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 */
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 /* War is peace. Verbosity is silence.
107 MS_VERBOSE is deprecated. */
108 #define MS_SILENT 32768
109 #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
110 #define MS_UNBINDABLE (1<<17) /* change to unbindable */
111 #define MS_PRIVATE (1<<18) /* change to private */
112 #define MS_SLAVE (1<<19) /* change to slave */
113 #define MS_SHARED (1<<20) /* change to shared */
114 #define MS_ACTIVE (1<<30)
115 #define MS_NOUSER (1<<31)
118 * Superblock flags that can be altered by MS_REMOUNT
120 #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK)
123 * Old magic mount flag and mask
125 #define MS_MGC_VAL 0xC0ED0000
126 #define MS_MGC_MSK 0xffff0000
128 /* Inode flags - they have nothing to superblock flags now */
130 #define S_SYNC 1 /* Writes are synced at once */
131 #define S_NOATIME 2 /* Do not update access times */
132 #define S_APPEND 4 /* Append-only file */
133 #define S_IMMUTABLE 8 /* Immutable file */
134 #define S_DEAD 16 /* removed, but still open directory */
135 #define S_NOQUOTA 32 /* Inode is not counted to quota */
136 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
137 #define S_NOCMTIME 128 /* Do not update file c/mtime */
138 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
139 #define S_PRIVATE 512 /* Inode is fs-internal */
142 * Note that nosuid etc flags are inode-specific: setting some file-system
143 * flags just means all the inodes inherit those flags by default. It might be
144 * possible to override it selectively if you really wanted to with some
145 * ioctl() that is not currently implemented.
147 * Exception: MS_RDONLY is always applied to the entire file system.
149 * Unfortunately, it is possible to change a filesystems flags with it mounted
150 * with files in use. This means that all of the inodes will not have their
151 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
152 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
154 #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
156 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
157 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
158 ((inode)->i_flags & S_SYNC))
159 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
160 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
161 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
163 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
164 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
165 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
166 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
168 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
169 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
170 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
171 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
173 /* the read-only stuff doesn't really belong here, but any other place is
174 probably as bad and I don't want to create yet another include file. */
176 #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
177 #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
178 #define BLKRRPART _IO(0x12,95) /* re-read partition table */
179 #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
180 #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
181 #define BLKRASET _IO(0x12,98) /* set read ahead for block device */
182 #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
183 #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
184 #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
185 #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
186 #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
187 #define BLKSSZGET _IO(0x12,104)/* get block device sector size */
189 #define BLKPG _IO(0x12,105)/* See blkpg.h */
191 /* Some people are morons. Do not use sizeof! */
193 #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
194 #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
195 /* This was here just to show that the number is taken -
196 probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
