, unsigned long);
1230 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1231 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1232 int (*mmap) (struct file *, struct vm_area_struct *);
1233 int (*open) (struct inode *, struct file *);
1234 int (*flush) (struct file *, fl_owner_t id);
1235 int (*release) (struct inode *, struct file *);
1236 int (*fsync) (struct file *, struct dentry *, int datasync);
1237 int (*aio_fsync) (struct kiocb *, int datasync);
1238 int (*fasync) (int, struct file *, int);
1239 int (*lock) (struct file *, int, struct file_lock *);
1240 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1241 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1242 int (*check_flags)(int);
1243 int (*dir_notify)(struct file *filp, unsigned long arg);
1244 int (*flock) (struct file *, int, struct file_lock *);
1245 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1246 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
1247 int (*setlease)(struct file *, long, struct file_lock **);
1250 struct inode_operations {
1251 int (*create) (struct inode *,struct dentry *,int, struct nameidata *);
1252 struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
1253 int (*link) (struct dentry *,struct inode *,struct dentry *);
1254 int (*unlink) (struct inode *,struct dentry *);
1255 int (*symlink) (struct inode *,struct dentry *,const char *);
1256 int (*mkdir) (struct inode *,struct dentry *,int);
1257 int (*rmdir) (struct inode *,struct dentry *);
1258 int (*mknod) (struct inode *,struct dentry *,int,dev_t);
1259 int (*rename) (struct inode *, struct dentry *,
1260 struct inode *, struct dentry *);
1261 int (*readlink) (struct dentry *, char __user *,int);
1262 void * (*follow_link) (struct dentry *, struct nameidata *);
1263 void (*put_link) (struct dentry *, struct nameidata *, void *);
1264 void (*truncate) (struct inode *);
1265 int (*permission) (struct inode *, int, struct nameidata *);
1266 int (*setattr) (struct dentry *, struct iattr *);
1267 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1268 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1269 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1270 ssize_t (*listxattr) (struct dentry *, char *, size_t);
1271 int (*removexattr) (struct dentry *, const char *);
1272 void (*truncate_range)(struct inode *, loff_t, loff_t);
1273 long (*fallocate)(struct inode *inode, int mode, loff_t offset,
1279 ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
1280 unsigned long nr_segs, unsigned long fast_segs,
1281 struct iovec *fast_pointer,
1282 struct iovec **ret_pointer);
1284 extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1285 extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1286 extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1287 unsigned long, loff_t *);
1288 extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1289 unsigned long, loff_t *);
1291 struct super_operations {
1292 struct inode *(*alloc_inode)(struct super_block *sb);
1293 void (*destroy_inode)(struct inode *);
1295 void (*dirty_inode) (struct inode *);
1296 int (*write_inode) (struct inode *, int);
1297 void (*drop_inode) (struct inode *);
1298 void (*delete_inode) (struct inode *);
1299 void (*put_super) (struct super_block *);
1300 void (*write_super) (struct super_block *);
1301 int (*sync_fs)(struct super_block *sb, int wait);
1302 void (*write_super_lockfs) (struct super_block *);
1303 void (*unlockfs) (struct super_block *);
1304 int (*statfs) (struct dentry *, struct kstatfs *);
1305 int (*remount_fs) (struct super_block *, int *, char *);
1306 void (*clear_inode) (struct inode *);
1307 void (*umount_begin) (struct super_block *);
1309 int (*show_options)(struct seq_file *, struct vfsmount *);
1310 int (*show_stats)(struct seq_file *, struct vfsmount *);
1312 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1313 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
1318 * Inode state bits. Protected by inode_lock.
1320 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1321 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1323 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1324 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1325 * various stages of removing an inode.
1327 * Two bits are used for locking and completion notification, I_LOCK and I_SYNC.
1329 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1330 * fdatasync(). i_atime is the usual cause.
1331 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1332 * these changes separately from I_DIRTY_SYNC so that we
1333 * don't have to write inode on fdatasync() when only
1334 * mtime has changed in it.
1335 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1336 * I_NEW get_new_inode() sets i_state to I_LOCK|I_NEW. Both
1337 * are cleared by unlock_new_inode(), called from iget().
