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>
14 * It's silly to have NR_OPEN bigger than NR_FILE, but you can change
15 * the file limit at runtime and only root can increase the per-process
16 * nr_file rlimit, so it's safe to set up a ridiculously high absolute
17 * upper limit on files-per-process.
19 * Some programs (notably those using select()) may have to be
20 * recompiled to take full advantage of the new limits..
23 /* Fixed constants first: */
25 #define NR_OPEN (1024*1024) /* Absolute upper limit on fd num */
26 #define INR_OPEN 1024 /* Initial setting for nfile rlimits */
28 #define BLOCK_SIZE_BITS 10
29 #define BLOCK_SIZE (1<<BLOCK_SIZE_BITS)
31 /* And dynamically-tunable limits and defaults: */
32 struct files_stat_struct {
33 int nr_files; /* read only */
34 int nr_free_files; /* read only */
35 int max_files; /* tunable */
37 extern struct files_stat_struct files_stat;
38 extern int get_max_files(void);
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 */
68 /* File is being opened for execution. Primary users of this flag are
69 distributed filesystems that can use it to achieve correct ETXTBUSY
70 behavior for cross-node execution/opening_for_writing of files */
77 #define READA 2 /* read-ahead - don't block if no resources */
78 #define SWRITE 3 /* for ll_rw_block() - wait for buffer lock */
79 #define SPECIAL 4 /* For non-blockdevice requests in request queue */
80 #define READ_SYNC (READ | (1 << BIO_RW_SYNC))
81 #define WRITE_SYNC (WRITE | (1 << BIO_RW_SYNC))
82 #define WRITE_BARRIER ((1 << BIO_RW) | (1 << BIO_RW_BARRIER))
88 /* public flags for file_system_type */
89 #define FS_REQUIRES_DEV 1
90 #define FS_BINARY_MOUNTDATA 2
91 #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
92 #define FS_ODD_RENAME 32768 /* Temporary stuff; will go away as soon
93 * as nfs_rename() will be cleaned up
96 * These are the fs-independent mount-flags: up to 32 flags are supported
98 #define MS_RDONLY 1 /* Mount read-only */
99 #define MS_NOSUID 2 /* Ignore suid and sgid bits */
100 #define MS_NODEV 4 /* Disallow access to device special files */
101 #define MS_NOEXEC 8 /* Disallow program execution */
102 #define MS_SYNCHRONOUS 16 /* Writes are synced at once */
103 #define MS_REMOUNT 32 /* Alter flags of a mounted FS */
104 #define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
105 #define MS_DIRSYNC 128 /* Directory modifications are synchronous */
106 #define MS_NOATIME 1024 /* Do not update access times. */
107 #define MS_NODIRATIME 2048 /* Do not update directory access times */
111 #define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
112 MS_VERBOSE is deprecated. */
113 #define MS_SILENT 32768
114 #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
115 #define MS_UNBINDABLE (1<<17) /* change to unbindable */
116 #define MS_PRIVATE (1<<18) /* change to private */
117 #define MS_SLAVE (1<<19) /* change to slave */
118 #define MS_SHARED (1<<20) /* change to shared */
119 #define MS_ACTIVE (1<<30)
120 #define MS_NOUSER (1<<31)
123 * Superblock flags that can be altered by MS_REMOUNT
125 #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK)
128 * Old magic mount flag and mask
130 #define MS_MGC_VAL 0xC0ED0000
131 #define MS_MGC_MSK 0xffff0000
133 /* Inode flags - they have nothing to superblock flags now */
135 #define S_SYNC 1 /* Writes are synced at once */
136 #define S_NOATIME 2 /* Do not update access times */
137 #define S_APPEND 4 /* Append-only file */
138 #define S_IMMUTABLE 8 /* Immutable file */
139 #define S_DEAD 16 /* removed, but still open directory */
140 #define S_NOQUOTA 32 /* Inode is not counted to quota */
141 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
142 #define S_NOCMTIME 128 /* Do not update file c/mtime */
143 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
144 #define S_PRIVATE 512 /* Inode is fs-internal */
147 * Note that nosuid etc flags are inode-specific: setting some file-system
148 * flags just means all the inodes inherit those flags by default. It might be
149 * possible to override it selectively if you really wanted to with some
150 * ioctl() that is not currently implemented.
152 * Exception: MS_RDONLY is always applied to the entire file system.
154 * Unfortunately, it is possible to change a filesystems flags with it mounted
155 * with files in use. This means that all of the inodes will not have their
156 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
157 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
159 #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
161 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
162 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
163 ((inode)->i_flags & S_SYNC))
164 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
165 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
166 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
168 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
169 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
170 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
171 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
173 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
174 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
175 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
176 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
178 /* the read-only stuff doesn't really belong here, but any other place is
179 probably as bad and I don't want to create yet another include file. */
181 #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
182 #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
183 #define BLKRRPART _IO(0x12,95) /* re-read partition table */
184 #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
185 #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
186 #define BLKRASET _IO(0x12,98) /* set read ahead for block device */
187 #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
188 #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
189 #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
190 #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
191 #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
192 #define BLKSSZGET _IO(0x12,104)/* get block device sector size */
194 #define BLKPG _IO(0x12,105)/* See blkpg.h */
196 /* Some people are morons. Do not use sizeof! */
198 #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
199 #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
200 /* This was here just to show that the number is taken -
201 probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
203 /* A jump here: 108-111 have been used for various private purposes. */
204 #define BLKBSZGET _IOR(0x12,112,size_t)
205 #define BLKBSZSET _IOW(0x12,113,size_t)
206 #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
207 #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
208 #define BLKTRACESTART _IO(0x12,116)
209 #define BLKTRACESTOP _IO(0x12,117)
210 #define BLKTRACETEARDOWN _IO(0x12,118)
212 #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
213 #define FIBMAP _IO(0x00,1) /* bmap access */
214 #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
218 #include <linux/linkage.h>
219 #include <linux/wait.h>
220 #include <linux/types.h>
221 #include <linux/kdev_t.h>
222 #include <linux/dcache.h>
223 #include <linux/stat.h>
224 #include <linux/cache.h>
225 #include <linux/kobject.h>
226 #include <linux/list.h>
227 #include <linux/radix-tree.h>
228 #include <linux/prio_tree.h>
229 #include <linux/init.h>
230 #include <linux/sched.h>
231 #include <linux/mutex.h>
233 #include <asm/atomic.h>
234 #include <asm/semaphore.h>
235 #include <asm/byteorder.h>
241 struct pipe_inode_info;
242 struct poll_table_struct;
244 struct vm_area_struct;
247 extern void __init inode_init(unsigned long);
248 extern void __init inode_init_early(void);
249 extern void __init mnt_init(unsigned long);
250 extern void __init files_init(unsigned long);
253 typedef int (get_block_t)(struct inode *inode, sector_t iblock,
254 struct buffer_head *bh_result, int create);
255 typedef int (get_blocks_t)(struct inode *inode, sector_t iblock,
256 unsigned long max_blocks,
257 struct buffer_head *bh_result, int create);
258 typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
259 ssize_t bytes, void *private);
262 * Attribute flags. These should be or-ed together to figure out what
269 #define ATTR_ATIME 16
270 #define ATTR_MTIME 32
271 #define ATTR_CTIME 64
272 #define ATTR_ATIME_SET 128
273 #define ATTR_MTIME_SET 256
274 #define ATTR_FORCE 512 /* Not a change, but a change it */
275 #define ATTR_ATTR_FLAG 1024
276 #define ATTR_KILL_SUID 2048
277 #define ATTR_KILL_SGID 4096
278 #define ATTR_FILE 8192
281 * This is the Inode Attributes structure, used for notify_change(). It
282 * uses the above definitions as flags, to know which values have changed.
283 * Also, in this manner, a Filesystem can look at only the values it cares
284 * about. Basically, these are the attributes that the VFS layer can
285 * request to change from the FS layer.
