5 * This file has definitions for some important file table
9 #include <linux/config.h>
10 #include <linux/limits.h>
11 #include <linux/ioctl.h>
12 #include <linux/rcuref.h>
15 * It's silly to have NR_OPEN bigger than NR_FILE, but you can change
16 * the file limit at runtime and only root can increase the per-process
17 * nr_file rlimit, so it's safe to set up a ridiculously high absolute
18 * upper limit on files-per-process.
20 * Some programs (notably those using select()) may have to be
21 * recompiled to take full advantage of the new limits..
24 /* Fixed constants first: */
26 #define NR_OPEN (1024*1024) /* Absolute upper limit on fd num */
27 #define INR_OPEN 1024 /* Initial setting for nfile rlimits */
29 #define BLOCK_SIZE_BITS 10
30 #define BLOCK_SIZE (1<<BLOCK_SIZE_BITS)
32 /* And dynamically-tunable limits and defaults: */
33 struct files_stat_struct
{
34 int nr_files
; /* read only */
35 int nr_free_files
; /* read only */
36 int max_files
; /* tunable */
38 extern struct files_stat_struct files_stat
;
40 struct inodes_stat_t
{
45 extern struct inodes_stat_t inodes_stat
;
47 extern int leases_enable
, lease_break_time
;
50 extern int dir_notify_enable
;
53 #define NR_FILE 8192 /* this can well be larger on a larger system */
63 /* Internal kernel extensions */
66 #define FMODE_PWRITE FMODE_PREAD /* These go hand in hand */
72 #define READA 2 /* read-ahead - don't block if no resources */
73 #define SWRITE 3 /* for ll_rw_block() - wait for buffer lock */
74 #define SPECIAL 4 /* For non-blockdevice requests in request queue */
75 #define READ_SYNC (READ | (1 << BIO_RW_SYNC))
76 #define WRITE_SYNC (WRITE | (1 << BIO_RW_SYNC))
77 #define WRITE_BARRIER ((1 << BIO_RW) | (1 << BIO_RW_BARRIER))
83 /* public flags for file_system_type */
84 #define FS_REQUIRES_DEV 1
85 #define FS_BINARY_MOUNTDATA 2
86 #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
87 #define FS_ODD_RENAME 32768 /* Temporary stuff; will go away as soon
88 * as nfs_rename() will be cleaned up
91 * These are the fs-independent mount-flags: up to 32 flags are supported
93 #define MS_RDONLY 1 /* Mount read-only */
94 #define MS_NOSUID 2 /* Ignore suid and sgid bits */
95 #define MS_NODEV 4 /* Disallow access to device special files */
96 #define MS_NOEXEC 8 /* Disallow program execution */
97 #define MS_SYNCHRONOUS 16 /* Writes are synced at once */
98 #define MS_REMOUNT 32 /* Alter flags of a mounted FS */
99 #define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
100 #define MS_DIRSYNC 128 /* Directory modifications are synchronous */
101 #define MS_NOATIME 1024 /* Do not update access times. */
102 #define MS_NODIRATIME 2048 /* Do not update directory access times */
106 #define MS_VERBOSE 32768
107 #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
108 #define MS_ACTIVE (1<<30)
109 #define MS_NOUSER (1<<31)
112 * Superblock flags that can be altered by MS_REMOUNT
114 #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK|MS_NOATIME|\
118 * Old magic mount flag and mask
120 #define MS_MGC_VAL 0xC0ED0000
121 #define MS_MGC_MSK 0xffff0000
123 /* Inode flags - they have nothing to superblock flags now */
125 #define S_SYNC 1 /* Writes are synced at once */
126 #define S_NOATIME 2 /* Do not update access times */
127 #define S_APPEND 4 /* Append-only file */
128 #define S_IMMUTABLE 8 /* Immutable file */
129 #define S_DEAD 16 /* removed, but still open directory */
130 #define S_NOQUOTA 32 /* Inode is not counted to quota */
131 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
132 #define S_NOCMTIME 128 /* Do not update file c/mtime */
133 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
134 #define S_PRIVATE 512 /* Inode is fs-internal */
137 * Note that nosuid etc flags are inode-specific: setting some file-system
138 * flags just means all the inodes inherit those flags by default. It might be
139 * possible to override it selectively if you really wanted to with some
140 * ioctl() that is not currently implemented.
142 * Exception: MS_RDONLY is always applied to the entire file system.
144 * Unfortunately, it is possible to change a filesystems flags with it mounted
145 * with files in use. This means that all of the inodes will not have their
146 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
147 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
149 #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
151 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
152 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
153 ((inode)->i_flags & S_SYNC))
154 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
155 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
156 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
158 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
159 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
160 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
161 #define IS_NOATIME(inode) (__IS_FLG(inode, MS_NOATIME) || ((inode)->i_flags & S_NOATIME))
162 #define IS_NODIRATIME(inode) __IS_FLG(inode, MS_NODIRATIME)
163 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
165 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
166 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
167 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
168 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
170 /* the read-only stuff doesn't really belong here, but any other place is
171 probably as bad and I don't want to create yet another include file. */
173 #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
174 #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
175 #define BLKRRPART _IO(0x12,95) /* re-read partition table */
176 #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
177 #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
178 #define BLKRASET _IO(0x12,98) /* set read ahead for block device */
179 #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
180 #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
181 #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
182 #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
183 #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
184 #define BLKSSZGET _IO(0x12,104)/* get block device sector size */
186 #define BLKPG _IO(0x12,105)/* See blkpg.h */
188 /* Some people are morons. Do not use sizeof! */
190 #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
191 #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
192 /* This was here just to show that the number is taken -
193 probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
195 /* A jump here: 108-111 have been used for various private purposes. */
196 #define BLKBSZGET _IOR(0x12,112,size_t)
197 #define BLKBSZSET _IOW(0x12,113,size_t)
198 #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
200 #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
201 #define FIBMAP _IO(0x00,1) /* bmap access */
202 #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
206 #include <linux/linkage.h>
207 #include <linux/wait.h>
208 #include <linux/types.h>
209 #include <linux/kdev_t.h>
210 #include <linux/dcache.h>
211 #include <linux/stat.h>
212 #include <linux/cache.h>
213 #include <linux/kobject.h>
214 #include <linux/list.h>
215 #include <linux/radix-tree.h>
216 #include <linux/prio_tree.h>
217 #include <linux/init.h>
218 #include <linux/sched.h>
220 #include <asm/atomic.h>
221 #include <asm/semaphore.h>
222 #include <asm/byteorder.h>
227 struct pipe_inode_info
;
228 struct poll_table_struct
;
230 struct vm_area_struct
;
233 /* Used to be a macro which just called the function, now just a function */
234 extern void update_atime (struct inode
*);
236 extern void __init
inode_init(unsigned long);
237 extern void __init
inode_init_early(void);
238 extern void __init
mnt_init(unsigned long);
239 extern void __init
files_init(unsigned long);
242 typedef int (get_block_t
)(struct inode
*inode
, sector_t iblock
,
243 struct buffer_head
*bh_result
, int create
);
244 typedef int (get_blocks_t
)(struct inode
*inode
, sector_t iblock
,
245 unsigned long max_blocks
,
246 struct buffer_head
*bh_result
, int create
);
247 typedef void (dio_iodone_t
)(struct kiocb
*iocb
, loff_t offset
,
248 ssize_t bytes
, void *private);
251 * Attribute flags. These should be or-ed together to figure out what
258 #define ATTR_ATIME 16
259 #define ATTR_MTIME 32
260 #define ATTR_CTIME 64
261 #define ATTR_ATIME_SET 128
262 #define ATTR_MTIME_SET 256
263 #define ATTR_FORCE 512 /* Not a change, but a change it */
264 #define ATTR_ATTR_FLAG 1024
265 #define ATTR_KILL_SUID 2048
266 #define ATTR_KILL_SGID 4096
269 * This is the Inode Attributes structure, used for notify_change(). It
270 * uses the above definitions as flags, to know which values have changed.
271 * Also, in this manner, a Filesystem can look at only the values it cares
272 * about. Basically, these are the attributes that the VFS layer can
273 * request to change from the FS layer.
275 * Derek Atkins <warlord@MIT.EDU> 94-10-20
278 unsigned int ia_valid
;
283 struct timespec ia_atime
;
284 struct timespec ia_mtime
;
285 struct timespec ia_ctime
;
289 * Includes for diskquotas.
291 #include <linux/quota.h>
294 * oh the beauties of C type declarations.