1338 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1339 * is zero. I_FREEING must be set when I_WILL_FREE is
1341 * I_FREEING Set when inode is about to be freed but still has dirty
1342 * pages or buffers attached or the inode itself is still
1344 * I_CLEAR Set by clear_inode(). In this state the inode is clean
1345 * and can be destroyed.
1347 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1348 * prohibited for many purposes. iget() must wait for
1349 * the inode to be completely released, then create it
1350 * anew. Other functions will just ignore such inodes,
1351 * if appropriate. I_LOCK is used for waiting.
1353 * I_LOCK Serves as both a mutex and completion notification.
1354 * New inodes set I_LOCK. If two processes both create
1355 * the same inode, one of them will release its inode and
1356 * wait for I_LOCK to be released before returning.
1357 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1358 * also cause waiting on I_LOCK, without I_LOCK actually
1359 * being set. find_inode() uses this to prevent returning
1360 * nearly-dead inodes.
1361 * I_SYNC Similar to I_LOCK, but limited in scope to writeback
1362 * of inode dirty data. Having a separate lock for this
1363 * purpose reduces latency and prevents some filesystem-
1364 * specific deadlocks.
1366 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1367 * Q: igrab() only checks on (I_FREEING|I_WILL_FREE). Should it also check on
1368 * I_CLEAR? If not, why?
1370 #define I_DIRTY_SYNC 1
1371 #define I_DIRTY_DATASYNC 2
1372 #define I_DIRTY_PAGES 4
1374 #define I_WILL_FREE 16
1375 #define I_FREEING 32
1378 #define I_LOCK (1 << __I_LOCK)
1380 #define I_SYNC (1 << __I_SYNC)
1382 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1384 extern void __mark_inode_dirty(struct inode *, int);
1385 static inline void mark_inode_dirty(struct inode *inode)
1387 __mark_inode_dirty(inode, I_DIRTY);
1390 static inline void mark_inode_dirty_sync(struct inode *inode)
1392 __mark_inode_dirty(inode, I_DIRTY_SYNC);
1396 * inc_nlink - directly increment an inode's link count
1399 * This is a low-level filesystem helper to replace any
1400 * direct filesystem manipulation of i_nlink. Currently,
1401 * it is only here for parity with dec_nlink().
1403 static inline void inc_nlink(struct inode *inode)
1408 static inline void inode_inc_link_count(struct inode *inode)
1411 mark_inode_dirty(inode);
1415 * drop_nlink - directly drop an inode's link count
1418 * This is a low-level filesystem helper to replace any
1419 * direct filesystem manipulation of i_nlink. In cases
1420 * where we are attempting to track writes to the
1421 * filesystem, a decrement to zero means an imminent
1422 * write when the file is truncated and actually unlinked
1423 * on the filesystem.
1425 static inline void drop_nlink(struct inode *inode)
1431 * clear_nlink - directly zero an inode's link count
1434 * This is a low-level filesystem helper to replace any
1435 * direct filesystem manipulation of i_nlink. See
1436 * drop_nlink() for why we care about i_nlink hitting zero.
1438 static inline void clear_nlink(struct inode *inode)
1443 static inline void inode_dec_link_count(struct inode *inode)
1446 mark_inode_dirty(inode);
1450 * inode_inc_iversion - increments i_version
1451 * @inode: inode that need to be updated
1453 * Every time the inode is modified, the i_version field will be incremented.