287 * Derek Atkins <warlord@MIT.EDU> 94-10-20
290 unsigned int ia_valid;
295 struct timespec ia_atime;
296 struct timespec ia_mtime;
297 struct timespec ia_ctime;
300 * Not an attribute, but an auxilary info for filesystems wanting to
301 * implement an ftruncate() like method. NOTE: filesystem should
302 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
304 struct file *ia_file;
308 * Includes for diskquotas.
310 #include <linux/quota.h>
313 * enum positive_aop_returns - aop return codes with specific semantics
315 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
316 * completed, that the page is still locked, and
317 * should be considered active. The VM uses this hint
318 * to return the page to the active list -- it won't
319 * be a candidate for writeback again in the near
320 * future. Other callers must be careful to unlock
321 * the page if they get this return. Returned by
324 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
325 * unlocked it and the page might have been truncated.
326 * The caller should back up to acquiring a new page and
327 * trying again. The aop will be taking reasonable
328 * precautions not to livelock. If the caller held a page
329 * reference, it should drop it before retrying. Returned
330 * by readpage(), prepare_write(), and commit_write().
332 * address_space_operation functions return these large constants to indicate
333 * special semantics to the caller. These are much larger than the bytes in a
334 * page to allow for functions that return the number of bytes operated on in a
338 enum positive_aop_returns {
339 AOP_WRITEPAGE_ACTIVATE = 0x80000,
340 AOP_TRUNCATED_PAGE = 0x80001,
344 * oh the beauties of C type declarations.
347 struct address_space;
348 struct writeback_control;
350 struct address_space_operations {
351 int (*writepage)(struct page *page, struct writeback_control *wbc);
352 int (*readpage)(struct file *, struct page *);
353 int (*sync_page)(struct page *);
355 /* Write back some dirty pages from this mapping. */
356 int (*writepages)(struct address_space *, struct writeback_control *);
358 /* Set a page dirty. Return true if this dirtied it */
359 int (*set_page_dirty)(struct page *page);
361 int (*readpages)(struct file *filp, struct address_space *mapping,
362 struct list_head *pages, unsigned nr_pages);
365 * ext3 requires that a successful prepare_write() call be followed
366 * by a commit_write() call - they must be balanced
368 int (*prepare_write)(struct file *, struct page *, unsigned, unsigned);
369 int (*commit_write)(struct file *, struct page *, unsigned, unsigned);
370 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
371 sector_t (*bmap)(struct address_space *, sector_t);
372 int (*invalidatepage) (struct page *, unsigned long);
373 int (*releasepage) (struct page *, gfp_t);
374 ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
375 loff_t offset, unsigned long nr_segs);
376 struct page* (*get_xip_page)(struct address_space *, sector_t,
378 /* migrate the contents of a page to the specified target */
379 int (*migratepage) (struct page *, struct page *);
382 struct backing_dev_info;
383 struct address_space {
384 struct inode *host; /* owner: inode, block_device */
385 struct radix_tree_root page_tree; /* radix tree of all pages */
386 rwlock_t tree_lock; /* and rwlock protecting it */
387 unsigned int i_mmap_writable;/* count VM_SHARED mappings */
388 struct prio_tree_root i_mmap; /* tree of private and shared mappings */
389 struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
390 spinlock_t i_mmap_lock; /* protect tree, count, list */
391 unsigned int truncate_count; /* Cover race condition with truncate */
392 unsigned long nrpages; /* number of total pages */
393 pgoff_t writeback_index;/* writeback starts here */
394 struct address_space_operations *a_ops; /* methods */
395 unsigned long flags; /* error bits/gfp mask */
396 struct backing_dev_info *backing_dev_info; /* device readahead, etc */
397 spinlock_t private_lock; /* for use by the address_space */
398 struct list_head private_list; /* ditto */
399 struct address_space *assoc_mapping; /* ditto */
400 } __attribute__((aligned(sizeof(long))));
402 * On most architectures that alignment is already the case; but
403 * must be enforced here for CRIS, to let the least signficant bit
404 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
407 struct block_device {
408 dev_t bd_dev; /* not a kdev_t - it's a search key */
409 struct inode * bd_inode; /* will die */
411 struct mutex bd_mutex; /* open/close mutex */
412 struct mutex bd_mount_mutex; /* mount mutex */
413 struct list_head bd_inodes;
416 struct block_device * bd_contains;
417 unsigned bd_block_size;
418 struct hd_struct * bd_part;
419 /* number of times partitions within this device have been opened. */
420 unsigned bd_part_count;
422 struct gendisk * bd_disk;
423 struct list_head bd_list;
424 struct backing_dev_info *bd_inode_backing_dev_info;
426 * Private data. You must have bd_claim'ed the block_device
427 * to use this. NOTE: bd_claim allows an owner to claim
428 * the same device multiple times, the owner must take special
429 * care to not mess up bd_private for that case.
431 unsigned long bd_private;
435 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
438 #define PAGECACHE_TAG_DIRTY 0
439 #define PAGECACHE_TAG_WRITEBACK 1
441 int mapping_tagged(struct address_space *mapping, int tag);
444 * Might pages of this file be mapped into userspace?
446 static inline int mapping_mapped(struct address_space *mapping)
448 return !prio_tree_empty(&mapping->i_mmap) ||
449 !list_empty(&mapping->i_mmap_nonlinear);
453 * Might pages of this file have been modified in userspace?
454 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
455 * marks vma as VM_SHARED if it is shared, and the file was opened for
456 * writing i.e. vma may be mprotected writable even if now readonly.
458 static inline int mapping_writably_mapped(struct address_space *mapping)
460 return mapping->i_mmap_writable != 0;
464 * Use sequence counter to get consistent i_size on 32-bit processors.
466 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
467 #include <linux/seqlock.h>
468 #define __NEED_I_SIZE_ORDERED
469 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
471 #define i_size_ordered_init(inode) do { } while (0)
475 struct hlist_node i_hash;
476 struct list_head i_list;
477 struct list_head i_sb_list;
478 struct list_head i_dentry;
482 unsigned int i_nlink;
487 struct timespec i_atime;
488 struct timespec i_mtime;
489 struct timespec i_ctime;
490 unsigned int i_blkbits;
491 unsigned long i_blksize;
492 unsigned long i_version;
493 unsigned long i_blocks;
494 unsigned short i_bytes;
495 spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
496 struct mutex i_mutex;
497 struct rw_semaphore i_alloc_sem;
498 struct inode_operations *i_op;
499 struct file_operations *i_fop; /* former ->i_op->default_file_ops */
500 struct super_block *i_sb;
501 struct file_lock *i_flock;
502 struct address_space *i_mapping;
503 struct address_space i_data;
505 struct dquot *i_dquot[MAXQUOTAS];
507 /* These three should probably be a union */
508 struct list_head i_devices;
509 struct pipe_inode_info *i_pipe;
510 struct block_device *i_bdev;
516 #ifdef CONFIG_DNOTIFY
517 unsigned long i_dnotify_mask; /* Directory notify events */
518 struct dnotify_struct *i_dnotify; /* for directory notifications */
521 #ifdef CONFIG_INOTIFY
522 struct list_head inotify_watches; /* watches on this inode */
523 struct mutex inotify_mutex; /* protects the watches list */
526 unsigned long i_state;
527 unsigned long dirtied_when; /* jiffies of first dirtying */
529 unsigned int i_flags;
531 atomic_t i_writecount;
536 #ifdef __NEED_I_SIZE_ORDERED
537 seqcount_t i_size_seqcount;
542 * NOTE: in a 32bit arch with a preemptable kernel and
543 * an UP compile the i_size_read/write must be atomic
544 * with respect to the local cpu (unlike with preempt disabled),
545 * but they don't need to be atomic with respect to other cpus like in
546 * true SMP (so they need either to either locally disable irq around
547 * the read or for example on x86 they can be still implemented as a
548 * cmpxchg8b without the need of the lock prefix). For SMP compiles
549 * and 64bit archs it makes no difference if preempt is enabled or not.