198 /* A jump here: 108-111 have been used for various private purposes. */
199 #define BLKBSZGET _IOR(0x12,112,size_t)
200 #define BLKBSZSET _IOW(0x12,113,size_t)
201 #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
202 #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
203 #define BLKTRACESTART _IO(0x12,116)
204 #define BLKTRACESTOP _IO(0x12,117)
205 #define BLKTRACETEARDOWN _IO(0x12,118)
207 #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
208 #define FIBMAP _IO(0x00,1) /* bmap access */
209 #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
213 #include <linux/linkage.h>
214 #include <linux/wait.h>
215 #include <linux/types.h>
216 #include <linux/kdev_t.h>
217 #include <linux/dcache.h>
218 #include <linux/stat.h>
219 #include <linux/cache.h>
220 #include <linux/kobject.h>
221 #include <linux/list.h>
222 #include <linux/radix-tree.h>
223 #include <linux/prio_tree.h>
224 #include <linux/init.h>
225 #include <linux/sched.h>
226 #include <linux/mutex.h>
228 #include <asm/atomic.h>
229 #include <asm/semaphore.h>
230 #include <asm/byteorder.h>
236 struct pipe_inode_info;
237 struct poll_table_struct;
239 struct vm_area_struct;
242 extern void __init inode_init(unsigned long);
243 extern void __init inode_init_early(void);
244 extern void __init mnt_init(unsigned long);
245 extern void __init files_init(unsigned long);
248 typedef int (get_block_t)(struct inode *inode, sector_t iblock,
249 struct buffer_head *bh_result, int create);
250 typedef int (get_blocks_t)(struct inode *inode, sector_t iblock,
251 unsigned long max_blocks,
252 struct buffer_head *bh_result, int create);
253 typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
254 ssize_t bytes, void *private);
257 * Attribute flags. These should be or-ed together to figure out what
264 #define ATTR_ATIME 16
265 #define ATTR_MTIME 32
266 #define ATTR_CTIME 64
267 #define ATTR_ATIME_SET 128
268 #define ATTR_MTIME_SET 256
269 #define ATTR_FORCE 512 /* Not a change, but a change it */
270 #define ATTR_ATTR_FLAG 1024
271 #define ATTR_KILL_SUID 2048
272 #define ATTR_KILL_SGID 4096
273 #define ATTR_FILE 8192
276 * This is the Inode Attributes structure, used for notify_change(). It
277 * uses the above definitions as flags, to know which values have changed.
278 * Also, in this manner, a Filesystem can look at only the values it cares
279 * about. Basically, these are the attributes that the VFS layer can
280 * request to change from the FS layer.
282 * Derek Atkins <warlord@MIT.EDU> 94-10-20
285 unsigned int ia_valid;
290 struct timespec ia_atime;
291 struct timespec ia_mtime;
292 struct timespec ia_ctime;
295 * Not an attribute, but an auxilary info for filesystems wanting to
296 * implement an ftruncate() like method. NOTE: filesystem should
297 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
299 struct file *ia_file;
303 * Includes for diskquotas.
305 #include <linux/quota.h>
308 * enum positive_aop_returns - aop return codes with specific semantics
310 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
311 * completed, that the page is still locked, and
312 * should be considered active. The VM uses this hint
313 * to return the page to the active list -- it won't
314 * be a candidate for writeback again in the near
315 * future. Other callers must be careful to unlock
316 * the page if they get this return. Returned by
319 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
320 * unlocked it and the page might have been truncated.
321 * The caller should back up to acquiring a new page and
322 * trying again. The aop will be taking reasonable
323 * precautions not to livelock. If the caller held a page
324 * reference, it should drop it before retrying. Returned
325 * by readpage(), prepare_write(), and commit_write().
327 * address_space_operation functions return these large constants to indicate
328 * special semantics to the caller. These are much larger than the bytes in a
329 * page to allow for functions that return the number of bytes operated on in a
333 enum positive_aop_returns {
334 AOP_WRITEPAGE_ACTIVATE = 0x80000,
335 AOP_TRUNCATED_PAGE = 0x80001,
339 * oh the beauties of C type declarations.