1454 * The filesystem has to be mounted with i_version flag
1457 static inline void inode_inc_iversion(struct inode *inode)
1459 spin_lock(&inode->i_lock);
1461 spin_unlock(&inode->i_lock);
1464 extern void touch_atime(struct vfsmount *mnt, struct dentry *dentry);
1465 static inline void file_accessed(struct file *file)
1467 if (!(file->f_flags & O_NOATIME))
1468 touch_atime(file->f_path.mnt, file->f_path.dentry);
1471 int sync_inode(struct inode *inode, struct writeback_control *wbc);
1473 struct file_system_type {
1476 int (*get_sb) (struct file_system_type *, int,
1477 const char *, void *, struct vfsmount *);
1478 void (*kill_sb) (struct super_block *);
1479 struct module *owner;
1480 struct file_system_type * next;
1481 struct list_head fs_supers;
1483 struct lock_class_key s_lock_key;
1484 struct lock_class_key s_umount_key;
1486 struct lock_class_key i_lock_key;
1487 struct lock_class_key i_mutex_key;
1488 struct lock_class_key i_mutex_dir_key;
1489 struct lock_class_key i_alloc_sem_key;
1492 extern int get_sb_bdev(struct file_system_type *fs_type,
1493 int flags, const char *dev_name, void *data,
1494 int (*fill_super)(struct super_block *, void *, int),
1495 struct vfsmount *mnt);
1496 extern int get_sb_single(struct file_system_type *fs_type,
1497 int flags, void *data,
1498 int (*fill_super)(struct super_block *, void *, int),
1499 struct vfsmount *mnt);
1500 extern int get_sb_nodev(struct file_system_type *fs_type,
1501 int flags, void *data,
1502 int (*fill_super)(struct super_block *, void *, int),
1503 struct vfsmount *mnt);
1504 void generic_shutdown_super(struct super_block *sb);
1505 void kill_block_super(struct super_block *sb);
1506 void kill_anon_super(struct super_block *sb);
1507 void kill_litter_super(struct super_block *sb);
1508 void deactivate_super(struct super_block *sb);
1509 int set_anon_super(struct super_block *s, void *data);
1510 struct super_block *sget(struct file_system_type *type,
1511 int (*test)(struct super_block *,void *),
1512 int (*set)(struct super_block *,void *),
1514 extern int get_sb_pseudo(struct file_system_type *, char *,
1515 const struct super_operations *ops, unsigned long,
1516 struct vfsmount *mnt);
1517 extern int simple_set_mnt(struct vfsmount *mnt, struct super_block *sb);
1518 int __put_super_and_need_restart(struct super_block *sb);
1519 void unnamed_dev_init(void);
1521 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1522 #define fops_get(fops) \
1523 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1524 #define fops_put(fops) \
1525 do { if (fops) module_put((fops)->owner); } while(0)
1527 extern int register_filesystem(struct file_system_type *);
1528 extern int unregister_filesystem(struct file_system_type *);
1529 extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
1530 #define kern_mount(type) kern_mount_data(type, NULL)
1531 extern int may_umount_tree(struct vfsmount *);
1532 extern int may_umount(struct vfsmount *);
1533 extern long do_mount(char *, char *, char *, unsigned long, void *);
1534 extern struct vfsmount *collect_mounts(struct vfsmount *, struct dentry *);
1535 extern void drop_collected_mounts(struct vfsmount *);
1537 extern int vfs_statfs(struct dentry *, struct kstatfs *);
1540 extern struct kobject *fs_kobj;
1542 #define FLOCK_VERIFY_READ 1
1543 #define FLOCK_VERIFY_WRITE 2
1545 extern int locks_mandatory_locked(struct inode *);
1546 extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
1549 * Candidates for mandatory locking have the setgid bit set
1550 * but no group execute bit - an otherwise meaningless combination.