297 struct address_space
;
298 struct writeback_control
;
300 struct address_space_operations
{
301 int (*writepage
)(struct page
*page
, struct writeback_control
*wbc
);
302 int (*readpage
)(struct file
*, struct page
*);
303 int (*sync_page
)(struct page
*);
305 /* Write back some dirty pages from this mapping. */
306 int (*writepages
)(struct address_space
*, struct writeback_control
*);
308 /* Set a page dirty */
309 int (*set_page_dirty
)(struct page
*page
);
311 int (*readpages
)(struct file
*filp
, struct address_space
*mapping
,
312 struct list_head
*pages
, unsigned nr_pages
);
315 * ext3 requires that a successful prepare_write() call be followed
316 * by a commit_write() call - they must be balanced
318 int (*prepare_write
)(struct file
*, struct page
*, unsigned, unsigned);
319 int (*commit_write
)(struct file
*, struct page
*, unsigned, unsigned);
320 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
321 sector_t (*bmap
)(struct address_space
*, sector_t
);
322 int (*invalidatepage
) (struct page
*, unsigned long);
323 int (*releasepage
) (struct page
*, int);
324 ssize_t (*direct_IO
)(int, struct kiocb
*, const struct iovec
*iov
,
325 loff_t offset
, unsigned long nr_segs
);
326 struct page
* (*get_xip_page
)(struct address_space
*, sector_t
,
330 struct backing_dev_info
;
331 struct address_space
{
332 struct inode
*host
; /* owner: inode, block_device */
333 struct radix_tree_root page_tree
; /* radix tree of all pages */
334 rwlock_t tree_lock
; /* and rwlock protecting it */
335 unsigned int i_mmap_writable
;/* count VM_SHARED mappings */
336 struct prio_tree_root i_mmap
; /* tree of private and shared mappings */
337 struct list_head i_mmap_nonlinear
;/*list VM_NONLINEAR mappings */
338 spinlock_t i_mmap_lock
; /* protect tree, count, list */
339 unsigned int truncate_count
; /* Cover race condition with truncate */
340 unsigned long nrpages
; /* number of total pages */
341 pgoff_t writeback_index
;/* writeback starts here */
342 struct address_space_operations
*a_ops
; /* methods */
343 unsigned long flags
; /* error bits/gfp mask */
344 struct backing_dev_info
*backing_dev_info
; /* device readahead, etc */
345 spinlock_t private_lock
; /* for use by the address_space */
346 struct list_head private_list
; /* ditto */
347 struct address_space
*assoc_mapping
; /* ditto */
348 } __attribute__((aligned(sizeof(long))));
350 * On most architectures that alignment is already the case; but
351 * must be enforced here for CRIS, to let the least signficant bit
352 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
355 struct block_device
{
356 dev_t bd_dev
; /* not a kdev_t - it's a search key */
357 struct inode
* bd_inode
; /* will die */
359 struct semaphore bd_sem
; /* open/close mutex */
360 struct semaphore bd_mount_sem
; /* mount mutex */
361 struct list_head bd_inodes
;
364 struct block_device
* bd_contains
;
365 unsigned bd_block_size
;
366 struct hd_struct
* bd_part
;
367 /* number of times partitions within this device have been opened. */
368 unsigned bd_part_count
;
370 struct gendisk
* bd_disk
;
371 struct list_head bd_list
;
372 struct backing_dev_info
*bd_inode_backing_dev_info
;
374 * Private data. You must have bd_claim'ed the block_device
375 * to use this. NOTE: bd_claim allows an owner to claim
376 * the same device multiple times, the owner must take special
377 * care to not mess up bd_private for that case.
379 unsigned long bd_private
;
383 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
386 #define PAGECACHE_TAG_DIRTY 0
387 #define PAGECACHE_TAG_WRITEBACK 1
389 int mapping_tagged(struct address_space
*mapping
, int tag
);
392 * Might pages of this file be mapped into userspace?
394 static inline int mapping_mapped(struct address_space
*mapping
)
396 return !prio_tree_empty(&mapping
->i_mmap
) ||
397 !list_empty(&mapping
->i_mmap_nonlinear
);
401 * Might pages of this file have been modified in userspace?
402 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
403 * marks vma as VM_SHARED if it is shared, and the file was opened for
404 * writing i.e. vma may be mprotected writable even if now readonly.
406 static inline int mapping_writably_mapped(struct address_space
*mapping
)
408 return mapping
->i_mmap_writable
!= 0;
412 * Use sequence counter to get consistent i_size on 32-bit processors.
414 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
415 #include <linux/seqlock.h>
416 #define __NEED_I_SIZE_ORDERED
417 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
419 #define i_size_ordered_init(inode) do { } while (0)
423 struct hlist_node i_hash
;
424 struct list_head i_list
;
425 struct list_head i_sb_list
;
426 struct list_head i_dentry
;
430 unsigned int i_nlink
;
435 struct timespec i_atime
;
436 struct timespec i_mtime
;
437 struct timespec i_ctime
;
438 unsigned int i_blkbits
;
439 unsigned long i_blksize
;
440 unsigned long i_version
;
441 unsigned long i_blocks
;
442 unsigned short i_bytes
;
443 spinlock_t i_lock
; /* i_blocks, i_bytes, maybe i_size */
444 struct semaphore i_sem
;
445 struct rw_semaphore i_alloc_sem
;
446 struct inode_operations
*i_op
;
447 struct file_operations
*i_fop
; /* former ->i_op->default_file_ops */
448 struct super_block
*i_sb
;
449 struct file_lock
*i_flock
;
450 struct address_space
*i_mapping
;
451 struct address_space i_data
;
453 struct dquot
*i_dquot
[MAXQUOTAS
];
455 /* These three should probably be a union */
456 struct list_head i_devices
;
457 struct pipe_inode_info
*i_pipe
;
458 struct block_device
*i_bdev
;
464 #ifdef CONFIG_DNOTIFY
465 unsigned long i_dnotify_mask
; /* Directory notify events */
466 struct dnotify_struct
*i_dnotify
; /* for directory notifications */
469 #ifdef CONFIG_INOTIFY
470 struct list_head inotify_watches
; /* watches on this inode */
471 struct semaphore inotify_sem
; /* protects the watches list */
474 unsigned long i_state
;
475 unsigned long dirtied_when
; /* jiffies of first dirtying */
477 unsigned int i_flags
;
479 atomic_t i_writecount
;
484 #ifdef __NEED_I_SIZE_ORDERED
485 seqcount_t i_size_seqcount
;
490 * NOTE: in a 32bit arch with a preemptable kernel and
491 * an UP compile the i_size_read/write must be atomic
492 * with respect to the local cpu (unlike with preempt disabled),
493 * but they don't need to be atomic with respect to other cpus like in
494 * true SMP (so they need either to either locally disable irq around
495 * the read or for example on x86 they can be still implemented as a
496 * cmpxchg8b without the need of the lock prefix). For SMP compiles
497 * and 64bit archs it makes no difference if preempt is enabled or not.