342 struct address_space;
343 struct writeback_control;
345 struct address_space_operations {
346 int (*writepage)(struct page *page, struct writeback_control *wbc);
347 int (*readpage)(struct file *, struct page *);
348 int (*sync_page)(struct page *);
350 /* Write back some dirty pages from this mapping. */
351 int (*writepages)(struct address_space *, struct writeback_control *);
353 /* Set a page dirty. Return true if this dirtied it */
354 int (*set_page_dirty)(struct page *page);
356 int (*readpages)(struct file *filp, struct address_space *mapping,
357 struct list_head *pages, unsigned nr_pages);
360 * ext3 requires that a successful prepare_write() call be followed
361 * by a commit_write() call - they must be balanced
363 int (*prepare_write)(struct file *, struct page *, unsigned, unsigned);
364 int (*commit_write)(struct file *, struct page *, unsigned, unsigned);
365 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
366 sector_t (*bmap)(struct address_space *, sector_t);
367 int (*invalidatepage) (struct page *, unsigned long);
368 int (*releasepage) (struct page *, gfp_t);
369 ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
370 loff_t offset, unsigned long nr_segs);
371 struct page* (*get_xip_page)(struct address_space *, sector_t,
373 /* migrate the contents of a page to the specified target */
374 int (*migratepage) (struct page *, struct page *);
377 struct backing_dev_info;
378 struct address_space {
379 struct inode *host; /* owner: inode, block_device */
380 struct radix_tree_root page_tree; /* radix tree of all pages */
381 rwlock_t tree_lock; /* and rwlock protecting it */
382 unsigned int i_mmap_writable;/* count VM_SHARED mappings */
383 struct prio_tree_root i_mmap; /* tree of private and shared mappings */
384 struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
385 spinlock_t i_mmap_lock; /* protect tree, count, list */
386 unsigned int truncate_count; /* Cover race condition with truncate */
387 unsigned long nrpages; /* number of total pages */
388 pgoff_t writeback_index;/* writeback starts here */
389 struct address_space_operations *a_ops; /* methods */
390 unsigned long flags; /* error bits/gfp mask */
391 struct backing_dev_info *backing_dev_info; /* device readahead, etc */
392 spinlock_t private_lock; /* for use by the address_space */
393 struct list_head private_list; /* ditto */
394 struct address_space *assoc_mapping; /* ditto */
395 } __attribute__((aligned(sizeof(long))));
397 * On most architectures that alignment is already the case; but
398 * must be enforced here for CRIS, to let the least signficant bit
399 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
402 struct block_device {
403 dev_t bd_dev; /* not a kdev_t - it's a search key */
404 struct inode * bd_inode; /* will die */
406 struct mutex bd_mutex; /* open/close mutex */
407 struct mutex bd_mount_mutex; /* mount mutex */
408 struct list_head bd_inodes;
411 struct block_device * bd_contains;
412 unsigned bd_block_size;
413 struct hd_struct * bd_part;
414 /* number of times partitions within this device have been opened. */
415 unsigned bd_part_count;
417 struct gendisk * bd_disk;
418 struct list_head bd_list;
419 struct backing_dev_info *bd_inode_backing_dev_info;
421 * Private data. You must have bd_claim'ed the block_device
422 * to use this. NOTE: bd_claim allows an owner to claim
423 * the same device multiple times, the owner must take special
424 * care to not mess up bd_private for that case.
426 unsigned long bd_private;
430 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
433 #define PAGECACHE_TAG_DIRTY 0
434 #define PAGECACHE_TAG_WRITEBACK 1
436 int mapping_tagged(struct address_space *mapping, int tag);
439 * Might pages of this file be mapped into userspace?
441 static inline int mapping_mapped(struct address_space *mapping)
443 return !prio_tree_empty(&mapping->i_mmap) ||
444 !list_empty(&mapping->i_mmap_nonlinear);
448 * Might pages of this file have been modified in userspace?
449 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
450 * marks vma as VM_SHARED if it is shared, and the file was opened for
451 * writing i.e. vma may be mprotected writable even if now readonly.
453 static inline int mapping_writably_mapped(struct address_space *mapping)
455 return mapping->i_mmap_writable != 0;
459 * Use sequence counter to get consistent i_size on 32-bit processors.