1553 static inline int __mandatory_lock(struct inode *ino)
1555 return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
1559 * ... and these candidates should be on MS_MANDLOCK mounted fs,
1560 * otherwise these will be advisory locks
1563 static inline int mandatory_lock(struct inode *ino)
1565 return IS_MANDLOCK(ino) && __mandatory_lock(ino);
1568 static inline int locks_verify_locked(struct inode *inode)
1570 if (mandatory_lock(inode))
1571 return locks_mandatory_locked(inode);
1575 extern int rw_verify_area(int, struct file *, loff_t *, size_t);
1577 static inline int locks_verify_truncate(struct inode *inode,
1581 if (inode->i_flock && mandatory_lock(inode))
1582 return locks_mandatory_area(
1583 FLOCK_VERIFY_WRITE, inode, filp,
1584 size < inode->i_size ? size : inode->i_size,
1585 (size < inode->i_size ? inode->i_size - size
1586 : size - inode->i_size)
1591 static inline int break_lease(struct inode *inode, unsigned int mode)
1594 return __break_lease(inode, mode);
1600 extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
1602 extern long do_sys_open(int dfd, const char __user *filename, int flags,
1604 extern struct file *filp_open(const char *, int, int);
1605 extern struct file * dentry_open(struct dentry *, struct vfsmount *, int);
1606 extern int filp_close(struct file *, fl_owner_t id);
1607 extern char * getname(const char __user *);
1610 extern void __init vfs_caches_init_early(void);
1611 extern void __init vfs_caches_init(unsigned long);
1613 extern struct kmem_cache *names_cachep;
1615 #define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL)
1616 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
1617 #ifndef CONFIG_AUDITSYSCALL
1618 #define putname(name) __putname(name)
1620 extern void putname(const char *name);
1624 extern int register_blkdev(unsigned int, const char *);
1625 extern void unregister_blkdev(unsigned int, const char *);
1626 extern struct block_device *bdget(dev_t);
1627 extern void bd_set_size(struct block_device *, loff_t size);
1628 extern void bd_forget(struct inode *inode);
1629 extern void bdput(struct block_device *);
1630 extern struct block_device *open_by_devnum(dev_t, unsigned);
1632 static inline void bd_forget(struct inode *inode) {}
1634 extern const struct file_operations def_blk_fops;
1635 extern const struct file_operations def_chr_fops;
1636 extern const struct file_operations bad_sock_fops;
1637 extern const struct file_operations def_fifo_fops;
1639 extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
1640 extern int blkdev_ioctl(struct inode *, struct file *, unsigned, unsigned long);
1641 extern int blkdev_driver_ioctl(struct inode *inode, struct file *file,
1642 struct gendisk *disk, unsigned cmd,
1644 extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
1645 extern int blkdev_get(struct block_device *, mode_t, unsigned);
1646 extern int blkdev_put(struct block_device *);
1647 extern int bd_claim(struct block_device *, void *);
1648 extern void bd_release(struct block_device *);
1650 extern int bd_claim_by_disk(struct block_device *, void *, struct gendisk *);
1651 extern void bd_release_from_disk(struct block_device *, struct gendisk *);
1653 #define bd_claim_by_disk(bdev, holder, disk) bd_claim(bdev, holder)
1654 #define bd_release_from_disk(bdev, disk) bd_release(bdev)
1659 #define CHRDEV_MAJOR_HASH_SIZE 255
1660 extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
1661 extern int register_chrdev_region(dev_t, unsigned, const char *);
1662 extern int register_chrdev(unsigned int, const char *,
1663 const struct file_operations *);
1664 extern void unregister_chrdev(unsigned int, const char *);
1665 extern void unregister_chrdev_region(dev_t, unsigned);
1666 extern void chrdev_show(struct seq_file *,off_t);
1668 /* fs/block_dev.c */
1669 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
1672 #define BLKDEV_MAJOR_HASH_SIZE 255
1673 extern const char *__bdevname(dev_t, char *buffer);
1674 extern const char *bdevname(struct block_device *bdev, char *buffer);
1675 extern struct block_device *lookup_bdev(const char *);
1676 extern struct block_device *open_bdev_excl(const char *, int, void *);
1677 extern void close_bdev_excl(struct block_device *);
1678 extern void blkdev_show(struct seq_file *,off_t);
1680 #define BLKDEV_MAJOR_HASH_SIZE 0
1683 extern void init_special_inode(struct inode *, umode_t, dev_t);
1685 /* Invalid inode operations -- fs/bad_inode.c */
1686 extern void make_bad_inode(struct inode *);