499 static inline loff_t
i_size_read(struct inode
*inode
)
501 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
506 seq
= read_seqcount_begin(&inode
->i_size_seqcount
);
507 i_size
= inode
->i_size
;
508 } while (read_seqcount_retry(&inode
->i_size_seqcount
, seq
));
510 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
514 i_size
= inode
->i_size
;
518 return inode
->i_size
;
523 static inline void i_size_write(struct inode
*inode
, loff_t i_size
)
525 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
526 write_seqcount_begin(&inode
->i_size_seqcount
);
527 inode
->i_size
= i_size
;
528 write_seqcount_end(&inode
->i_size_seqcount
);
529 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
531 inode
->i_size
= i_size
;
534 inode
->i_size
= i_size
;
538 static inline unsigned iminor(struct inode
*inode
)
540 return MINOR(inode
->i_rdev
);
543 static inline unsigned imajor(struct inode
*inode
)
545 return MAJOR(inode
->i_rdev
);
548 extern struct block_device
*I_BDEV(struct inode
*inode
);
551 rwlock_t lock
; /* protects pid, uid, euid fields */
552 int pid
; /* pid or -pgrp where SIGIO should be sent */
553 uid_t uid
, euid
; /* uid/euid of process setting the owner */
555 int signum
; /* posix.1b rt signal to be delivered on IO */
559 * Track a single file's readahead state
561 struct file_ra_state
{
562 unsigned long start
; /* Current window */
564 unsigned long flags
; /* ra flags RA_FLAG_xxx*/
565 unsigned long cache_hit
; /* cache hit count*/
566 unsigned long prev_page
; /* Cache last read() position */
567 unsigned long ahead_start
; /* Ahead window */
568 unsigned long ahead_size
;
569 unsigned long ra_pages
; /* Maximum readahead window */
570 unsigned long mmap_hit
; /* Cache hit stat for mmap accesses */
571 unsigned long mmap_miss
; /* Cache miss stat for mmap accesses */
573 #define RA_FLAG_MISS 0x01 /* a cache miss occured against this file */
574 #define RA_FLAG_INCACHE 0x02 /* file is already in cache */
577 struct list_head f_list
;
578 struct dentry
*f_dentry
;
579 struct vfsmount
*f_vfsmnt
;
580 struct file_operations
*f_op
;
582 unsigned int f_flags
;
585 struct fown_struct f_owner
;
586 unsigned int f_uid
, f_gid
;
587 struct file_ra_state f_ra
;
589 unsigned long f_version
;
592 /* needed for tty driver, and maybe others */
596 /* Used by fs/eventpoll.c to link all the hooks to this file */
597 struct list_head f_ep_links
;
598 spinlock_t f_ep_lock
;
599 #endif /* #ifdef CONFIG_EPOLL */
600 struct address_space
*f_mapping
;
601 struct rcu_head f_rcuhead
;
603 extern spinlock_t files_lock
;
604 #define file_list_lock() spin_lock(&files_lock);
605 #define file_list_unlock() spin_unlock(&files_lock);
607 #define get_file(x) rcuref_inc(&(x)->f_count)
608 #define file_count(x) atomic_read(&(x)->f_count)
610 #define MAX_NON_LFS ((1UL<<31) - 1)
612 /* Page cache limit. The filesystems should put that into their s_maxbytes
613 limits, otherwise bad things can happen in VM. */
614 #if BITS_PER_LONG==32
615 #define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
616 #elif BITS_PER_LONG==64
617 #define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
622 #define FL_ACCESS 8 /* not trying to lock, just looking */
623 #define FL_LOCKD 16 /* lock held by rpc.lockd */
624 #define FL_LEASE 32 /* lease held on this file */
625 #define FL_SLEEP 128 /* A blocking lock */
628 * The POSIX file lock owner is determined by
629 * the "struct files_struct" in the thread group
630 * (or NULL for no owner - BSD locks).
632 * Lockd stuffs a "host" pointer into this.
634 typedef struct files_struct
*fl_owner_t
;
636 struct file_lock_operations
{
637 void (*fl_insert
)(struct file_lock
*); /* lock insertion callback */
638 void (*fl_remove
)(struct file_lock
*); /* lock removal callback */
639 void (*fl_copy_lock
)(struct file_lock
*, struct file_lock
*);
640 void (*fl_release_private
)(struct file_lock
*);
643 struct lock_manager_operations
{
644 int (*fl_compare_owner
)(struct file_lock
*, struct file_lock
*);
645 void (*fl_notify
)(struct file_lock
*); /* unblock callback */
646 void (*fl_copy_lock
)(struct file_lock
*, struct file_lock
*);
647 void (*fl_release_private
)(struct file_lock
*);
648 void (*fl_break
)(struct file_lock
*);
649 int (*fl_mylease
)(struct file_lock
*, struct file_lock
*);
650 int (*fl_change
)(struct file_lock
**, int);
653 /* that will die - we need it for nfs_lock_info */
654 #include <linux/nfs_fs_i.h>
657 struct file_lock
*fl_next
; /* singly linked list for this inode */
658 struct list_head fl_link
; /* doubly linked list of all locks */
659 struct list_head fl_block
; /* circular list of blocked processes */
662 wait_queue_head_t fl_wait
;
663 struct file
*fl_file
;
664 unsigned char fl_flags
;
665 unsigned char fl_type
;
669 struct fasync_struct
* fl_fasync
; /* for lease break notifications */
670 unsigned long fl_break_time
; /* for nonblocking lease breaks */
672 struct file_lock_operations
*fl_ops
; /* Callbacks for filesystems */
673 struct lock_manager_operations
*fl_lmops
; /* Callbacks for lockmanagers */
675 struct nfs_lock_info nfs_fl
;
676 struct nfs4_lock_info nfs4_fl
;
680 /* The following constant reflects the upper bound of the file/locking space */
682 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
683 #define OFFSET_MAX INT_LIMIT(loff_t)
684 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
687 extern struct list_head file_lock_list
;
689 #include <linux/fcntl.h>
691 extern int fcntl_getlk(struct file
*, struct flock __user
*);
692 extern int fcntl_setlk(unsigned int, struct file
*, unsigned int,
693 struct flock __user
*);
695 #if BITS_PER_LONG == 32
696 extern int fcntl_getlk64(struct file
*, struct flock64 __user
*);
697 extern int fcntl_setlk64(unsigned int, struct file
*, unsigned int,
698 struct flock64 __user
*);
701 extern void send_sigio(struct fown_struct
*fown
, int fd
, int band
);
702 extern int fcntl_setlease(unsigned int fd
, struct file
*filp
, long arg
);
703 extern int fcntl_getlease(struct file
*filp
);
706 extern void locks_init_lock(struct file_lock
*);
707 extern void locks_copy_lock(struct file_lock
*, struct file_lock
*);
708 extern void locks_remove_posix(struct file
*, fl_owner_t
);
709 extern void locks_remove_flock(struct file
*);
710 extern struct file_lock
*posix_test_lock(struct file
*, struct file_lock
*);
711 extern int posix_lock_file(struct file
*, struct file_lock
*);
712 extern int posix_lock_file_wait(struct file
*, struct file_lock
*);
713 extern void posix_block_lock(struct file_lock
*, struct file_lock
*);
714 extern void posix_unblock_lock(struct file
*, struct file_lock
*);
715 extern int posix_locks_deadlock(struct file_lock
*, struct file_lock
*);
716 extern int flock_lock_file_wait(struct file
*filp
, struct file_lock
*fl
);
717 extern int __break_lease(struct inode
*inode
, unsigned int flags
);
718 extern void lease_get_mtime(struct inode
*, struct timespec
*time
);
719 extern int setlease(struct file
*, long, struct file_lock
**);
720 extern int lease_modify(struct file_lock
**, int);
721 extern int lock_may_read(struct inode
*, loff_t start
, unsigned long count
);
722 extern int lock_may_write(struct inode
*, loff_t start
, unsigned long count
);
723 extern void steal_locks(fl_owner_t from
);
725 struct fasync_struct
{
728 struct fasync_struct
*fa_next
; /* singly linked list */
729 struct file
*fa_file
;
732 #define FASYNC_MAGIC 0x4601
734 /* SMP safe fasync helpers: */
735 extern int fasync_helper(int, struct file
*, int, struct fasync_struct
**);
736 /* can be called from interrupts */
737 extern void kill_fasync(struct fasync_struct
**, int, int);
738 /* only for net: no internal synchronization */
739 extern void __kill_fasync(struct fasync_struct
*, int, int);
741 extern int f_setown(struct file
*filp
, unsigned long arg
, int force
);
742 extern void f_delown(struct file
*filp
);
743 extern int send_sigurg(struct fown_struct
*fown
);
749 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
750 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
751 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
753 extern struct list_head super_blocks
;
754 extern spinlock_t sb_lock
;
756 #define sb_entry(list) list_entry((list), struct super_block, s_list)
757 #define S_BIAS (1<<30)
759 struct list_head s_list
; /* Keep this first */
760 dev_t s_dev
; /* search index; _not_ kdev_t */
761 unsigned long s_blocksize
;
762 unsigned long s_old_blocksize
;
763 unsigned char s_blocksize_bits
;
764 unsigned char s_dirt
;
765 unsigned long long s_maxbytes
; /* Max file size */
766 struct file_system_type
*s_type
;
767 struct super_operations
*s_op
;
768 struct dquot_operations
*dq_op
;
769 struct quotactl_ops
*s_qcop
;
770 struct export_operations
*s_export_op
;
771 unsigned long s_flags
;
772 unsigned long s_magic
;
773 struct dentry
*s_root
;
774 struct rw_semaphore s_umount
;
775 struct semaphore s_lock
;
781 struct xattr_handler
**s_xattr
;
783 struct list_head s_inodes
; /* all inodes */
784 struct list_head s_dirty
; /* dirty inodes */
785 struct list_head s_io
; /* parked for writeback */
786 struct hlist_head s_anon
; /* anonymous dentries for (nfs) exporting */
787 struct list_head s_files
;
789 struct block_device
*s_bdev
;
790 struct list_head s_instances
;
791 struct quota_info s_dquot
; /* Diskquota specific options */
794 wait_queue_head_t s_wait_unfrozen
;
796 char s_id
[32]; /* Informational name */
798 void *s_fs_info
; /* Filesystem private info */
801 * The next field is for VFS *only*. No filesystems have any business
802 * even looking at it. You had been warned.