461 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
462 #include <linux/seqlock.h>
463 #define __NEED_I_SIZE_ORDERED
464 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
466 #define i_size_ordered_init(inode) do { } while (0)
470 struct hlist_node i_hash;
471 struct list_head i_list;
472 struct list_head i_sb_list;
473 struct list_head i_dentry;
477 unsigned int i_nlink;
482 struct timespec i_atime;
483 struct timespec i_mtime;
484 struct timespec i_ctime;
485 unsigned int i_blkbits;
486 unsigned long i_blksize;
487 unsigned long i_version;
488 unsigned long i_blocks;
489 unsigned short i_bytes;
490 spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
491 struct mutex i_mutex;
492 struct rw_semaphore i_alloc_sem;
493 struct inode_operations *i_op;
494 struct file_operations *i_fop; /* former ->i_op->default_file_ops */
495 struct super_block *i_sb;
496 struct file_lock *i_flock;
497 struct address_space *i_mapping;
498 struct address_space i_data;
500 struct dquot *i_dquot[MAXQUOTAS];
502 /* These three should probably be a union */
503 struct list_head i_devices;
504 struct pipe_inode_info *i_pipe;
505 struct block_device *i_bdev;
511 #ifdef CONFIG_DNOTIFY
512 unsigned long i_dnotify_mask; /* Directory notify events */
513 struct dnotify_struct *i_dnotify; /* for directory notifications */
516 #ifdef CONFIG_INOTIFY
517 struct list_head inotify_watches; /* watches on this inode */
518 struct mutex inotify_mutex; /* protects the watches list */
521 unsigned long i_state;
522 unsigned long dirtied_when; /* jiffies of first dirtying */
524 unsigned int i_flags;
526 atomic_t i_writecount;
531 #ifdef __NEED_I_SIZE_ORDERED
532 seqcount_t i_size_seqcount;
537 * NOTE: in a 32bit arch with a preemptable kernel and
538 * an UP compile the i_size_read/write must be atomic
539 * with respect to the local cpu (unlike with preempt disabled),
540 * but they don't need to be atomic with respect to other cpus like in
541 * true SMP (so they need either to either locally disable irq around
542 * the read or for example on x86 they can be still implemented as a
543 * cmpxchg8b without the need of the lock prefix). For SMP compiles
544 * and 64bit archs it makes no difference if preempt is enabled or not.
546 static inline loff_t i_size_read(struct inode *inode)
548 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
553 seq = read_seqcount_begin(&inode->i_size_seqcount);
554 i_size = inode->i_size;
555 } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
557 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
561 i_size = inode->i_size;
565 return inode->i_size;
570 static inline void i_size_write(struct inode *inode, loff_t i_size)
572 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
573 write_seqcount_begin(&inode->i_size_seqcount);
574 inode->i_size = i_size;
575 write_seqcount_end(&inode->i_size_seqcount);
576 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
578 inode->i_size = i_size;
581 inode->i_size = i_size;
585 static inline unsigned iminor(struct inode *inode)
587 return MINOR(inode->i_rdev);
590 static inline unsigned imajor(struct inode *inode)
592 return MAJOR(inode->i_rdev);
595 extern struct block_device *I_BDEV(struct inode *inode);
598 rwlock_t lock; /* protects pid, uid, euid fields */
599 int pid; /* pid or -pgrp where SIGIO should be sent */
600 uid_t uid, euid; /* uid/euid of process setting the owner */
602 int signum; /* posix.1b rt signal to be delivered on IO */
606 * Track a single file's readahead state
608 struct file_ra_state {
609 unsigned long start; /* Current window */
611 unsigned long flags; /* ra flags RA_FLAG_xxx*/
612 unsigned long cache_hit; /* cache hit count*/
613 unsigned long prev_page; /* Cache last read() position */
614 unsigned long ahead_start; /* Ahead window */
615 unsigned long ahead_size;
616 unsigned long ra_pages; /* Maximum readahead window */
617 unsigned long mmap_hit; /* Cache hit stat for mmap accesses */
618 unsigned long mmap_miss; /* Cache miss stat for mmap accesses */
620 #define RA_FLAG_MISS 0x01 /* a cache miss occured against this file */