1687 extern int is_bad_inode(struct inode *);
1689 extern const struct file_operations read_fifo_fops;
1690 extern const struct file_operations write_fifo_fops;
1691 extern const struct file_operations rdwr_fifo_fops;
1693 extern int fs_may_remount_ro(struct super_block *);
1697 * return READ, READA, or WRITE
1699 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
1702 * return data direction, READ or WRITE
1704 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
1706 extern int check_disk_change(struct block_device *);
1707 extern int __invalidate_device(struct block_device *);
1708 extern int invalidate_partition(struct gendisk *, int);
1710 extern int invalidate_inodes(struct super_block *);
1711 unsigned long __invalidate_mapping_pages(struct address_space *mapping,
1712 pgoff_t start, pgoff_t end,
1714 unsigned long invalidate_mapping_pages(struct address_space *mapping,
1715 pgoff_t start, pgoff_t end);
1717 static inline unsigned long __deprecated
1718 invalidate_inode_pages(struct address_space *mapping)
1720 return invalidate_mapping_pages(mapping, 0, ~0UL);
1723 static inline void invalidate_remote_inode(struct inode *inode)
1725 if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
1726 S_ISLNK(inode->i_mode))
1727 invalidate_mapping_pages(inode->i_mapping, 0, -1);
1729 extern int invalidate_inode_pages2(struct address_space *mapping);
1730 extern int invalidate_inode_pages2_range(struct address_space *mapping,
1731 pgoff_t start, pgoff_t end);
1732 extern void generic_sync_sb_inodes(struct super_block *sb,
1733 struct writeback_control *wbc);
1734 extern int write_inode_now(struct inode *, int);
1735 extern int filemap_fdatawrite(struct address_space *);
1736 extern int filemap_flush(struct address_space *);
1737 extern int filemap_fdatawait(struct address_space *);
1738 extern int filemap_write_and_wait(struct address_space *mapping);
1739 extern int filemap_write_and_wait_range(struct address_space *mapping,
1740 loff_t lstart, loff_t lend);
1741 extern int wait_on_page_writeback_range(struct address_space *mapping,
1742 pgoff_t start, pgoff_t end);
1743 extern int __filemap_fdatawrite_range(struct address_space *mapping,
1744 loff_t start, loff_t end, int sync_mode);
1745 extern int filemap_fdatawrite_range(struct address_space *mapping,
1746 loff_t start, loff_t end);
1748 extern long do_fsync(struct file *file, int datasync);
1749 extern void sync_supers(void);
1750 extern void sync_filesystems(int wait);
1751 extern void __fsync_super(struct super_block *sb);
1752 extern void emergency_sync(void);
1753 extern void emergency_remount(void);
1754 extern int do_remount_sb(struct super_block *sb, int flags,
1755 void *data, int force);
1757 extern sector_t bmap(struct inode *, sector_t);
1759 extern int notify_change(struct dentry *, struct iattr *);
1760 extern int permission(struct inode *, int, struct nameidata *);
1761 extern int generic_permission(struct inode *, int,
1762 int (*check_acl)(struct inode *, int));
1764 extern int get_write_access(struct inode *);
1765 extern int deny_write_access(struct file *);
1766 static inline void put_write_access(struct inode * inode)
1768 atomic_dec(&inode->i_writecount);
1770 static inline void allow_write_access(struct file *file)
1773 atomic_inc(&file->f_path.dentry->d_inode->i_writecount);
1775 extern int do_pipe(int *);
1776 extern struct file *create_read_pipe(struct file *f);
1777 extern struct file *create_write_pipe(void);
1778 extern void free_write_pipe(struct file *);
1780 extern struct file *do_filp_open(int dfd, const char *pathname,
1781 int open_flag, int mode);
1782 extern int may_open(struct nameidata *, int, int);
1784 extern int kernel_read(struct file *, unsigned long, char *, unsigned long);
1785 extern struct file * open_exec(const char *);
1787 /* fs/dcache.c -- generic fs support functions */
1788 extern int is_subdir(struct dentry *, struct dentry *);
1789 extern ino_t find_inode_number(struct dentry *, struct qstr *);
1791 #include <linux/err.h>
1793 /* needed for stackable file system support */
1794 extern loff_t default_llseek(struct file *file, loff_t offset, int origin);
1796 extern loff_t vfs_llseek(struct file *file, loff_t offset, int origin);
1798 extern void inode_init_once(struct inode *);
1799 extern void iput(struct inode *);
1800 extern struct inode * igrab(struct inode *);
1801 extern ino_t iunique(struct super_block *, ino_t);
1802 extern int inode_needs_sync(struct inode *inode);
1803 extern void generic_delete_inode(struct inode *inode);