804 struct semaphore s_vfs_rename_sem
; /* Kludge */
806 /* Granuality of c/m/atime in ns.
807 Cannot be worse than a second */
811 extern struct timespec
current_fs_time(struct super_block
*sb
);
814 * Snapshotting support.
822 #define vfs_check_frozen(sb, level) \
823 wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
825 static inline void get_fs_excl(void)
827 atomic_inc(¤t
->fs_excl
);
830 static inline void put_fs_excl(void)
832 atomic_dec(¤t
->fs_excl
);
835 static inline int has_fs_excl(void)
837 return atomic_read(¤t
->fs_excl
);
842 * Superblock locking.
844 static inline void lock_super(struct super_block
* sb
)
850 static inline void unlock_super(struct super_block
* sb
)
857 * VFS helper functions..
859 extern int vfs_create(struct inode
*, struct dentry
*, int, struct nameidata
*);
860 extern int vfs_mkdir(struct inode
*, struct dentry
*, int);
861 extern int vfs_mknod(struct inode
*, struct dentry
*, int, dev_t
);
862 extern int vfs_symlink(struct inode
*, struct dentry
*, const char *, int);
863 extern int vfs_link(struct dentry
*, struct inode
*, struct dentry
*);
864 extern int vfs_rmdir(struct inode
*, struct dentry
*);
865 extern int vfs_unlink(struct inode
*, struct dentry
*);
866 extern int vfs_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
869 * VFS dentry helper functions.
871 extern void dentry_unhash(struct dentry
*dentry
);
876 * NOTE! These match bits 12..15 of stat.st_mode
877 * (ie "(i_mode >> 12) & 15").
889 #define OSYNC_METADATA (1<<0)
890 #define OSYNC_DATA (1<<1)
891 #define OSYNC_INODE (1<<2)
892 int generic_osync_inode(struct inode
*, struct address_space
*, int);
895 * This is the "filldir" function type, used by readdir() to let
896 * the kernel specify what kind of dirent layout it wants to have.
897 * This allows the kernel to read directories into kernel space or
898 * to have different dirent layouts depending on the binary type.
900 typedef int (*filldir_t
)(void *, const char *, int, loff_t
, ino_t
, unsigned);
902 struct block_device_operations
{
903 int (*open
) (struct inode
*, struct file
*);
904 int (*release
) (struct inode
*, struct file
*);
905 int (*ioctl
) (struct inode
*, struct file
*, unsigned, unsigned long);
906 long (*unlocked_ioctl
) (struct file
*, unsigned, unsigned long);
907 long (*compat_ioctl
) (struct file
*, unsigned, unsigned long);
908 int (*direct_access
) (struct block_device
*, sector_t
, unsigned long *);
909 int (*media_changed
) (struct gendisk
*);
910 int (*revalidate_disk
) (struct gendisk
*);
911 struct module
*owner
;
915 * "descriptor" for what we're up to with a read for sendfile().
916 * This allows us to use the same read code yet
917 * have multiple different users of the data that
918 * we read from a file.
920 * The simplest case just copies the data to user
933 typedef int (*read_actor_t
)(read_descriptor_t
*, struct page
*, unsigned long, unsigned long);
935 /* These macros are for out of kernel modules to test that
936 * the kernel supports the unlocked_ioctl and compat_ioctl
937 * fields in struct file_operations. */
938 #define HAVE_COMPAT_IOCTL 1
939 #define HAVE_UNLOCKED_IOCTL 1
943 * read, write, poll, fsync, readv, writev, unlocked_ioctl and compat_ioctl
944 * can be called without the big kernel lock held in all filesystems.
946 struct file_operations
{
947 struct module
*owner
;
948 loff_t (*llseek
) (struct file
*, loff_t
, int);
949 ssize_t (*read
) (struct file
*, char __user
*, size_t, loff_t
*);
950 ssize_t (*aio_read
) (struct kiocb
*, char __user
*, size_t, loff_t
);
951 ssize_t (*write
) (struct file
*, const char __user
*, size_t, loff_t
*);
952 ssize_t (*aio_write
) (struct kiocb
*, const char __user
*, size_t, loff_t
);
953 int (*readdir
) (struct file
*, void *, filldir_t
);
954 unsigned int (*poll
) (struct file
*, struct poll_table_struct
*);
955 int (*ioctl
) (struct inode
*, struct file
*, unsigned int, unsigned long);
956 long (*unlocked_ioctl
) (struct file
*, unsigned int, unsigned long);
957 long (*compat_ioctl
) (struct file
*, unsigned int, unsigned long);
958 int (*mmap
) (struct file
*, struct vm_area_struct
*);
959 int (*open
) (struct inode
*, struct file
*);
960 int (*flush
) (struct file
*);
961 int (*release
) (struct inode
*, struct file
*);
962 int (*fsync
) (struct file
*, struct dentry
*, int datasync
);
963 int (*aio_fsync
) (struct kiocb
*, int datasync
);
964 int (*fasync
) (int, struct file
*, int);
965 int (*lock
) (struct file
*, int, struct file_lock
*);
966 ssize_t (*readv
) (struct file
*, const struct iovec
*, unsigned long, loff_t
*);
967 ssize_t (*writev
) (struct file
*, const struct iovec
*, unsigned long, loff_t
*);
968 ssize_t (*sendfile
) (struct file
*, loff_t
*, size_t, read_actor_t
, void *);
969 ssize_t (*sendpage
) (struct file
*, struct page
*, int, size_t, loff_t
*, int);
970 unsigned long (*get_unmapped_area
)(struct file
*, unsigned long, unsigned long, unsigned long, unsigned long);
971 int (*check_flags
)(int);
972 int (*dir_notify
)(struct file
*filp
, unsigned long arg
);
973 int (*flock
) (struct file
*, int, struct file_lock
*);
976 struct inode_operations
{
977 int (*create
) (struct inode
*,struct dentry
*,int, struct nameidata
*);
978 struct dentry
* (*lookup
) (struct inode
*,struct dentry
*, struct nameidata
*);
979 int (*link
) (struct dentry
*,struct inode
*,struct dentry
*);
980 int (*unlink
) (struct inode
*,struct dentry
*);
981 int (*symlink
) (struct inode
*,struct dentry
*,const char *);
982 int (*mkdir
) (struct inode
*,struct dentry
*,int);
983 int (*rmdir
) (struct inode
*,struct dentry
*);
984 int (*mknod
) (struct inode
*,struct dentry
*,int,dev_t
);
985 int (*rename
) (struct inode
*, struct dentry
*,
986 struct inode
*, struct dentry
*);
987 int (*readlink
) (struct dentry
*, char __user
*,int);
988 void * (*follow_link
) (struct dentry
*, struct nameidata
*);
989 void (*put_link
) (struct dentry
*, struct nameidata
*, void *);
990 void (*truncate
) (struct inode
*);
991 int (*permission
) (struct inode
*, int, struct nameidata
*);
992 int (*setattr
) (struct dentry
*, struct iattr
*);
993 int (*getattr
) (struct vfsmount
*mnt
, struct dentry
*, struct kstat
*);
994 int (*setxattr
) (struct dentry
*, const char *,const void *,size_t,int);
995 ssize_t (*getxattr
) (struct dentry
*, const char *, void *, size_t);
996 ssize_t (*listxattr
) (struct dentry
*, char *, size_t);
997 int (*removexattr
) (struct dentry
*, const char *);
1002 extern ssize_t
vfs_read(struct file
*, char __user
*, size_t, loff_t
*);
1003 extern ssize_t
vfs_write(struct file
*, const char __user
*, size_t, loff_t
*);
1004 extern ssize_t
vfs_readv(struct file
*, const struct iovec __user
*,
1005 unsigned long, loff_t
*);
1006 extern ssize_t
vfs_writev(struct file
*, const struct iovec __user
*,
1007 unsigned long, loff_t
*);
1010 * NOTE: write_inode, delete_inode, clear_inode, put_inode can be called
1011 * without the big kernel lock held in all filesystems.