551 static inline loff_t i_size_read(struct inode *inode)
553 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
558 seq = read_seqcount_begin(&inode->i_size_seqcount);
559 i_size = inode->i_size;
560 } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
562 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
566 i_size = inode->i_size;
570 return inode->i_size;
575 static inline void i_size_write(struct inode *inode, loff_t i_size)
577 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
578 write_seqcount_begin(&inode->i_size_seqcount);
579 inode->i_size = i_size;
580 write_seqcount_end(&inode->i_size_seqcount);
581 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
583 inode->i_size = i_size;
586 inode->i_size = i_size;
590 static inline unsigned iminor(struct inode *inode)
592 return MINOR(inode->i_rdev);
595 static inline unsigned imajor(struct inode *inode)
597 return MAJOR(inode->i_rdev);
600 extern struct block_device *I_BDEV(struct inode *inode);
603 rwlock_t lock; /* protects pid, uid, euid fields */
604 int pid; /* pid or -pgrp where SIGIO should be sent */
605 uid_t uid, euid; /* uid/euid of process setting the owner */
607 int signum; /* posix.1b rt signal to be delivered on IO */
611 * Track a single file's readahead state
613 struct file_ra_state {
614 unsigned long start; /* Current window */
616 unsigned long flags; /* ra flags RA_FLAG_xxx*/
617 unsigned long cache_hit; /* cache hit count*/
618 unsigned long prev_page; /* Cache last read() position */
619 unsigned long ahead_start; /* Ahead window */
620 unsigned long ahead_size;
621 unsigned long ra_pages; /* Maximum readahead window */
622 unsigned long mmap_hit; /* Cache hit stat for mmap accesses */
623 unsigned long mmap_miss; /* Cache miss stat for mmap accesses */
625 #define RA_FLAG_MISS 0x01 /* a cache miss occured against this file */
626 #define RA_FLAG_INCACHE 0x02 /* file is already in cache */
630 * fu_list becomes invalid after file_free is called and queued via
631 * fu_rcuhead for RCU freeing
634 struct list_head fu_list;
635 struct rcu_head fu_rcuhead;
637 struct dentry *f_dentry;
638 struct vfsmount *f_vfsmnt;
639 struct file_operations *f_op;
641 unsigned int f_flags;
644 struct fown_struct f_owner;
645 unsigned int f_uid, f_gid;
646 struct file_ra_state f_ra;
648 unsigned long f_version;
651 /* needed for tty driver, and maybe others */
655 /* Used by fs/eventpoll.c to link all the hooks to this file */
656 struct list_head f_ep_links;
657 spinlock_t f_ep_lock;
658 #endif /* #ifdef CONFIG_EPOLL */
659 struct address_space *f_mapping;
661 extern spinlock_t files_lock;
662 #define file_list_lock() spin_lock(&files_lock);
663 #define file_list_unlock() spin_unlock(&files_lock);
665 #define get_file(x) atomic_inc(&(x)->f_count)
666 #define file_count(x) atomic_read(&(x)->f_count)
668 #define MAX_NON_LFS ((1UL<<31) - 1)
670 /* Page cache limit. The filesystems should put that into their s_maxbytes
671 limits, otherwise bad things can happen in VM. */
672 #if BITS_PER_LONG==32
673 #define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
674 #elif BITS_PER_LONG==64
675 #define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
680 #define FL_ACCESS 8 /* not trying to lock, just looking */
681 #define FL_LEASE 32 /* lease held on this file */
682 #define FL_SLEEP 128 /* A blocking lock */
685 * The POSIX file lock owner is determined by
686 * the "struct files_struct" in the thread group
687 * (or NULL for no owner - BSD locks).
689 * Lockd stuffs a "host" pointer into this.
691 typedef struct files_struct *fl_owner_t;
693 struct file_lock_operations {
694 void (*fl_insert)(struct file_lock *); /* lock insertion callback */
695 void (*fl_remove)(struct file_lock *); /* lock removal callback */
696 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
697 void (*fl_release_private)(struct file_lock *);
700 struct lock_manager_operations {
701 int (*fl_compare_owner)(struct file_lock *, struct file_lock *);
702 void (*fl_notify)(struct file_lock *); /* unblock callback */
703 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
704 void (*fl_release_private)(struct file_lock *);
705 void (*fl_break)(struct file_lock *);
706 int (*fl_mylease)(struct file_lock *, struct file_lock *);
707 int (*fl_change)(struct file_lock **, int);
710 /* that will die - we need it for nfs_lock_info */
711 #include <linux/nfs_fs_i.h>
714 struct file_lock *fl_next; /* singly linked list for this inode */
715 struct list_head fl_link; /* doubly linked list of all locks */
716 struct list_head fl_block; /* circular list of blocked processes */
719 wait_queue_head_t fl_wait;
720 struct file *fl_file;
721 unsigned char fl_flags;
722 unsigned char fl_type;
726 struct fasync_struct * fl_fasync; /* for lease break notifications */
727 unsigned long fl_break_time; /* for nonblocking lease breaks */
729 struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
730 struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
732 struct nfs_lock_info nfs_fl;
733 struct nfs4_lock_info nfs4_fl;
737 /* The following constant reflects the upper bound of the file/locking space */
739 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
740 #define OFFSET_MAX INT_LIMIT(loff_t)
741 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
744 #include <linux/fcntl.h>
746 extern int fcntl_getlk(struct file *, struct flock __user *);
747 extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
748 struct flock __user *);
750 #if BITS_PER_LONG == 32
751 extern int fcntl_getlk64(struct file *, struct flock64 __user *);
752 extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
753 struct flock64 __user *);
756 extern void send_sigio(struct fown_struct *fown, int fd, int band);
757 extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
758 extern int fcntl_getlease(struct file *filp);
761 extern void locks_init_lock(struct file_lock *);
762 extern void locks_copy_lock(struct file_lock *, struct file_lock *);
763 extern void locks_remove_posix(struct file *, fl_owner_t);
764 extern void locks_remove_flock(struct file *);
765 extern int posix_test_lock(struct file *, struct file_lock *, struct file_lock *);
766 extern int posix_lock_file(struct file *, struct file_lock *);
767 extern int posix_lock_file_wait(struct file *, struct file_lock *);
768 extern int posix_unblock_lock(struct file *, struct file_lock *);
769 extern int posix_locks_deadlock(struct file_lock *, struct file_lock *);
770 extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
771 extern int __break_lease(struct inode *inode, unsigned int flags);
772 extern void lease_get_mtime(struct inode *, struct timespec *time);
773 extern int setlease(struct file *, long, struct file_lock **);
774 extern int lease_modify(struct file_lock **, int);
775 extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
776 extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
777 extern void steal_locks(fl_owner_t from);
779 struct fasync_struct {
782 struct fasync_struct *fa_next; /* singly linked list */
783 struct file *fa_file;
786 #define FASYNC_MAGIC 0x4601
788 /* SMP safe fasync helpers: */
789 extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
790 /* can be called from interrupts */
791 extern void kill_fasync(struct fasync_struct **, int, int);
792 /* only for net: no internal synchronization */
793 extern void __kill_fasync(struct fasync_struct *, int, int);
795 extern int f_setown(struct file *filp, unsigned long arg, int force);
796 extern void f_delown(struct file *filp);
797 extern int send_sigurg(struct fown_struct *fown);
803 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
804 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
805 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
807 extern struct list_head super_blocks;
808 extern spinlock_t sb_lock;
810 #define sb_entry(list) list_entry((list), struct super_block, s_list)
811 #define S_BIAS (1<<30)
813 struct list_head s_list; /* Keep this first */
814 dev_t s_dev; /* search index; _not_ kdev_t */
815 unsigned long s_blocksize;
816 unsigned char s_blocksize_bits;
817 unsigned char s_dirt;
818 unsigned long long s_maxbytes; /* Max file size */
819 struct file_system_type *s_type;
820 struct super_operations *s_op;
821 struct dquot_operations *dq_op;
822 struct quotactl_ops *s_qcop;
823 struct export_operations *s_export_op;
824 unsigned long s_flags;
825 unsigned long s_magic;
826 struct dentry *s_root;
827 struct rw_semaphore s_umount;
834 struct xattr_handler **s_xattr;
836 struct list_head s_inodes; /* all inodes */
837 struct list_head s_dirty; /* dirty inodes */
838 struct list_head s_io; /* parked for writeback */
839 struct hlist_head s_anon; /* anonymous dentries for (nfs) exporting */
840 struct list_head s_files;
842 struct block_device *s_bdev;
843 struct list_head s_instances;
844 struct quota_info s_dquot; /* Diskquota specific options */
847 wait_queue_head_t s_wait_unfrozen;
849 char s_id[32]; /* Informational name */
851 void *s_fs_info; /* Filesystem private info */
854 * The next field is for VFS *only*. No filesystems have any business
855 * even looking at it. You had been warned.