1804 extern void generic_drop_inode(struct inode *inode);
1806 extern struct inode *ilookup5_nowait(struct super_block *sb,
1807 unsigned long hashval, int (*test)(struct inode *, void *),
1809 extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
1810 int (*test)(struct inode *, void *), void *data);
1811 extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
1813 extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
1814 extern struct inode * iget_locked(struct super_block *, unsigned long);
1815 extern void unlock_new_inode(struct inode *);
1817 extern void __iget(struct inode * inode);
1818 extern void iget_failed(struct inode *);
1819 extern void clear_inode(struct inode *);
1820 extern void destroy_inode(struct inode *);
1821 extern struct inode *new_inode(struct super_block *);
1822 extern int should_remove_suid(struct dentry *);
1823 extern int remove_suid(struct dentry *);
1825 extern void __insert_inode_hash(struct inode *, unsigned long hashval);
1826 extern void remove_inode_hash(struct inode *);
1827 static inline void insert_inode_hash(struct inode *inode) {
1828 __insert_inode_hash(inode, inode->i_ino);
1831 extern struct file * get_empty_filp(void);
1832 extern void file_move(struct file *f, struct list_head *list);
1833 extern void file_kill(struct file *f);
1836 extern void submit_bio(int, struct bio *);
1837 extern int bdev_read_only(struct block_device *);
1839 extern int set_blocksize(struct block_device *, int);
1840 extern int sb_set_blocksize(struct super_block *, int);
1841 extern int sb_min_blocksize(struct super_block *, int);
1842 extern int sb_has_dirty_inodes(struct super_block *);
1844 extern int generic_file_mmap(struct file *, struct vm_area_struct *);
1845 extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
1846 extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
1847 int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
1848 extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t);
1849 extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t);
1850 extern ssize_t generic_file_aio_write_nolock(struct kiocb *, const struct iovec *,
1851 unsigned long, loff_t);
1852 extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
1853 unsigned long *, loff_t, loff_t *, size_t, size_t);
1854 extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
1855 unsigned long, loff_t, loff_t *, size_t, ssize_t);
1856 extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
1857 extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
1858 extern int generic_segment_checks(const struct iovec *iov,
1859 unsigned long *nr_segs, size_t *count, int access_flags);
1862 extern ssize_t generic_file_splice_read(struct file *, loff_t *,
1863 struct pipe_inode_info *, size_t, unsigned int);
1864 extern ssize_t generic_file_splice_write(struct pipe_inode_info *,
1865 struct file *, loff_t *, size_t, unsigned int);
1866 extern ssize_t generic_file_splice_write_nolock(struct pipe_inode_info *,
1867 struct file *, loff_t *, size_t, unsigned int);
1868 extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
1869 struct file *out, loff_t *, size_t len, unsigned int flags);
1870 extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
1871 size_t len, unsigned int flags);
1874 file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
1875 extern loff_t no_llseek(struct file *file, loff_t offset, int origin);
1876 extern loff_t generic_file_llseek(struct file *file, loff_t offset, int origin);
1877 extern loff_t generic_file_llseek_unlocked(struct file *file, loff_t offset,
1879 extern int generic_file_open(struct inode * inode, struct file * filp);
1880 extern int nonseekable_open(struct inode * inode, struct file * filp);
1882 #ifdef CONFIG_FS_XIP
1883 extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
1885 extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
1886 extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
1887 size_t len, loff_t *ppos);
1888 extern int xip_truncate_page(struct address_space *mapping, loff_t from);
1890 static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
1897 ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
1898 struct block_device *bdev, const struct iovec *iov, loff_t offset,
1899 unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
1903 DIO_LOCKING = 1, /* need locking between buffered and direct access */
1904 DIO_NO_LOCKING, /* bdev; no locking at all between buffered/direct */
1905 DIO_OWN_LOCKING, /* filesystem locks buffered and direct internally */