1013 struct super_operations
{
1014 struct inode
*(*alloc_inode
)(struct super_block
*sb
);
1015 void (*destroy_inode
)(struct inode
*);
1017 void (*read_inode
) (struct inode
*);
1019 void (*dirty_inode
) (struct inode
*);
1020 int (*write_inode
) (struct inode
*, int);
1021 void (*put_inode
) (struct inode
*);
1022 void (*drop_inode
) (struct inode
*);
1023 void (*delete_inode
) (struct inode
*);
1024 void (*put_super
) (struct super_block
*);
1025 void (*write_super
) (struct super_block
*);
1026 int (*sync_fs
)(struct super_block
*sb
, int wait
);
1027 void (*write_super_lockfs
) (struct super_block
*);
1028 void (*unlockfs
) (struct super_block
*);
1029 int (*statfs
) (struct super_block
*, struct kstatfs
*);
1030 int (*remount_fs
) (struct super_block
*, int *, char *);
1031 void (*clear_inode
) (struct inode
*);
1032 void (*umount_begin
) (struct super_block
*);
1034 int (*show_options
)(struct seq_file
*, struct vfsmount
*);
1036 ssize_t (*quota_read
)(struct super_block
*, int, char *, size_t, loff_t
);
1037 ssize_t (*quota_write
)(struct super_block
*, int, const char *, size_t, loff_t
);
1040 /* Inode state bits. Protected by inode_lock. */
1041 #define I_DIRTY_SYNC 1 /* Not dirty enough for O_DATASYNC */
1042 #define I_DIRTY_DATASYNC 2 /* Data-related inode changes pending */
1043 #define I_DIRTY_PAGES 4 /* Data-related inode changes pending */
1045 #define I_LOCK (1 << __I_LOCK)
1046 #define I_FREEING 16
1049 #define I_WILL_FREE 128
1051 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1053 extern void __mark_inode_dirty(struct inode
*, int);
1054 static inline void mark_inode_dirty(struct inode
*inode
)
1056 __mark_inode_dirty(inode
, I_DIRTY
);
1059 static inline void mark_inode_dirty_sync(struct inode
*inode
)
1061 __mark_inode_dirty(inode
, I_DIRTY_SYNC
);
1064 static inline void touch_atime(struct vfsmount
*mnt
, struct dentry
*dentry
)
1066 /* per-mountpoint checks will go here */
1067 update_atime(dentry
->d_inode
);
1070 static inline void file_accessed(struct file
*file
)
1072 if (!(file
->f_flags
& O_NOATIME
))
1073 touch_atime(file
->f_vfsmnt
, file
->f_dentry
);
1076 int sync_inode(struct inode
*inode
, struct writeback_control
*wbc
);
1079 * struct export_operations - for nfsd to communicate with file systems
1080 * @decode_fh: decode a file handle fragment and return a &struct dentry
1081 * @encode_fh: encode a file handle fragment from a dentry
1082 * @get_name: find the name for a given inode in a given directory
1083 * @get_parent: find the parent of a given directory
1084 * @get_dentry: find a dentry for the inode given a file handle sub-fragment
1087 * The export_operations structure provides a means for nfsd to communicate
1088 * with a particular exported file system - particularly enabling nfsd and
1089 * the filesystem to co-operate when dealing with file handles.
1091 * export_operations contains two basic operation for dealing with file
1092 * handles, decode_fh() and encode_fh(), and allows for some other
1093 * operations to be defined which standard helper routines use to get
1094 * specific information from the filesystem.
1096 * nfsd encodes information use to determine which filesystem a filehandle
1097 * applies to in the initial part of the file handle. The remainder, termed
1098 * a file handle fragment, is controlled completely by the filesystem. The
1099 * standard helper routines assume that this fragment will contain one or
1100 * two sub-fragments, one which identifies the file, and one which may be
1101 * used to identify the (a) directory containing the file.
1103 * In some situations, nfsd needs to get a dentry which is connected into a
1104 * specific part of the file tree. To allow for this, it passes the
1105 * function acceptable() together with a @context which can be used to see
1106 * if the dentry is acceptable. As there can be multiple dentrys for a
1107 * given file, the filesystem should check each one for acceptability before
1108 * looking for the next. As soon as an acceptable one is found, it should
1112 * @decode_fh is given a &struct super_block (@sb), a file handle fragment
1113 * (@fh, @fh_len) and an acceptability testing function (@acceptable,
1114 * @context). It should return a &struct dentry which refers to the same
1115 * file that the file handle fragment refers to, and which passes the
1116 * acceptability test. If it cannot, it should return a %NULL pointer if
1117 * the file was found but no acceptable &dentries were available, or a
1118 * %ERR_PTR error code indicating why it couldn't be found (e.g. %ENOENT or
1122 * @encode_fh should store in the file handle fragment @fh (using at most
1123 * @max_len bytes) information that can be used by @decode_fh to recover the
1124 * file refered to by the &struct dentry @de. If the @connectable flag is
1125 * set, the encode_fh() should store sufficient information so that a good
1126 * attempt can be made to find not only the file but also it's place in the
1127 * filesystem. This typically means storing a reference to de->d_parent in
1128 * the filehandle fragment. encode_fh() should return the number of bytes
1129 * stored or a negative error code such as %-ENOSPC
1132 * @get_name should find a name for the given @child in the given @parent
1133 * directory. The name should be stored in the @name (with the
1134 * understanding that it is already pointing to a a %NAME_MAX+1 sized
1135 * buffer. get_name() should return %0 on success, a negative error code
1136 * or error. @get_name will be called without @parent->i_sem held.
1139 * @get_parent should find the parent directory for the given @child which
1140 * is also a directory. In the event that it cannot be found, or storage
1141 * space cannot be allocated, a %ERR_PTR should be returned.
1144 * Given a &super_block (@sb) and a pointer to a file-system specific inode
1145 * identifier, possibly an inode number, (@inump) get_dentry() should find
1146 * the identified inode and return a dentry for that inode. Any suitable
1147 * dentry can be returned including, if necessary, a new dentry created with
1148 * d_alloc_root. The caller can then find any other extant dentrys by
1149 * following the d_alias links. If a new dentry was created using
1150 * d_alloc_root, DCACHE_NFSD_DISCONNECTED should be set, and the dentry
1151 * should be d_rehash()ed.
1153 * If the inode cannot be found, either a %NULL pointer or an %ERR_PTR code
1154 * can be returned. The @inump will be whatever was passed to
1155 * nfsd_find_fh_dentry() in either the @obj or @parent parameters.
1158 * get_parent is called with child->d_inode->i_sem down
1159 * get_name is not (which is possibly inconsistent)
1162 struct export_operations
{
1163 struct dentry
*(*decode_fh
)(struct super_block
*sb
, __u32
*fh
, int fh_len
, int fh_type
,
1164 int (*acceptable
)(void *context
, struct dentry
*de
),
1166 int (*encode_fh
)(struct dentry
*de
, __u32
*fh
, int *max_len
,
1169 /* the following are only called from the filesystem itself */
1170 int (*get_name
)(struct dentry
*parent
, char *name
,
1171 struct dentry
*child
);
1172 struct dentry
* (*get_parent
)(struct dentry
*child
);
1173 struct dentry
* (*get_dentry
)(struct super_block
*sb
, void *inump
);
1175 /* This is set by the exporting module to a standard helper */
1176 struct dentry
* (*find_exported_dentry
)(
1177 struct super_block
*sb
, void *obj
, void *parent
,
1178 int (*acceptable
)(void *context
, struct dentry
*de
),
1184 extern struct dentry
*
1185 find_exported_dentry(struct super_block
*sb
, void *obj
, void *parent
,
1186 int (*acceptable
)(void *context
, struct dentry
*de
),
1189 struct file_system_type
{
1192 struct super_block
*(*get_sb
) (struct file_system_type
*, int,
1193 const char *, void *);
1194 void (*kill_sb
) (struct super_block
*);
1195 struct module
*owner
;
1196 struct file_system_type
* next
;
1197 struct list_head fs_supers
;
1200 struct super_block
*get_sb_bdev(struct file_system_type
*fs_type
,
1201 int flags
, const char *dev_name
, void *data
,
1202 int (*fill_super
)(struct super_block
*, void *, int));
1203 struct super_block
*get_sb_single(struct file_system_type
*fs_type
,
1204 int flags
, void *data
,
1205 int (*fill_super
)(struct super_block
*, void *, int));
1206 struct super_block
*get_sb_nodev(struct file_system_type
*fs_type
,
1207 int flags
, void *data
,
1208 int (*fill_super
)(struct super_block
*, void *, int));
1209 void generic_shutdown_super(struct super_block
*sb
);
1210 void kill_block_super(struct super_block
*sb
);
1211 void kill_anon_super(struct super_block
*sb
);
1212 void kill_litter_super(struct super_block
*sb
);
1213 void deactivate_super(struct super_block
*sb
);
1214 int set_anon_super(struct super_block
*s
, void *data
);
1215 struct super_block
*sget(struct file_system_type
*type
,
1216 int (*test
)(struct super_block
*,void *),
1217 int (*set
)(struct super_block
*,void *),
1219 struct super_block
*get_sb_pseudo(struct file_system_type
*, char *,
1220 struct super_operations
*ops
, unsigned long);
1221 int __put_super(struct super_block
*sb
);
1222 int __put_super_and_need_restart(struct super_block
*sb
);
1223 void unnamed_dev_init(void);
1225 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1226 #define fops_get(fops) \
1227 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1228 #define fops_put(fops) \
1229 do { if (fops) module_put((fops)->owner); } while(0)
1231 extern int register_filesystem(struct file_system_type
*);
1232 extern int unregister_filesystem(struct file_system_type
*);
1233 extern struct vfsmount
*kern_mount(struct file_system_type
*);
1234 extern int may_umount_tree(struct vfsmount
*);
1235 extern int may_umount(struct vfsmount
*);
1236 extern long do_mount(char *, char *, char *, unsigned long, void *);
1238 extern int vfs_statfs(struct super_block
*, struct kstatfs
*);
1240 #define FLOCK_VERIFY_READ 1
1241 #define FLOCK_VERIFY_WRITE 2
1243 extern int locks_mandatory_locked(struct inode
*);
1244 extern int locks_mandatory_area(int, struct inode
*, struct file
*, loff_t
, size_t);
1247 * Candidates for mandatory locking have the setgid bit set
1248 * but no group execute bit - an otherwise meaningless combination.