857 struct mutex s_vfs_rename_mutex; /* Kludge */
859 /* Granuality of c/m/atime in ns.
860 Cannot be worse than a second */
864 extern struct timespec current_fs_time(struct super_block *sb);
867 * Snapshotting support.
875 #define vfs_check_frozen(sb, level) \
876 wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
878 static inline void get_fs_excl(void)
880 atomic_inc(¤t->fs_excl);
883 static inline void put_fs_excl(void)
885 atomic_dec(¤t->fs_excl);
888 static inline int has_fs_excl(void)
890 return atomic_read(¤t->fs_excl);
895 * Superblock locking.
897 static inline void lock_super(struct super_block * sb)
900 mutex_lock(&sb->s_lock);
903 static inline void unlock_super(struct super_block * sb)
906 mutex_unlock(&sb->s_lock);
910 * VFS helper functions..
912 extern int vfs_permission(struct nameidata *, int);
913 extern int vfs_create(struct inode *, struct dentry *, int, struct nameidata *);
914 extern int vfs_mkdir(struct inode *, struct dentry *, int);
915 extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t);
916 extern int vfs_symlink(struct inode *, struct dentry *, const char *, int);
917 extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
918 extern int vfs_rmdir(struct inode *, struct dentry *);
919 extern int vfs_unlink(struct inode *, struct dentry *);
920 extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
923 * VFS dentry helper functions.
925 extern void dentry_unhash(struct dentry *dentry);
928 * VFS file helper functions.
930 extern int file_permission(struct file *, int);
935 * NOTE! These match bits 12..15 of stat.st_mode
936 * (ie "(i_mode >> 12) & 15").
948 #define OSYNC_METADATA (1<<0)
949 #define OSYNC_DATA (1<<1)
950 #define OSYNC_INODE (1<<2)
951 int generic_osync_inode(struct inode *, struct address_space *, int);
954 * This is the "filldir" function type, used by readdir() to let
955 * the kernel specify what kind of dirent layout it wants to have.
956 * This allows the kernel to read directories into kernel space or
957 * to have different dirent layouts depending on the binary type.
959 typedef int (*filldir_t)(void *, const char *, int, loff_t, ino_t, unsigned);
961 struct block_device_operations {
962 int (*open) (struct inode *, struct file *);
963 int (*release) (struct inode *, struct file *);
964 int (*ioctl) (struct inode *, struct file *, unsigned, unsigned long);
965 long (*unlocked_ioctl) (struct file *, unsigned, unsigned long);
966 long (*compat_ioctl) (struct file *, unsigned, unsigned long);
967 int (*direct_access) (struct block_device *, sector_t, unsigned long *);
968 int (*media_changed) (struct gendisk *);
969 int (*revalidate_disk) (struct gendisk *);
970 int (*getgeo)(struct block_device *, struct hd_geometry *);
971 struct module *owner;
975 * "descriptor" for what we're up to with a read for sendfile().
976 * This allows us to use the same read code yet
977 * have multiple different users of the data that
978 * we read from a file.
980 * The simplest case just copies the data to user
993 typedef int (*read_actor_t)(read_descriptor_t *, struct page *, unsigned long, unsigned long);
995 /* These macros are for out of kernel modules to test that
996 * the kernel supports the unlocked_ioctl and compat_ioctl
997 * fields in struct file_operations. */
998 #define HAVE_COMPAT_IOCTL 1
999 #define HAVE_UNLOCKED_IOCTL 1
1003 * read, write, poll, fsync, readv, writev, unlocked_ioctl and compat_ioctl
1004 * can be called without the big kernel lock held in all filesystems.
1006 struct file_operations {
1007 struct module *owner;
1008 loff_t (*llseek) (struct file *, loff_t, int);
1009 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1010 ssize_t (*aio_read) (struct kiocb *, char __user *, size_t, loff_t);
1011 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
1012 ssize_t (*aio_write) (struct kiocb *, const char __user *, size_t, loff_t);
1013 int (*readdir) (struct file *, void *, filldir_t);
1014 unsigned int (*poll) (struct file *, struct poll_table_struct *);
1015 int (*ioctl) (struct inode *, struct file *, unsigned int, unsigned long);
1016 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1017 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1018 int (*mmap) (struct file *, struct vm_area_struct *);
1019 int (*open) (struct inode *, struct file *);
1020 int (*flush) (struct file *);
1021 int (*release) (struct inode *, struct file *);
1022 int (*fsync) (struct file *, struct dentry *, int datasync);
1023 int (*aio_fsync) (struct kiocb *, int datasync);
1024 int (*fasync) (int, struct file *, int);
1025 int (*lock) (struct file *, int, struct file_lock *);
1026 ssize_t (*readv) (struct file *, const struct iovec *, unsigned long, loff_t *);
1027 ssize_t (*writev) (struct file *, const struct iovec *, unsigned long, loff_t *);
1028 ssize_t (*sendfile) (struct file *, loff_t *, size_t, read_actor_t, void *);
1029 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1030 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1031 int (*check_flags)(int);
1032 int (*dir_notify)(struct file *filp, unsigned long arg);
1033 int (*flock) (struct file *, int, struct file_lock *);
1036 struct inode_operations {
1037 int (*create) (struct inode *,struct dentry *,int, struct nameidata *);
1038 struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
1039 int (*link) (struct dentry *,struct inode *,struct dentry *);
1040 int (*unlink) (struct inode *,struct dentry *);
1041 int (*symlink) (struct inode *,struct dentry *,const char *);
1042 int (*mkdir) (struct inode *,struct dentry *,int);
1043 int (*rmdir) (struct inode *,struct dentry *);
1044 int (*mknod) (struct inode *,struct dentry *,int,dev_t);
1045 int (*rename) (struct inode *, struct dentry *,
1046 struct inode *, struct dentry *);
1047 int (*readlink) (struct dentry *, char __user *,int);
1048 void * (*follow_link) (struct dentry *, struct nameidata *);
1049 void (*put_link) (struct dentry *, struct nameidata *, void *);
1050 void (*truncate) (struct inode *);
1051 int (*permission) (struct inode *, int, struct nameidata *);
1052 int (*setattr) (struct dentry *, struct iattr *);
1053 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1054 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1055 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1056 ssize_t (*listxattr) (struct dentry *, char *, size_t);
1057 int (*removexattr) (struct dentry *, const char *);
1058 void (*truncate_range)(struct inode *, loff_t, loff_t);
1063 extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1064 extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1065 extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1066 unsigned long, loff_t *);
1067 extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1068 unsigned long, loff_t *);
1071 * NOTE: write_inode, delete_inode, clear_inode, put_inode can be called
1072 * without the big kernel lock held in all filesystems.