1908 static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
1909 struct inode *inode, struct block_device *bdev, const struct iovec *iov,
1910 loff_t offset, unsigned long nr_segs, get_block_t get_block,
1911 dio_iodone_t end_io)
1913 return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset,
1914 nr_segs, get_block, end_io, DIO_LOCKING);
1917 static inline ssize_t blockdev_direct_IO_no_locking(int rw, struct kiocb *iocb,
1918 struct inode *inode, struct block_device *bdev, const struct iovec *iov,
1919 loff_t offset, unsigned long nr_segs, get_block_t get_block,
1920 dio_iodone_t end_io)
1922 return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset,
1923 nr_segs, get_block, end_io, DIO_NO_LOCKING);
1926 static inline ssize_t blockdev_direct_IO_own_locking(int rw, struct kiocb *iocb,
1927 struct inode *inode, struct block_device *bdev, const struct iovec *iov,
1928 loff_t offset, unsigned long nr_segs, get_block_t get_block,
1929 dio_iodone_t end_io)
1931 return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset,
1932 nr_segs, get_block, end_io, DIO_OWN_LOCKING);
1936 extern const struct file_operations generic_ro_fops;
1938 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
1940 extern int vfs_readlink(struct dentry *, char __user *, int, const char *);
1941 extern int vfs_follow_link(struct nameidata *, const char *);
1942 extern int page_readlink(struct dentry *, char __user *, int);
1943 extern void *page_follow_link_light(struct dentry *, struct nameidata *);
1944 extern void page_put_link(struct dentry *, struct nameidata *, void *);
1945 extern int __page_symlink(struct inode *inode, const char *symname, int len,
1947 extern int page_symlink(struct inode *inode, const char *symname, int len);
1948 extern const struct inode_operations page_symlink_inode_operations;
1949 extern int generic_readlink(struct dentry *, char __user *, int);
1950 extern void generic_fillattr(struct inode *, struct kstat *);
1951 extern int vfs_getattr(struct vfsmount *, struct dentry *, struct kstat *);
1952 void inode_add_bytes(struct inode *inode, loff_t bytes);
1953 void inode_sub_bytes(struct inode *inode, loff_t bytes);
1954 loff_t inode_get_bytes(struct inode *inode);
1955 void inode_set_bytes(struct inode *inode, loff_t bytes);
1957 extern int vfs_readdir(struct file *, filldir_t, void *);
1959 extern int vfs_stat(char __user *, struct kstat *);
1960 extern int vfs_lstat(char __user *, struct kstat *);
1961 extern int vfs_stat_fd(int dfd, char __user *, struct kstat *);
1962 extern int vfs_lstat_fd(int dfd, char __user *, struct kstat *);
1963 extern int vfs_fstat(unsigned int, struct kstat *);
1965 extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
1968 extern void get_filesystem(struct file_system_type *fs);
1969 extern void put_filesystem(struct file_system_type *fs);
1970 extern struct file_system_type *get_fs_type(const char *name);
1971 extern struct super_block *get_super(struct block_device *);
1972 extern struct super_block *user_get_super(dev_t);
1973 extern void drop_super(struct super_block *sb);
1975 extern int dcache_dir_open(struct inode *, struct file *);
1976 extern int dcache_dir_close(struct inode *, struct file *);
1977 extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
1978 extern int dcache_readdir(struct file *, void *, filldir_t);
1979 extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
1980 extern int simple_statfs(struct dentry *, struct kstatfs *);
1981 extern int simple_link(struct dentry *, struct inode *, struct dentry *);
1982 extern int simple_unlink(struct inode *, struct dentry *);
1983 extern int simple_rmdir(struct inode *, struct dentry *);
1984 extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
1985 extern int simple_sync_file(struct file *, struct dentry *, int);
1986 extern int simple_empty(struct dentry *);
1987 extern int simple_readpage(struct file *file, struct page *page);
1988 extern int simple_prepare_write(struct file *file, struct page *page,
1989 unsigned offset, unsigned to);
1990 extern int simple_write_begin(struct file *file, struct address_space *mapping,
1991 loff_t pos, unsigned len, unsigned flags,
1992 struct page **pagep, void **fsdata);
1993 extern int simple_write_end(struct file *file, struct address_space *mapping,
1994 loff_t pos, unsigned len, unsigned copied,
1995 struct page *page, void *fsdata);
1997 extern struct dentry *simple_lookup(struct inode *, struct dentry *, struct nameidata *);