1250 #define MANDATORY_LOCK(inode) \
1251 (IS_MANDLOCK(inode) && ((inode)->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID)
1253 static inline int locks_verify_locked(struct inode
*inode
)
1255 if (MANDATORY_LOCK(inode
))
1256 return locks_mandatory_locked(inode
);
1260 extern int rw_verify_area(int, struct file
*, loff_t
*, size_t);
1262 static inline int locks_verify_truncate(struct inode
*inode
,
1266 if (inode
->i_flock
&& MANDATORY_LOCK(inode
))
1267 return locks_mandatory_area(
1268 FLOCK_VERIFY_WRITE
, inode
, filp
,
1269 size
< inode
->i_size
? size
: inode
->i_size
,
1270 (size
< inode
->i_size
? inode
->i_size
- size
1271 : size
- inode
->i_size
)
1276 static inline int break_lease(struct inode
*inode
, unsigned int mode
)
1279 return __break_lease(inode
, mode
);
1285 extern int do_truncate(struct dentry
*, loff_t start
);
1286 extern long do_sys_open(const char __user
*filename
, int flags
, int mode
);
1287 extern struct file
*filp_open(const char *, int, int);
1288 extern struct file
* dentry_open(struct dentry
*, struct vfsmount
*, int);
1289 extern int filp_close(struct file
*, fl_owner_t id
);
1290 extern char * getname(const char __user
*);
1293 extern void __init
vfs_caches_init_early(void);
1294 extern void __init
vfs_caches_init(unsigned long);
1296 #define __getname() kmem_cache_alloc(names_cachep, SLAB_KERNEL)
1297 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
1298 #ifndef CONFIG_AUDITSYSCALL
1299 #define putname(name) __putname(name)
1301 extern void putname(const char *name
);
1304 extern int register_blkdev(unsigned int, const char *);
1305 extern int unregister_blkdev(unsigned int, const char *);
1306 extern struct block_device
*bdget(dev_t
);
1307 extern void bd_set_size(struct block_device
*, loff_t size
);
1308 extern void bd_forget(struct inode
*inode
);
1309 extern void bdput(struct block_device
*);
1310 extern struct block_device
*open_by_devnum(dev_t
, unsigned);
1311 extern struct file_operations def_blk_fops
;
1312 extern struct address_space_operations def_blk_aops
;
1313 extern struct file_operations def_chr_fops
;
1314 extern struct file_operations bad_sock_fops
;
1315 extern struct file_operations def_fifo_fops
;
1316 extern int ioctl_by_bdev(struct block_device
*, unsigned, unsigned long);
1317 extern int blkdev_ioctl(struct inode
*, struct file
*, unsigned, unsigned long);
1318 extern long compat_blkdev_ioctl(struct file
*, unsigned, unsigned long);
1319 extern int blkdev_get(struct block_device
*, mode_t
, unsigned);
1320 extern int blkdev_put(struct block_device
*);
1321 extern int bd_claim(struct block_device
*, void *);
1322 extern void bd_release(struct block_device
*);
1325 extern int alloc_chrdev_region(dev_t
*, unsigned, unsigned, const char *);
1326 extern int register_chrdev_region(dev_t
, unsigned, const char *);
1327 extern int register_chrdev(unsigned int, const char *,
1328 struct file_operations
*);
1329 extern int unregister_chrdev(unsigned int, const char *);
1330 extern void unregister_chrdev_region(dev_t
, unsigned);
1331 extern int chrdev_open(struct inode
*, struct file
*);
1333 /* fs/block_dev.c */
1334 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
1335 extern const char *__bdevname(dev_t
, char *buffer
);
1336 extern const char *bdevname(struct block_device
*bdev
, char *buffer
);
1337 extern struct block_device
*lookup_bdev(const char *);
1338 extern struct block_device
*open_bdev_excl(const char *, int, void *);
1339 extern void close_bdev_excl(struct block_device
*);
1341 extern void init_special_inode(struct inode
*, umode_t
, dev_t
);
1343 /* Invalid inode operations -- fs/bad_inode.c */
1344 extern void make_bad_inode(struct inode
*);
1345 extern int is_bad_inode(struct inode
*);
1347 extern struct file_operations read_fifo_fops
;
1348 extern struct file_operations write_fifo_fops
;
1349 extern struct file_operations rdwr_fifo_fops
;
1350 extern struct file_operations read_pipe_fops
;
1351 extern struct file_operations write_pipe_fops
;
1352 extern struct file_operations rdwr_pipe_fops
;
1354 extern int fs_may_remount_ro(struct super_block
*);
1357 * return READ, READA, or WRITE
1359 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
1362 * return data direction, READ or WRITE
1364 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
1366 extern int check_disk_change(struct block_device
*);
1367 extern int invalidate_inodes(struct super_block
*);
1368 extern int __invalidate_device(struct block_device
*);
1369 extern int invalidate_partition(struct gendisk
*, int);
1370 unsigned long invalidate_mapping_pages(struct address_space
*mapping
,
1371 pgoff_t start
, pgoff_t end
);
1372 unsigned long invalidate_inode_pages(struct address_space
*mapping
);
1373 static inline void invalidate_remote_inode(struct inode
*inode
)
1375 if (S_ISREG(inode
->i_mode
) || S_ISDIR(inode
->i_mode
) ||
1376 S_ISLNK(inode
->i_mode
))
1377 invalidate_inode_pages(inode
->i_mapping
);
1379 extern int invalidate_inode_pages2(struct address_space
*mapping
);
1380 extern int invalidate_inode_pages2_range(struct address_space
*mapping
,
1381 pgoff_t start
, pgoff_t end
);
1382 extern int write_inode_now(struct inode
*, int);
1383 extern int filemap_fdatawrite(struct address_space
*);
1384 extern int filemap_flush(struct address_space
*);
1385 extern int filemap_fdatawait(struct address_space
*);
1386 extern int filemap_write_and_wait(struct address_space
*mapping
);
1387 extern int filemap_write_and_wait_range(struct address_space
*mapping
,
1388 loff_t lstart
, loff_t lend
);
1389 extern void sync_supers(void);
1390 extern void sync_filesystems(int wait
);
1391 extern void emergency_sync(void);
1392 extern void emergency_remount(void);
1393 extern int do_remount_sb(struct super_block
*sb
, int flags
,
1394 void *data
, int force
);
1395 extern sector_t
bmap(struct inode
*, sector_t
);
1396 extern int notify_change(struct dentry
*, struct iattr
*);
1397 extern int permission(struct inode
*, int, struct nameidata
*);
1398 extern int generic_permission(struct inode
*, int,
1399 int (*check_acl
)(struct inode
*, int));
1401 extern int get_write_access(struct inode
*);
1402 extern int deny_write_access(struct file
*);
1403 static inline void put_write_access(struct inode
* inode
)
1405 atomic_dec(&inode
->i_writecount
);
1407 static inline void allow_write_access(struct file
*file
)
1410 atomic_inc(&file
->f_dentry
->d_inode
->i_writecount
);
1412 extern int do_pipe(int *);
1414 extern int open_namei(const char *, int, int, struct nameidata
*);
1415 extern int may_open(struct nameidata
*, int, int);
1417 extern int kernel_read(struct file
*, unsigned long, char *, unsigned long);
1418 extern struct file
* open_exec(const char *);
1420 /* fs/dcache.c -- generic fs support functions */
1421 extern int is_subdir(struct dentry
*, struct dentry
*);
1422 extern ino_t
find_inode_number(struct dentry
*, struct qstr
*);
1424 #include <linux/err.h>
1426 /* needed for stackable file system support */
1427 extern loff_t
default_llseek(struct file
*file
, loff_t offset
, int origin
);
1429 extern loff_t
vfs_llseek(struct file
*file
, loff_t offset
, int origin
);