1074 struct super_operations {
1075 struct inode *(*alloc_inode)(struct super_block *sb);
1076 void (*destroy_inode)(struct inode *);
1078 void (*read_inode) (struct inode *);
1080 void (*dirty_inode) (struct inode *);
1081 int (*write_inode) (struct inode *, int);
1082 void (*put_inode) (struct inode *);
1083 void (*drop_inode) (struct inode *);
1084 void (*delete_inode) (struct inode *);
1085 void (*put_super) (struct super_block *);
1086 void (*write_super) (struct super_block *);
1087 int (*sync_fs)(struct super_block *sb, int wait);
1088 void (*write_super_lockfs) (struct super_block *);
1089 void (*unlockfs) (struct super_block *);
1090 int (*statfs) (struct super_block *, struct kstatfs *);
1091 int (*remount_fs) (struct super_block *, int *, char *);
1092 void (*clear_inode) (struct inode *);
1093 void (*umount_begin) (struct super_block *);
1095 int (*show_options)(struct seq_file *, struct vfsmount *);
1096 int (*show_stats)(struct seq_file *, struct vfsmount *);
1098 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1099 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
1102 /* Inode state bits. Protected by inode_lock. */
1103 #define I_DIRTY_SYNC 1 /* Not dirty enough for O_DATASYNC */
1104 #define I_DIRTY_DATASYNC 2 /* Data-related inode changes pending */
1105 #define I_DIRTY_PAGES 4 /* Data-related inode changes pending */
1107 #define I_LOCK (1 << __I_LOCK)
1108 #define I_FREEING 16
1111 #define I_WILL_FREE 128
1113 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1115 extern void __mark_inode_dirty(struct inode *, int);
1116 static inline void mark_inode_dirty(struct inode *inode)
1118 __mark_inode_dirty(inode, I_DIRTY);
1121 static inline void mark_inode_dirty_sync(struct inode *inode)
1123 __mark_inode_dirty(inode, I_DIRTY_SYNC);
1126 static inline void inode_inc_link_count(struct inode *inode)
1129 mark_inode_dirty(inode);
1132 static inline void inode_dec_link_count(struct inode *inode)
1135 mark_inode_dirty(inode);
1138 extern void touch_atime(struct vfsmount *mnt, struct dentry *dentry);
1139 static inline void file_accessed(struct file *file)
1141 if (!(file->f_flags & O_NOATIME))
1142 touch_atime(file->f_vfsmnt, file->f_dentry);
1145 int sync_inode(struct inode *inode, struct writeback_control *wbc);
1148 * struct export_operations - for nfsd to communicate with file systems
1149 * @decode_fh: decode a file handle fragment and return a &struct dentry
1150 * @encode_fh: encode a file handle fragment from a dentry
1151 * @get_name: find the name for a given inode in a given directory
1152 * @get_parent: find the parent of a given directory
1153 * @get_dentry: find a dentry for the inode given a file handle sub-fragment
1154 * @find_exported_dentry:
1155 * set by the exporting module to a standard helper function.
1158 * The export_operations structure provides a means for nfsd to communicate
1159 * with a particular exported file system - particularly enabling nfsd and
1160 * the filesystem to co-operate when dealing with file handles.
1162 * export_operations contains two basic operation for dealing with file
1163 * handles, decode_fh() and encode_fh(), and allows for some other
1164 * operations to be defined which standard helper routines use to get
1165 * specific information from the filesystem.
1167 * nfsd encodes information use to determine which filesystem a filehandle
1168 * applies to in the initial part of the file handle. The remainder, termed
1169 * a file handle fragment, is controlled completely by the filesystem. The
1170 * standard helper routines assume that this fragment will contain one or
1171 * two sub-fragments, one which identifies the file, and one which may be
1172 * used to identify the (a) directory containing the file.
1174 * In some situations, nfsd needs to get a dentry which is connected into a
1175 * specific part of the file tree. To allow for this, it passes the
1176 * function acceptable() together with a @context which can be used to see
1177 * if the dentry is acceptable. As there can be multiple dentrys for a
1178 * given file, the filesystem should check each one for acceptability before
1179 * looking for the next. As soon as an acceptable one is found, it should
1183 * @decode_fh is given a &struct super_block (@sb), a file handle fragment
1184 * (@fh, @fh_len) and an acceptability testing function (@acceptable,
1185 * @context). It should return a &struct dentry which refers to the same
1186 * file that the file handle fragment refers to, and which passes the
1187 * acceptability test. If it cannot, it should return a %NULL pointer if
1188 * the file was found but no acceptable &dentries were available, or a
1189 * %ERR_PTR error code indicating why it couldn't be found (e.g. %ENOENT or
1193 * @encode_fh should store in the file handle fragment @fh (using at most
1194 * @max_len bytes) information that can be used by @decode_fh to recover the
1195 * file refered to by the &struct dentry @de. If the @connectable flag is
1196 * set, the encode_fh() should store sufficient information so that a good
1197 * attempt can be made to find not only the file but also it's place in the
1198 * filesystem. This typically means storing a reference to de->d_parent in
1199 * the filehandle fragment. encode_fh() should return the number of bytes
1200 * stored or a negative error code such as %-ENOSPC
1203 * @get_name should find a name for the given @child in the given @parent
1204 * directory. The name should be stored in the @name (with the
1205 * understanding that it is already pointing to a a %NAME_MAX+1 sized
1206 * buffer. get_name() should return %0 on success, a negative error code
1207 * or error. @get_name will be called without @parent->i_mutex held.
1210 * @get_parent should find the parent directory for the given @child which
1211 * is also a directory. In the event that it cannot be found, or storage
1212 * space cannot be allocated, a %ERR_PTR should be returned.
1215 * Given a &super_block (@sb) and a pointer to a file-system specific inode
1216 * identifier, possibly an inode number, (@inump) get_dentry() should find
1217 * the identified inode and return a dentry for that inode. Any suitable
1218 * dentry can be returned including, if necessary, a new dentry created with
1219 * d_alloc_root. The caller can then find any other extant dentrys by
1220 * following the d_alias links. If a new dentry was created using
1221 * d_alloc_root, DCACHE_NFSD_DISCONNECTED should be set, and the dentry
1222 * should be d_rehash()ed.
1224 * If the inode cannot be found, either a %NULL pointer or an %ERR_PTR code
1225 * can be returned. The @inump will be whatever was passed to
1226 * nfsd_find_fh_dentry() in either the @obj or @parent parameters.