1998 extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
1999 extern const struct file_operations simple_dir_operations;
2000 extern const struct inode_operations simple_dir_inode_operations;
2001 struct tree_descr { char *name; const struct file_operations *ops; int mode; };
2002 struct dentry *d_alloc_name(struct dentry *, const char *);
2003 extern int simple_fill_super(struct super_block *, int, struct tree_descr *);
2004 extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
2005 extern void simple_release_fs(struct vfsmount **mount, int *count);
2007 extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
2008 loff_t *ppos, const void *from, size_t available);
2009 extern ssize_t memory_read_from_buffer(void *to, size_t count, loff_t *ppos,
2010 const void *from, size_t available);
2012 #ifdef CONFIG_MIGRATION
2013 extern int buffer_migrate_page(struct address_space *,
2014 struct page *, struct page *);
2016 #define buffer_migrate_page NULL
2019 extern int inode_change_ok(struct inode *, struct iattr *);
2020 extern int __must_check inode_setattr(struct inode *, struct iattr *);
2022 extern void file_update_time(struct file *file);
2024 extern int generic_show_options(struct seq_file *m, struct vfsmount *mnt);
2025 extern void save_mount_options(struct super_block *sb, char *options);
2027 static inline ino_t parent_ino(struct dentry *dentry)
2031 spin_lock(&dentry->d_lock);
2032 res = dentry->d_parent->d_inode->i_ino;
2033 spin_unlock(&dentry->d_lock);
2037 /* Transaction based IO helpers */
2040 * An argresp is stored in an allocated page and holds the
2041 * size of the argument or response, along with its content
2043 struct simple_transaction_argresp {
2048 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2050 char *simple_transaction_get(struct file *file, const char __user *buf,
2052 ssize_t simple_transaction_read(struct file *file, char __user *buf,
2053 size_t size, loff_t *pos);
2054 int simple_transaction_release(struct inode *inode, struct file *file);
2056 static inline void simple_transaction_set(struct file *file, size_t n)
2058 struct simple_transaction_argresp *ar = file->private_data;
2060 BUG_ON(n > SIMPLE_TRANSACTION_LIMIT);
2063 * The barrier ensures that ar->size will really remain zero until
2064 * ar->data is ready for reading.
2071 * simple attribute files
2073 * These attributes behave similar to those in sysfs:
2075 * Writing to an attribute immediately sets a value, an open file can be
2076 * written to multiple times.
2078 * Reading from an attribute creates a buffer from the value that might get
2079 * read with multiple read calls. When the attribute has been read
2080 * completely, no further read calls are possible until the file is opened
2083 * All attributes contain a text representation of a numeric value
2084 * that are accessed with the get() and set() functions.
2086 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2087 static int __fops ## _open(struct inode *inode, struct file *file) \
2089 __simple_attr_check_format(__fmt, 0ull); \
2090 return simple_attr_open(inode, file, __get, __set, __fmt); \
2092 static struct file_operations __fops = { \
2093 .owner = THIS_MODULE, \
2094 .open = __fops ## _open, \
2095 .release = simple_attr_release, \
2096 .read = simple_attr_read, \
2097 .write = simple_attr_write, \
2100 static inline void __attribute__((format(printf, 1, 2)))
2101 __simple_attr_check_format(const char *fmt, ...)
2103 /* don't do anything, just let the compiler check the arguments; */
2106 int simple_attr_open(struct inode *inode, struct file *file,
2107 int (*get)(void *, u64 *), int (*set)(void *, u64),
2109 int simple_attr_release(struct inode *inode, struct file *file);
2110 ssize_t simple_attr_read(struct file *file, char __user *buf,
2111 size_t len, loff_t *ppos);
2112 ssize_t simple_attr_write(struct file *file, const char __user *buf,
2113 size_t len, loff_t *ppos);
2116 #ifdef CONFIG_SECURITY
2117 static inline char *alloc_secdata(void)
2119 return (char *)get_zeroed_page(GFP_KERNEL);
2122 static inline void free_secdata(void *secdata)
2124 free_page((unsigned long)secdata);
2127 static inline char *alloc_secdata(void)
2132 static inline void free_secdata(void *secdata)
2134 #endif /* CONFIG_SECURITY */
2137 int proc_nr_files(struct ctl_table *table, int write, struct file *filp,
2138 void __user *buffer, size_t *lenp, loff_t *ppos);
2140 int get_filesystem_list(char * buf);
2142 #endif /* __KERNEL__ */
2143 #endif /* _LINUX_FS_H */