1431 extern void inode_init_once(struct inode
*);
1432 extern void iput(struct inode
*);
1433 extern struct inode
* igrab(struct inode
*);
1434 extern ino_t
iunique(struct super_block
*, ino_t
);
1435 extern int inode_needs_sync(struct inode
*inode
);
1436 extern void generic_delete_inode(struct inode
*inode
);
1437 extern void generic_drop_inode(struct inode
*inode
);
1439 extern struct inode
*ilookup5_nowait(struct super_block
*sb
,
1440 unsigned long hashval
, int (*test
)(struct inode
*, void *),
1442 extern struct inode
*ilookup5(struct super_block
*sb
, unsigned long hashval
,
1443 int (*test
)(struct inode
*, void *), void *data
);
1444 extern struct inode
*ilookup(struct super_block
*sb
, unsigned long ino
);
1446 extern struct inode
* iget5_locked(struct super_block
*, unsigned long, int (*test
)(struct inode
*, void *), int (*set
)(struct inode
*, void *), void *);
1447 extern struct inode
* iget_locked(struct super_block
*, unsigned long);
1448 extern void unlock_new_inode(struct inode
*);
1450 static inline struct inode
*iget(struct super_block
*sb
, unsigned long ino
)
1452 struct inode
*inode
= iget_locked(sb
, ino
);
1454 if (inode
&& (inode
->i_state
& I_NEW
)) {
1455 sb
->s_op
->read_inode(inode
);
1456 unlock_new_inode(inode
);
1462 extern void __iget(struct inode
* inode
);
1463 extern void clear_inode(struct inode
*);
1464 extern void destroy_inode(struct inode
*);
1465 extern struct inode
*new_inode(struct super_block
*);
1466 extern int remove_suid(struct dentry
*);
1467 extern void remove_dquot_ref(struct super_block
*, int, struct list_head
*);
1468 extern struct semaphore iprune_sem
;
1470 extern void __insert_inode_hash(struct inode
*, unsigned long hashval
);
1471 extern void remove_inode_hash(struct inode
*);
1472 static inline void insert_inode_hash(struct inode
*inode
) {
1473 __insert_inode_hash(inode
, inode
->i_ino
);
1476 extern struct file
* get_empty_filp(void);
1477 extern void file_move(struct file
*f
, struct list_head
*list
);
1478 extern void file_kill(struct file
*f
);
1480 extern void submit_bio(int, struct bio
*);
1481 extern int bdev_read_only(struct block_device
*);
1482 extern int set_blocksize(struct block_device
*, int);
1483 extern int sb_set_blocksize(struct super_block
*, int);
1484 extern int sb_min_blocksize(struct super_block
*, int);
1486 extern int generic_file_mmap(struct file
*, struct vm_area_struct
*);
1487 extern int generic_file_readonly_mmap(struct file
*, struct vm_area_struct
*);
1488 extern int file_read_actor(read_descriptor_t
* desc
, struct page
*page
, unsigned long offset
, unsigned long size
);
1489 extern int file_send_actor(read_descriptor_t
* desc
, struct page
*page
, unsigned long offset
, unsigned long size
);
1490 extern ssize_t
generic_file_read(struct file
*, char __user
*, size_t, loff_t
*);
1491 int generic_write_checks(struct file
*file
, loff_t
*pos
, size_t *count
, int isblk
);
1492 extern ssize_t
generic_file_write(struct file
*, const char __user
*, size_t, loff_t
*);
1493 extern ssize_t
generic_file_aio_read(struct kiocb
*, char __user
*, size_t, loff_t
);
1494 extern ssize_t
__generic_file_aio_read(struct kiocb
*, const struct iovec
*, unsigned long, loff_t
*);
1495 extern ssize_t
generic_file_aio_write(struct kiocb
*, const char __user
*, size_t, loff_t
);
1496 extern ssize_t
generic_file_aio_write_nolock(struct kiocb
*, const struct iovec
*,
1497 unsigned long, loff_t
*);
1498 extern ssize_t
generic_file_direct_write(struct kiocb
*, const struct iovec
*,
1499 unsigned long *, loff_t
, loff_t
*, size_t, size_t);
1500 extern ssize_t
generic_file_buffered_write(struct kiocb
*, const struct iovec
*,
1501 unsigned long, loff_t
, loff_t
*, size_t, ssize_t
);
1502 extern ssize_t
do_sync_read(struct file
*filp
, char __user
*buf
, size_t len
, loff_t
*ppos
);
1503 extern ssize_t
do_sync_write(struct file
*filp
, const char __user
*buf
, size_t len
, loff_t
*ppos
);
1504 ssize_t
generic_file_write_nolock(struct file
*file
, const struct iovec
*iov
,
1505 unsigned long nr_segs
, loff_t
*ppos
);
1506 extern ssize_t
generic_file_sendfile(struct file
*, loff_t
*, size_t, read_actor_t
, void *);
1507 extern void do_generic_mapping_read(struct address_space
*mapping
,
1508 struct file_ra_state
*, struct file
*,
1509 loff_t
*, read_descriptor_t
*, read_actor_t
);
1511 file_ra_state_init(struct file_ra_state
*ra
, struct address_space
*mapping
);
1512 extern ssize_t
generic_file_readv(struct file
*filp
, const struct iovec
*iov
,
1513 unsigned long nr_segs
, loff_t
*ppos
);
1514 ssize_t
generic_file_writev(struct file
*filp
, const struct iovec
*iov
,
1515 unsigned long nr_segs
, loff_t
*ppos
);
1516 extern loff_t
no_llseek(struct file
*file
, loff_t offset
, int origin
);
1517 extern loff_t
generic_file_llseek(struct file
*file
, loff_t offset
, int origin
);
1518 extern loff_t
remote_llseek(struct file
*file
, loff_t offset
, int origin
);
1519 extern int generic_file_open(struct inode
* inode
, struct file
* filp
);
1520 extern int nonseekable_open(struct inode
* inode
, struct file
* filp
);
1522 #ifdef CONFIG_FS_XIP
1523 extern ssize_t
xip_file_read(struct file
*filp
, char __user
*buf
, size_t len
,
1525 extern ssize_t
xip_file_sendfile(struct file
*in_file
, loff_t
*ppos
,
1526 size_t count
, read_actor_t actor
,
1528 extern int xip_file_mmap(struct file
* file
, struct vm_area_struct
* vma
);
1529 extern ssize_t
xip_file_write(struct file
*filp
, const char __user
*buf
,
1530 size_t len
, loff_t
*ppos
);
1531 extern int xip_truncate_page(struct address_space
*mapping
, loff_t from
);
1533 static inline int xip_truncate_page(struct address_space
*mapping
, loff_t from
)
1539 static inline void do_generic_file_read(struct file
* filp
, loff_t
*ppos
,
1540 read_descriptor_t
* desc
,
1543 do_generic_mapping_read(filp
->f_mapping
,
1551 ssize_t
__blockdev_direct_IO(int rw
, struct kiocb
*iocb
, struct inode
*inode
,
1552 struct block_device
*bdev
, const struct iovec
*iov
, loff_t offset
,
1553 unsigned long nr_segs
, get_blocks_t get_blocks
, dio_iodone_t end_io
,
1557 DIO_LOCKING
= 1, /* need locking between buffered and direct access */
1558 DIO_NO_LOCKING
, /* bdev; no locking at all between buffered/direct */
1559 DIO_OWN_LOCKING
, /* filesystem locks buffered and direct internally */
1562 static inline ssize_t
blockdev_direct_IO(int rw
, struct kiocb
*iocb
,
1563 struct inode
*inode
, struct block_device
*bdev
, const struct iovec
*iov
,
1564 loff_t offset
, unsigned long nr_segs
, get_blocks_t get_blocks
,
1565 dio_iodone_t end_io
)
1567 return __blockdev_direct_IO(rw
, iocb
, inode
, bdev
, iov
, offset
,
1568 nr_segs
, get_blocks
, end_io
, DIO_LOCKING
);
1571 static inline ssize_t
blockdev_direct_IO_no_locking(int rw
, struct kiocb
*iocb
,
1572 struct inode
*inode
, struct block_device
*bdev
, const struct iovec
*iov
,
1573 loff_t offset
, unsigned long nr_segs
, get_blocks_t get_blocks
,
1574 dio_iodone_t end_io
)
1576 return __blockdev_direct_IO(rw
, iocb
, inode
, bdev
, iov
, offset
,
1577 nr_segs
, get_blocks
, end_io
, DIO_NO_LOCKING
);
1580 static inline ssize_t
blockdev_direct_IO_own_locking(int rw
, struct kiocb
*iocb
,
1581 struct inode