1229 * get_parent is called with child->d_inode->i_mutex down
1230 * get_name is not (which is possibly inconsistent)
1233 struct export_operations {
1234 struct dentry *(*decode_fh)(struct super_block *sb, __u32 *fh, int fh_len, int fh_type,
1235 int (*acceptable)(void *context, struct dentry *de),
1237 int (*encode_fh)(struct dentry *de, __u32 *fh, int *max_len,
1240 /* the following are only called from the filesystem itself */
1241 int (*get_name)(struct dentry *parent, char *name,
1242 struct dentry *child);
1243 struct dentry * (*get_parent)(struct dentry *child);
1244 struct dentry * (*get_dentry)(struct super_block *sb, void *inump);
1246 /* This is set by the exporting module to a standard helper */
1247 struct dentry * (*find_exported_dentry)(
1248 struct super_block *sb, void *obj, void *parent,
1249 int (*acceptable)(void *context, struct dentry *de),
1255 extern struct dentry *
1256 find_exported_dentry(struct super_block *sb, void *obj, void *parent,
1257 int (*acceptable)(void *context, struct dentry *de),
1260 struct file_system_type {
1263 struct super_block *(*get_sb) (struct file_system_type *, int,
1264 const char *, void *);
1265 void (*kill_sb) (struct super_block *);
1266 struct module *owner;
1267 struct file_system_type * next;
1268 struct list_head fs_supers;
1271 struct super_block *get_sb_bdev(struct file_system_type *fs_type,
1272 int flags, const char *dev_name, void *data,
1273 int (*fill_super)(struct super_block *, void *, int));
1274 struct super_block *get_sb_single(struct file_system_type *fs_type,
1275 int flags, void *data,
1276 int (*fill_super)(struct super_block *, void *, int));
1277 struct super_block *get_sb_nodev(struct file_system_type *fs_type,
1278 int flags, void *data,
1279 int (*fill_super)(struct super_block *, void *, int));
1280 void generic_shutdown_super(struct super_block *sb);
1281 void kill_block_super(struct super_block *sb);
1282 void kill_anon_super(struct super_block *sb);
1283 void kill_litter_super(struct super_block *sb);
1284 void deactivate_super(struct super_block *sb);
1285 int set_anon_super(struct super_block *s, void *data);
1286 struct super_block *sget(struct file_system_type *type,
1287 int (*test)(struct super_block *,void *),
1288 int (*set)(struct super_block *,void *),
1290 struct super_block *get_sb_pseudo(struct file_system_type *, char *,
1291 struct super_operations *ops, unsigned long);
1292 int __put_super(struct super_block *sb);
1293 int __put_super_and_need_restart(struct super_block *sb);
1294 void unnamed_dev_init(void);
1296 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1297 #define fops_get(fops) \
1298 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1299 #define fops_put(fops) \
1300 do { if (fops) module_put((fops)->owner); } while(0)
1302 extern int register_filesystem(struct file_system_type *);
1303 extern int unregister_filesystem(struct file_system_type *);
1304 extern struct vfsmount *kern_mount(struct file_system_type *);
1305 extern int may_umount_tree(struct vfsmount *);
1306 extern int may_umount(struct vfsmount *);
1307 extern void umount_tree(struct vfsmount *, int, struct list_head *);
1308 extern void release_mounts(struct list_head *);
1309 extern long do_mount(char *, char *, char *, unsigned long, void *);
1310 extern struct vfsmount *copy_tree(struct vfsmount *, struct dentry *, int);
1311 extern void mnt_set_mountpoint(struct vfsmount *, struct dentry *,
1314 extern int vfs_statfs(struct super_block *, struct kstatfs *);
1317 extern struct subsystem fs_subsys;
1319 #define FLOCK_VERIFY_READ 1
1320 #define FLOCK_VERIFY_WRITE 2
1322 extern int locks_mandatory_locked(struct inode *);
1323 extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
1326 * Candidates for mandatory locking have the setgid bit set
1327 * but no group execute bit - an otherwise meaningless combination.
1329 #define MANDATORY_LOCK(inode) \
1330 (IS_MANDLOCK(inode) && ((inode)->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID)
1332 static inline int locks_verify_locked(struct inode *inode)
1334 if (MANDATORY_LOCK(inode))
1335 return locks_mandatory_locked(inode);
1339 extern int rw_verify_area(int, struct file *, loff_t *, size_t);
1341 static inline int locks_verify_truncate(struct inode *inode,
1345 if (inode->i_flock && MANDATORY_LOCK(inode))
1346 return locks_mandatory_area(
1347 FLOCK_VERIFY_WRITE, inode, filp,
1348 size < inode->i_size ? size : inode->i_size,
1349 (size < inode->i_size ? inode->i_size - size
1350 : size - inode->i_size)
1355 static inline int break_lease(struct inode *inode, unsigned int mode)
1358 return __break_lease(inode, mode);
1364 extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
1366 extern long do_sys_open(int fdf, const char __user *filename, int flags,
1368 extern struct file *filp_open(const char *, int, int);
1369 extern struct file * dentry_open(struct dentry *, struct vfsmount *, int);
1370 extern int filp_close(struct file *, fl_owner_t id);
1371 extern char * getname(const char __user *);
1374 extern void __init vfs_caches_init_early(void);
1375 extern void __init vfs_caches_init(unsigned long);
1377 #define __getname() kmem_cache_alloc(names_cachep, SLAB_KERNEL)
1378 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
1379 #ifndef CONFIG_AUDITSYSCALL
1380 #define putname(name) __putname(name)
1382 extern void putname(const char *name);
1385 extern int register_blkdev(unsigned int, const char *);
1386 extern int unregister_blkdev(unsigned int, const char *);
1387 extern struct block_device *bdget(dev_t);
1388 extern void bd_set_size(struct block_device *, loff_t size);
1389 extern void bd_forget(struct inode *inode);
1390 extern void bdput(struct block_device *);
1391 extern struct block_device *open_by_devnum(dev_t, unsigned);
1392 extern struct file_operations def_blk_fops;
1393 extern struct address_space_operations def_blk_aops;
1394 extern struct file_operations def_chr_fops;
1395 extern struct file_operations bad_sock_fops;
1396 extern struct file_operations def_fifo_fops;
1397 extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
1398 extern int blkdev_ioctl(struct inode *, struct file *, unsigned, unsigned long);
1399 extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
1400 extern int blkdev_get(struct block_device *, mode_t, unsigned);
1401 extern int blkdev_put(struct block_device *);
1402 extern int bd_claim(struct block_device *, void *);
1403 extern void bd_release(struct block_device *);
1406 extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
1407 extern int register_chrdev_region(dev_t, unsigned, const char *);
1408 extern int register_chrdev(unsigned int, const char *,
1409 struct file_operations *);
1410 extern int unregister_chrdev(unsigned int, const char *);
1411 extern void unregister_chrdev_region(dev_t, unsigned);
1412 extern int chrdev_open(struct inode *, struct file *);
1413 extern int get_chrdev_list(char *);
1414 extern void *acquire_chrdev_list(void);
1415 extern int count_chrdev_list(void);
1416 extern void *get_next_chrdev(void *);
1417 extern int get_chrdev_info(void *, int *, char **);
1418 extern void release_chrdev_list(void *);
1420 /* fs/block_dev.c */
1421 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
1422 extern const char *__bdevname(dev_t, char *buffer);
1423 extern const char *bdevname(struct block_device *bdev, char *buffer);
1424 extern struct block_device *lookup_bdev(const char *);
1425 extern struct block_device *open_bdev_excl(const char *, int, void *);
1426 extern void close_bdev_excl(struct block_device *);
1427 extern void *acquire_blkdev_list(void);
1428 extern int count_blkdev_list(void);
1429 extern void *get_next_blkdev(void *);
1430 extern int get_blkdev_info(void *, int *, char **);
1431 extern void release_blkdev_list(void *);
1433 extern void init_special_inode(struct inode *, umode_t, dev_t);
1435 /* Invalid inode operations -- fs/bad_inode.c */
1436 extern void make_bad_inode(struct inode *);
1437 extern int is_bad_inode(struct inode *);
1439 extern struct file_operations read_fifo_fops;
1440 extern struct file_operations write_fifo_fops;
1441 extern struct file_operations rdwr_fifo_fops;
1443 extern int fs_may_remount_ro(struct super_block *);
1446 * return READ, READA, or WRITE
1448 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
1451 * return data direction, READ or WRITE
1453 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