621 #define RA_FLAG_INCACHE 0x02 /* file is already in cache */
625 * fu_list becomes invalid after file_free is called and queued via
626 * fu_rcuhead for RCU freeing
629 struct list_head fu_list;
630 struct rcu_head fu_rcuhead;
632 struct dentry *f_dentry;
633 struct vfsmount *f_vfsmnt;
634 struct file_operations *f_op;
636 unsigned int f_flags;
639 struct fown_struct f_owner;
640 unsigned int f_uid, f_gid;
641 struct file_ra_state f_ra;
643 unsigned long f_version;
646 /* needed for tty driver, and maybe others */
650 /* Used by fs/eventpoll.c to link all the hooks to this file */
651 struct list_head f_ep_links;
652 spinlock_t f_ep_lock;
653 #endif /* #ifdef CONFIG_EPOLL */
654 struct address_space *f_mapping;
656 extern spinlock_t files_lock;
657 #define file_list_lock() spin_lock(&files_lock);
658 #define file_list_unlock() spin_unlock(&files_lock);
660 #define get_file(x) atomic_inc(&(x)->f_count)
661 #define file_count(x) atomic_read(&(x)->f_count)
663 #define MAX_NON_LFS ((1UL<<31) - 1)
665 /* Page cache limit. The filesystems should put that into their s_maxbytes
666 limits, otherwise bad things can happen in VM. */
667 #if BITS_PER_LONG==32
668 #define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
669 #elif BITS_PER_LONG==64
670 #define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
675 #define FL_ACCESS 8 /* not trying to lock, just looking */
676 #define FL_LOCKD 16 /* lock held by rpc.lockd */
677 #define FL_LEASE 32 /* lease held on this file */
678 #define FL_SLEEP 128 /* A blocking lock */
681 * The POSIX file lock owner is determined by
682 * the "struct files_struct" in the thread group
683 * (or NULL for no owner - BSD locks).
685 * Lockd stuffs a "host" pointer into this.
687 typedef struct files_struct *fl_owner_t;
689 struct file_lock_operations {
690 void (*fl_insert)(struct file_lock *); /* lock insertion callback */
691 void (*fl_remove)(struct file_lock *); /* lock removal callback */
692 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
693 void (*fl_release_private)(struct file_lock *);
696 struct lock_manager_operations {
697 int (*fl_compare_owner)(struct file_lock *, struct file_lock *);
698 void (*fl_notify)(struct file_lock *); /* unblock callback */
699 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
700 void (*fl_release_private)(struct file_lock *);
701 void (*fl_break)(struct file_lock *);
702 int (*fl_mylease)(struct file_lock *, struct file_lock *);
703 int (*fl_change)(struct file_lock **, int);
706 /* that will die - we need it for nfs_lock_info */
707 #include <linux/nfs_fs_i.h>
710 struct file_lock *fl_next; /* singly linked list for this inode */
711 struct list_head fl_link; /* doubly linked list of all locks */
712 struct list_head fl_block; /* circular list of blocked processes */
715 wait_queue_head_t fl_wait;
716 struct file *fl_file;
717 unsigned char fl_flags;
718 unsigned char fl_type;
722 struct fasync_struct * fl_fasync; /* for lease break notifications */
723 unsigned long fl_break_time; /* for nonblocking lease breaks */
725 struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
726 struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
728 struct nfs_lock_info nfs_fl;
729 struct nfs4_lock_info nfs4_fl;
733 /* The following constant reflects the upper bound of the file/locking space */
735 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
736 #define OFFSET_MAX INT_LIMIT(loff_t)
737 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
740 extern struct list_head file_lock_list;
742 #include <linux/fcntl.h>
744 extern int fcntl_getlk(struct file *, struct flock __user *);
745 extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
746 struct flock __user *);
748 #if BITS_PER_LONG == 32
749 extern int fcntl_getlk64(struct file *, struct flock64 __user *);
750 extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
751 struct flock64 __user *);
754 extern void send_sigio(struct fown_struct *fown, int fd, int band);
755 extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
756 extern int fcntl_getlease(struct file *filp);
759 extern void locks_init_lock(struct file_lock *);