*inode
, struct block_device
*bdev
, const struct iovec
*iov
,
1582 loff_t offset
, unsigned long nr_segs
, get_blocks_t get_blocks
,
1583 dio_iodone_t end_io
)
1585 return __blockdev_direct_IO(rw
, iocb
, inode
, bdev
, iov
, offset
,
1586 nr_segs
, get_blocks
, end_io
, DIO_OWN_LOCKING
);
1589 extern struct file_operations generic_ro_fops
;
1591 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
1593 extern int vfs_readlink(struct dentry
*, char __user
*, int, const char *);
1594 extern int vfs_follow_link(struct nameidata
*, const char *);
1595 extern int page_readlink(struct dentry
*, char __user
*, int);
1596 extern void *page_follow_link_light(struct dentry
*, struct nameidata
*);
1597 extern void page_put_link(struct dentry
*, struct nameidata
*, void *);
1598 extern int page_symlink(struct inode
*inode
, const char *symname
, int len
);
1599 extern struct inode_operations page_symlink_inode_operations
;
1600 extern int generic_readlink(struct dentry
*, char __user
*, int);
1601 extern void generic_fillattr(struct inode
*, struct kstat
*);
1602 extern int vfs_getattr(struct vfsmount
*, struct dentry
*, struct kstat
*);
1603 void inode_add_bytes(struct inode
*inode
, loff_t bytes
);
1604 void inode_sub_bytes(struct inode
*inode
, loff_t bytes
);
1605 loff_t
inode_get_bytes(struct inode
*inode
);
1606 void inode_set_bytes(struct inode
*inode
, loff_t bytes
);
1608 extern int vfs_readdir(struct file
*, filldir_t
, void *);
1610 extern int vfs_stat(char __user
*, struct kstat
*);
1611 extern int vfs_lstat(char __user
*, struct kstat
*);
1612 extern int vfs_fstat(unsigned int, struct kstat
*);
1614 extern int vfs_ioctl(struct file
*, unsigned int, unsigned int, unsigned long);
1616 extern struct file_system_type
*get_fs_type(const char *name
);
1617 extern struct super_block
*get_super(struct block_device
*);
1618 extern struct super_block
*user_get_super(dev_t
);
1619 extern void drop_super(struct super_block
*sb
);
1621 extern int dcache_dir_open(struct inode
*, struct file
*);
1622 extern int dcache_dir_close(struct inode
*, struct file
*);
1623 extern loff_t
dcache_dir_lseek(struct file
*, loff_t
, int);
1624 extern int dcache_readdir(struct file
*, void *, filldir_t
);
1625 extern int simple_getattr(struct vfsmount
*, struct dentry
*, struct kstat
*);
1626 extern int simple_statfs(struct super_block
*, struct kstatfs
*);
1627 extern int simple_link(struct dentry
*, struct inode
*, struct dentry
*);
1628 extern int simple_unlink(struct inode
*, struct dentry
*);
1629 extern int simple_rmdir(struct inode
*, struct dentry
*);
1630 extern int simple_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
1631 extern int simple_sync_file(struct file
*, struct dentry
*, int);
1632 extern int simple_empty(struct dentry
*);
1633 extern int simple_readpage(struct file
*file
, struct page
*page
);
1634 extern int simple_prepare_write(struct file
*file
, struct page
*page
,
1635 unsigned offset
, unsigned to
);
1636 extern int simple_commit_write(struct file
*file
, struct page
*page
,
1637 unsigned offset
, unsigned to
);
1639 extern struct dentry
*simple_lookup(struct inode
*, struct dentry
*, struct nameidata
*);
1640 extern ssize_t
generic_read_dir(struct file
*, char __user
*, size_t, loff_t
*);
1641 extern struct file_operations simple_dir_operations
;
1642 extern struct inode_operations simple_dir_inode_operations
;
1643 struct tree_descr
{ char *name
; struct file_operations
*ops
; int mode
; };
1644 struct dentry
*d_alloc_name(struct dentry
*, const char *);
1645 extern int simple_fill_super(struct super_block
*, int, struct tree_descr
*);
1646 extern int simple_pin_fs(char *name
, struct vfsmount
**mount
, int *count
);
1647 extern void simple_release_fs(struct vfsmount
**mount
, int *count
);
1649 extern ssize_t
simple_read_from_buffer(void __user
*, size_t, loff_t
*, const void *, size_t);
1651 extern int inode_change_ok(struct inode
*, struct iattr
*);
1652 extern int __must_check
inode_setattr(struct inode
*, struct iattr
*);
1654 extern void inode_update_time(struct inode
*inode
, int ctime_too
);
1656 static inline ino_t
parent_ino(struct dentry
*dentry
)
1660 spin_lock(&dentry
->d_lock
);
1661 res
= dentry
->d_parent
->d_inode
->i_ino
;
1662 spin_unlock(&dentry
->d_lock
);
1667 extern int unshare_files(void);
1669 /* Transaction based IO helpers */
1672 * An argresp is stored in an allocated page and holds the
1673 * size of the argument or response, along with its content
1675 struct simple_transaction_argresp
{
1680 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
1682 char *simple_transaction_get(struct file
*file
, const char __user
*buf
,
1684 ssize_t
simple_transaction_read(struct file
*file
, char __user
*buf
,
1685 size_t size
, loff_t
*pos
);
1686 int simple_transaction_release(struct inode
*inode
, struct file
*file
);
1688 static inline void simple_transaction_set(struct file
*file
, size_t n
)
1690 struct simple_transaction_argresp
*ar
= file
->private_data
;
1692 BUG_ON(n
> SIMPLE_TRANSACTION_LIMIT
);
1695 * The barrier ensures that ar->size will really remain zero until
1696 * ar->data is ready for reading.
1703 * simple attribute files
1705 * These attributes behave similar to those in sysfs:
1707 * Writing to an attribute immediately sets a value, an open file can be
1708 * written to multiple times.
1710 * Reading from an attribute creates a buffer from the value that might get
1711 * read with multiple read calls. When the attribute has been read
1712 * completely, no further read calls are possible until the file is opened
1715 * All attributes contain a text representation of a numeric value
1716 * that are accessed with the get() and set() functions.
1718 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
1719 static int __fops ## _open(struct inode *inode, struct file *file) \
1721 __simple_attr_check_format(__fmt, 0ull); \
1722 return simple_attr_open(inode, file, __get, __set, __fmt); \
1724 static struct file_operations __fops = { \
1725 .owner = THIS_MODULE, \
1726 .open = __fops ## _open, \
1727 .release = simple_attr_close, \
1728 .read = simple_attr_read, \
1729 .write = simple_attr_write, \
1732 static inline void __attribute__((format(printf
, 1, 2)))
1733 __simple_attr_check_format(const char *fmt
, ...)
1735 /* don't do anything, just let the compiler check the arguments; */
1738 int simple_attr_open(struct inode
*inode
, struct file
*file
,
1739 u64 (*get
)(void *), void (*set
)(void *, u64
),
1741 int simple_attr_close(struct inode
*inode
, struct file
*file
);
1742 ssize_t
simple_attr_read(struct file
*file
, char __user
*buf
,
1743 size_t len
, loff_t
*ppos
);
1744 ssize_t
simple_attr_write(struct file
*file
, const char __user
*buf
,
1745 size_t len
, loff_t
*ppos
);
1748 #ifdef CONFIG_SECURITY
1749 static inline char *alloc_secdata(void)
1751 return (char *)get_zeroed_page(GFP_KERNEL
);
1754 static inline void free_secdata(void *secdata
)
1756 free_page((unsigned long)secdata
);
1759 static inline char *alloc_secdata(void)
1764 static inline void free_secdata(void *secdata
)
1766 #endif /* CONFIG_SECURITY */
1768 #endif /* __KERNEL__ */
1769 #endif /* _LINUX_FS_H */