1455 extern int check_disk_change(struct block_device *);
1456 extern int invalidate_inodes(struct super_block *);
1457 extern int __invalidate_device(struct block_device *);
1458 extern int invalidate_partition(struct gendisk *, int);
1459 unsigned long invalidate_mapping_pages(struct address_space *mapping,
1460 pgoff_t start, pgoff_t end);
1461 unsigned long invalidate_inode_pages(struct address_space *mapping);
1462 static inline void invalidate_remote_inode(struct inode *inode)
1464 if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
1465 S_ISLNK(inode->i_mode))
1466 invalidate_inode_pages(inode->i_mapping);
1468 extern int invalidate_inode_pages2(struct address_space *mapping);
1469 extern int invalidate_inode_pages2_range(struct address_space *mapping,
1470 pgoff_t start, pgoff_t end);
1471 extern int write_inode_now(struct inode *, int);
1472 extern int filemap_fdatawrite(struct address_space *);
1473 extern int filemap_flush(struct address_space *);
1474 extern int filemap_fdatawait(struct address_space *);
1475 extern int filemap_write_and_wait(struct address_space *mapping);
1476 extern int filemap_write_and_wait_range(struct address_space *mapping,
1477 loff_t lstart, loff_t lend);
1478 extern int wait_on_page_writeback_range(struct address_space *mapping,
1479 pgoff_t start, pgoff_t end);
1480 extern int __filemap_fdatawrite_range(struct address_space *mapping,
1481 loff_t start, loff_t end, int sync_mode);
1483 extern long do_fsync(struct file *file, int datasync);
1484 extern void sync_supers(void);
1485 extern void sync_filesystems(int wait);
1486 extern void emergency_sync(void);
1487 extern void emergency_remount(void);
1488 extern int do_remount_sb(struct super_block *sb, int flags,
1489 void *data, int force);
1490 extern sector_t bmap(struct inode *, sector_t);
1491 extern int notify_change(struct dentry *, struct iattr *);
1492 extern int permission(struct inode *, int, struct nameidata *);
1493 extern int generic_permission(struct inode *, int,
1494 int (*check_acl)(struct inode *, int));
1496 extern int get_write_access(struct inode *);
1497 extern int deny_write_access(struct file *);
1498 static inline void put_write_access(struct inode * inode)
1500 atomic_dec(&inode->i_writecount);
1502 static inline void allow_write_access(struct file *file)
1505 atomic_inc(&file->f_dentry->d_inode->i_writecount);
1507 extern int do_pipe(int *);
1509 extern int open_namei(int dfd, const char *, int, int, struct nameidata *);
1510 extern int may_open(struct nameidata *, int, int);
1512 extern int kernel_read(struct file *, unsigned long, char *, unsigned long);
1513 extern struct file * open_exec(const char *);
1515 /* fs/dcache.c -- generic fs support functions */
1516 extern int is_subdir(struct dentry *, struct dentry *);
1517 extern ino_t find_inode_number(struct dentry *, struct qstr *);
1519 #include <linux/err.h>
1521 /* needed for stackable file system support */
1522 extern loff_t default_llseek(struct file *file, loff_t offset, int origin);
1524 extern loff_t vfs_llseek(struct file *file, loff_t offset, int origin);
1526 extern void inode_init_once(struct inode *);
1527 extern void iput(struct inode *);
1528 extern struct inode * igrab(struct inode *);
1529 extern ino_t iunique(struct super_block *, ino_t);
1530 extern int inode_needs_sync(struct inode *inode);
1531 extern void generic_delete_inode(struct inode *inode);
1532 extern void generic_drop_inode(struct inode *inode);
1534 extern struct inode *ilookup5_nowait(struct super_block *sb,
1535 unsigned long hashval, int (*test)(struct inode *, void *),
1537 extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
1538 int (*test)(struct inode *, void *), void *data);
1539 extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
1541 extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
1542 extern struct inode * iget_locked(struct super_block *, unsigned long);
1543 extern void unlock_new_inode(struct inode *);
1545 static inline struct inode *iget(struct super_block *sb, unsigned long ino)
1547 struct inode *inode = iget_locked(sb, ino);
1549 if (inode && (inode->i_state & I_NEW)) {
1550 sb->s_op->read_inode(inode);
1551 unlock_new_inode(inode);
1557 extern void __iget(struct inode * inode);
1558 extern void clear_inode(struct inode *);
1559 extern void destroy_inode(struct inode *);
1560 extern struct inode *new_inode(struct super_block *);
1561 extern int remove_suid(struct dentry *);
1562 extern void remove_dquot_ref(struct super_block *, int, struct list_head *);
1564 extern void __insert_inode_hash(struct inode *, unsigned long hashval);
1565 extern void remove_inode_hash(struct inode *);
1566 static inline void insert_inode_hash(struct inode *inode) {
1567 __insert_inode_hash(inode, inode->i_ino);
1570 extern struct file * get_empty_filp(void);
1571 extern void file_move(struct file *f, struct list_head *list);
1572 extern void file_kill(struct file *f);
1574 extern void submit_bio(int, struct bio *);
1575 extern int bdev_read_only(struct block_device *);
1576 extern int set_blocksize(struct block_device *, int);
1577 extern int sb_set_blocksize(struct super_block *, int);
1578 extern int sb_min_blocksize(struct super_block *, int);
1580 extern int generic_file_mmap(struct file *, struct vm_area_struct *);
1581 extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
1582 extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
1583 extern int file_send_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
1584 extern ssize_t generic_file_read(struct file *, char __user *, size_t, loff_t *);
1585 int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
1586 extern ssize_t generic_file_write(struct file *, const char __user *, size_t, loff_t *);
1587 extern ssize_t generic_file_aio_read(struct kiocb *, char __user *, size_t, loff_t);
1588 extern ssize_t __generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t *);
1589 extern ssize_t generic_file_aio_write(struct kiocb *, const char __user *, size_t, loff_t);
1590 extern ssize_t generic_file_aio_write_nolock(struct kiocb *, const struct iovec *,
1591 unsigned long, loff_t *);
1592 extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
1593 unsigned long *, loff_t, loff_t *, size_t, size_t);
1594 extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
1595 unsigned long, loff_t, loff_t *, size_t, ssize_t);
1596 extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
1597 extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
1598 ssize_t generic_file_write_nolock(struct file *file, const struct iovec *iov,
1599 unsigned long nr_segs, loff_t *ppos);
1600 extern ssize_t generic_file_sendfile(struct file *, loff_t *, size_t, read_actor_t, void *);
1601 extern void do_generic_mapping_read(struct address_space *mapping,
1602 struct file_ra_state *, struct file *,
1603 loff_t *, read_descriptor_t *, read_actor_t);
1605 file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
1606 extern ssize_t generic_file_readv(struct file *filp, const struct iovec *iov,
1607 unsigned long nr_segs, loff_t *ppos);
1608 ssize_t generic_file_writev(struct file *filp, const struct iovec *iov,
1609 unsigned long nr_segs, loff_t *ppos);
1610 extern loff_t no_llseek(struct file *file, loff_t offset, int origin);
1611 extern loff_t generic_file_llseek(struct file *file, loff_t offset, int origin);
1612 extern loff_t remote_llseek(struct file *file, loff_t offset, int origin);
1613 extern int generic_file_open(struct inode * inode, struct file * filp);
1614 extern int nonseekable_open(struct inode * inode, struct file * filp);
1616 #ifdef CONFIG_FS_XIP
1617 extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
1619 extern ssize_t xip_file_sendfile(struct file *in_file, loff_t *ppos,
1620 size_t count, read_actor_t actor,
1622 extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
1623 extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
1624 size_t len, loff_t *ppos);
1625 extern int xip_truncate_page(struct address_space *mapping, loff_t from);
1627 static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
1633 static inline void do_generic_file_read(struct file * filp, loff_t *ppos,
1634 read_descriptor_t * desc,
1637 do_generic_mapping_read(filp->f_mapping,
1645 ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
1646 struct block_device *bdev, const struct iovec *iov, loff_t offset,
1647 unsigned long nr_segs, get_blocks_t get_blocks, dio_iodone_t end_io,