760 extern void locks_copy_lock(struct file_lock *, struct file_lock *);
761 extern void locks_remove_posix(struct file *, fl_owner_t);
762 extern void locks_remove_flock(struct file *);
763 extern struct file_lock *posix_test_lock(struct file *, struct file_lock *);
764 extern int posix_lock_file(struct file *, struct file_lock *);
765 extern int posix_lock_file_wait(struct file *, struct file_lock *);
766 extern void posix_block_lock(struct file_lock *, struct file_lock *);
767 extern int posix_unblock_lock(struct file *, struct file_lock *);
768 extern int posix_locks_deadlock(struct file_lock *, struct file_lock *);
769 extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
770 extern int __break_lease(struct inode *inode, unsigned int flags);
771 extern void lease_get_mtime(struct inode *, struct timespec *time);
772 extern int setlease(struct file *, long, struct file_lock **);
773 extern int lease_modify(struct file_lock **, int);
774 extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
775 extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
776 extern void steal_locks(fl_owner_t from);
778 struct fasync_struct {
781 struct fasync_struct *fa_next; /* singly linked list */
782 struct file *fa_file;
785 #define FASYNC_MAGIC 0x4601
787 /* SMP safe fasync helpers: */
788 extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
789 /* can be called from interrupts */
790 extern void kill_fasync(struct fasync_struct **, int, int);
791 /* only for net: no internal synchronization */
792 extern void __kill_fasync(struct fasync_struct *, int, int);
794 extern int f_setown(struct file *filp, unsigned long arg, int force);
795 extern void f_delown(struct file *filp);
796 extern int send_sigurg(struct fown_struct *fown);
802 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
803 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
804 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
806 extern struct list_head super_blocks;
807 extern spinlock_t sb_lock;
809 #define sb_entry(list) list_entry((list), struct super_block, s_list)
810 #define S_BIAS (1<<30)
812 struct list_head s_list; /* Keep this first */
813 dev_t s_dev; /* search index; _not_ kdev_t */
814 unsigned long s_blocksize;
815 unsigned char s_blocksize_bits;
816 unsigned char s_dirt;
817 unsigned long long s_maxbytes; /* Max file size */
818 struct file_system_type *s_type;
819 struct super_operations *s_op;
820 struct dquot_operations *dq_op;
821 struct quotactl_ops *s_qcop;
822 struct export_operations *s_export_op;
823 unsigned long s_flags;
824 unsigned long s_magic;
825 struct dentry *s_root;
826 struct rw_semaphore s_umount;
833 struct xattr_handler **s_xattr;
835 struct list_head s_inodes; /* all inodes */
836 struct list_head s_dirty; /* dirty inodes */
837 struct list_head s_io; /* parked for writeback */
838 struct hlist_head s_anon; /* anonymous dentries for (nfs) exporting */
839 struct list_head s_files;
841 struct block_device *s_bdev;
842 struct list_head s_instances;
843 struct quota_info s_dquot; /* Diskquota specific options */
846 wait_queue_head_t s_wait_unfrozen;
848 char s_id[32]; /* Informational name */
850 void *s_fs_info; /* Filesystem private info */
853 * The next field is for VFS *only*. No filesystems have any business
854 * even looking at it. You had been warned.
856 struct mutex s_vfs_rename_mutex; /* Kludge */
858 /* Granuality of c/m/atime in ns.
859 Cannot be worse than a second */
863 extern struct timespec current_fs_time(struct super_block *sb);
866 * Snapshotting support.
874 #define vfs_check_frozen(sb, level) \
875 wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
877 static inline void get_fs_excl(void)
879 atomic_inc(¤t->fs_excl);
882 static inline void put_fs_excl(void)
884 atomic_dec(¤t->fs_excl);
887 static inline int has_fs_excl(void)
889 return atomic_read(¤t->fs_excl);
894 * Superblock locking.
896 static inline void lock_super(struct super_block * sb)
899 mutex_lock(&sb->s_lock);
902 static inline void unlock_super(struct super_block * sb)
905 mutex_unlock(&sb->s_lock);
909 * VFS helper functions..