1651 DIO_LOCKING = 1, /* need locking between buffered and direct access */
1652 DIO_NO_LOCKING, /* bdev; no locking at all between buffered/direct */
1653 DIO_OWN_LOCKING, /* filesystem locks buffered and direct internally */
1656 static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
1657 struct inode *inode, struct block_device *bdev, const struct iovec *iov,
1658 loff_t offset, unsigned long nr_segs, get_blocks_t get_blocks,
1659 dio_iodone_t end_io)
1661 return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset,
1662 nr_segs, get_blocks, end_io, DIO_LOCKING);
1665 static inline ssize_t blockdev_direct_IO_no_locking(int rw, struct kiocb *iocb,
1666 struct inode *inode, struct block_device *bdev, const struct iovec *iov,
1667 loff_t offset, unsigned long nr_segs, get_blocks_t get_blocks,
1668 dio_iodone_t end_io)
1670 return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset,
1671 nr_segs, get_blocks, end_io, DIO_NO_LOCKING);
1674 static inline ssize_t blockdev_direct_IO_own_locking(int rw, struct kiocb *iocb,
1675 struct inode *inode, struct block_device *bdev, const struct iovec *iov,
1676 loff_t offset, unsigned long nr_segs, get_blocks_t get_blocks,
1677 dio_iodone_t end_io)
1679 return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset,
1680 nr_segs, get_blocks, end_io, DIO_OWN_LOCKING);
1683 extern struct file_operations generic_ro_fops;
1685 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
1687 extern int vfs_readlink(struct dentry *, char __user *, int, const char *);
1688 extern int vfs_follow_link(struct nameidata *, const char *);
1689 extern int page_readlink(struct dentry *, char __user *, int);
1690 extern void *page_follow_link_light(struct dentry *, struct nameidata *);
1691 extern void page_put_link(struct dentry *, struct nameidata *, void *);
1692 extern int __page_symlink(struct inode *inode, const char *symname, int len,
1694 extern int page_symlink(struct inode *inode, const char *symname, int len);
1695 extern struct inode_operations page_symlink_inode_operations;
1696 extern int generic_readlink(struct dentry *, char __user *, int);
1697 extern void generic_fillattr(struct inode *, struct kstat *);
1698 extern int vfs_getattr(struct vfsmount *, struct dentry *, struct kstat *);
1699 void inode_add_bytes(struct inode *inode, loff_t bytes);
1700 void inode_sub_bytes(struct inode *inode, loff_t bytes);
1701 loff_t inode_get_bytes(struct inode *inode);
1702 void inode_set_bytes(struct inode *inode, loff_t bytes);
1704 extern int vfs_readdir(struct file *, filldir_t, void *);
1706 extern int vfs_stat(char __user *, struct kstat *);
1707 extern int vfs_lstat(char __user *, struct kstat *);
1708 extern int vfs_stat_fd(int dfd, char __user *, struct kstat *);
1709 extern int vfs_lstat_fd(int dfd, char __user *, struct kstat *);
1710 extern int vfs_fstat(unsigned int, struct kstat *);
1712 extern int vfs_ioctl(struct file *, unsigned int, unsigned int, unsigned long);
1714 extern struct file_system_type *get_fs_type(const char *name);
1715 extern struct super_block *get_super(struct block_device *);
1716 extern struct super_block *user_get_super(dev_t);
1717 extern void drop_super(struct super_block *sb);
1719 extern int dcache_dir_open(struct inode *, struct file *);
1720 extern int dcache_dir_close(struct inode *, struct file *);
1721 extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
1722 extern int dcache_readdir(struct file *, void *, filldir_t);
1723 extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
1724 extern int simple_statfs(struct super_block *, struct kstatfs *);
1725 extern int simple_link(struct dentry *, struct inode *, struct dentry *);
1726 extern int simple_unlink(struct inode *, struct dentry *);
1727 extern int simple_rmdir(struct inode *, struct dentry *);
1728 extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
1729 extern int simple_sync_file(struct file *, struct dentry *, int);
1730 extern int simple_empty(struct dentry *);
1731 extern int simple_readpage(struct file *file, struct page *page);
1732 extern int simple_prepare_write(struct file *file, struct page *page,
1733 unsigned offset, unsigned to);
1734 extern int simple_commit_write(struct file *file, struct page *page,
1735 unsigned offset, unsigned to);
1737 extern struct dentry *simple_lookup(struct inode *, struct dentry *, struct nameidata *);
1738 extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
1739 extern struct file_operations simple_dir_operations;
1740 extern struct inode_operations simple_dir_inode_operations;
1741 struct tree_descr { char *name; struct file_operations *ops; int mode; };
1742 struct dentry *d_alloc_name(struct dentry *, const char *);
1743 extern int simple_fill_super(struct super_block *, int, struct tree_descr *);
1744 extern int simple_pin_fs(char *name, struct vfsmount **mount, int *count);
1745 extern void simple_release_fs(struct vfsmount **mount, int *count);
1747 extern ssize_t simple_read_from_buffer(void __user *, size_t, loff_t *, const void *, size_t);
1749 #ifdef CONFIG_MIGRATION
1750 extern int buffer_migrate_page(struct page *, struct page *);
1752 #define buffer_migrate_page NULL
1755 extern int inode_change_ok(struct inode *, struct iattr *);
1756 extern int __must_check inode_setattr(struct inode *, struct iattr *);
1758 extern void file_update_time(struct file *file);
1760 static inline ino_t parent_ino(struct dentry *dentry)
1764 spin_lock(&dentry->d_lock);
1765 res = dentry->d_parent->d_inode->i_ino;
1766 spin_unlock(&dentry->d_lock);
1771 extern int unshare_files(void);
1773 /* Transaction based IO helpers */
1776 * An argresp is stored in an allocated page and holds the
1777 * size of the argument or response, along with its content
1779 struct simple_transaction_argresp {
1784 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
1786 char *simple_transaction_get(struct file *file, const char __user *buf,
1788 ssize_t simple_transaction_read(struct file *file, char __user *buf,
1789 size_t size, loff_t *pos);
1790 int simple_transaction_release(struct inode *inode, struct file *file);
1792 static inline void simple_transaction_set(struct file *file, size_t n)
1794 struct simple_transaction_argresp *ar = file->private_data;
1796 BUG_ON(n > SIMPLE_TRANSACTION_LIMIT);
1799 * The barrier ensures that ar->size will really remain zero until
1800 * ar->data is ready for reading.
1807 * simple attribute files
1809 * These attributes behave similar to those in sysfs:
1811 * Writing to an attribute immediately sets a value, an open file can be
1812 * written to multiple times.
1814 * Reading from an attribute creates a buffer from the value that might get
1815 * read with multiple read calls. When the attribute has been read
1816 * completely, no further read calls are possible until the file is opened
1819 * All attributes contain a text representation of a numeric value
1820 * that are accessed with the get() and set() functions.
1822 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
1823 static int __fops ## _open(struct inode *inode, struct file *file) \
1825 __simple_attr_check_format(__fmt, 0ull); \
1826 return simple_attr_open(inode, file, __get, __set, __fmt); \
1828 static struct file_operations __fops = { \
1829 .owner = THIS_MODULE, \
1830 .open = __fops ## _open, \
1831 .release = simple_attr_close, \
1832 .read = simple_attr_read, \
1833 .write = simple_attr_write, \
1836 static inline void __attribute__((format(printf, 1, 2)))
1837 __simple_attr_check_format(const char *fmt, ...)
1839 /* don't do anything, just let the compiler check the arguments; */
1842 int simple_attr_open(struct inode *inode, struct file *file,
1843 u64 (*get)(void *), void (*set)(void *, u64),
1845 int simple_attr_close(struct inode *inode, struct file *file);
1846 ssize_t simple_attr_read(struct file *file, char __user *buf,
1847 size_t len, loff_t *ppos);
1848 ssize_t simple_attr_write(struct file *file, const char __user *buf,
1849 size_t len, loff_t *ppos);
1852 #ifdef CONFIG_SECURITY
1853 static inline char *alloc_secdata(void)
1855 return (char *)get_zeroed_page(GFP_KERNEL);
1858 static inline void free_secdata(void *secdata)
1860 free_page((unsigned long)secdata);
1863 static inline char *alloc_secdata(void)
1868 static inline void free_secdata(void *secdata)
1870 #endif /* CONFIG_SECURITY */
1872 #endif /* __KERNEL__ */
1873 #endif /* _LINUX_FS_H */