911 extern int vfs_permission(struct nameidata *, int);
912 extern int vfs_create(struct inode *, struct dentry *, int, struct nameidata *);
913 extern int vfs_mkdir(struct inode *, struct dentry *, int);
914 extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t);
915 extern int vfs_symlink(struct inode *, struct dentry *, const char *, int);
916 extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
917 extern int vfs_rmdir(struct inode *, struct dentry *);
918 extern int vfs_unlink(struct inode *, struct dentry *);
919 extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
922 * VFS dentry helper functions.
924 extern void dentry_unhash(struct dentry *dentry);
927 * VFS file helper functions.
929 extern int file_permission(struct file *, int);
934 * NOTE! These match bits 12..15 of stat.st_mode
935 * (ie "(i_mode >> 12) & 15").
947 #define OSYNC_METADATA (1<<0)
948 #define OSYNC_DATA (1<<1)
949 #define OSYNC_INODE (1<<2)
950 int generic_osync_inode(struct inode *, struct address_space *, int);
953 * This is the "filldir" function type, used by readdir() to let
954 * the kernel specify what kind of dirent layout it wants to have.
955 * This allows the kernel to read directories into kernel space or
956 * to have different dirent layouts depending on the binary type.
958 typedef int (*filldir_t)(void *, const char *, int, loff_t, ino_t, unsigned);
960 struct block_device_operations {
961 int (*open) (struct inode *, struct file *);
962 int (*release) (struct inode *, struct file *);
963 int (*ioctl) (struct inode *, struct file *, unsigned, unsigned long);
964 long (*unlocked_ioctl) (struct file *, unsigned, unsigned long);
965 long (*compat_ioctl) (struct file *, unsigned, unsigned long);
966 int (*direct_access) (struct block_device *, sector_t, unsigned long *);
967 int (*media_changed) (struct gendisk *);
968 int (*revalidate_disk) (struct gendisk *);
969 int (*getgeo)(struct block_device *, struct hd_geometry *);
970 struct module *owner;
974 * "descriptor" for what we're up to with a read for sendfile().
975 * This allows us to use the same read code yet
976 * have multiple different users of the data that
977 * we read from a file.
979 * The simplest case just copies the data to user
992 typedef int (*read_actor_t)(read_descriptor_t *, struct page *, unsigned long, unsigned long);
994 /* These macros are for out of kernel modules to test that
995 * the kernel supports the unlocked_ioctl and compat_ioctl
996 * fields in struct file_operations. */
997 #define HAVE_COMPAT_IOCTL 1
998 #define HAVE_UNLOCKED_IOCTL 1
1002 * read, write, poll, fsync, readv, writev, unlocked_ioctl and compat_ioctl
1003 * can be called without the big kernel lock held in all filesystems.
1005 struct file_operations {
1006 struct module *owner;
1007 loff_t (*llseek) (struct file *, loff_t, int);
1008 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1009 ssize_t (*aio_read) (struct kiocb *, char __user *, size_t, loff_t);
1010 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
1011 ssize_t (*aio_write) (struct kiocb *, const char __user *, size_t, loff_t);
1012 int (*readdir) (struct file *, void *, filldir_t);
1013 unsigned int (*poll) (struct file *, struct poll_table_struct *);
1014 int (*ioctl) (struct inode *, struct file *, unsigned int, unsigned long);
1015 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1016 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1017 int (*mmap) (struct file *, struct vm_area_struct *);
1018 int (*open) (struct inode *, struct file *);
1019 int (*flush) (struct file *);
1020 int (*release) (struct inode *, struct file *);
1021 int (*fsync) (struct file *, struct dentry *, int datasync);
1022 int (*aio_fsync) (struct kiocb *, int datasync);
1023 int (*fasync) (int, struct file *, int);
1024 int (*lock) (struct file *, int, struct file_lock *);
1025 ssize_t (*readv) (struct file *, const struct iovec *, unsigned long, loff_t *);
1026 ssize_t (*writev) (struct file *, const struct iovec *, unsigned long, loff_t *);
1027 ssize_t (*sendfile) (struct file *, loff_t *, size_t, read_actor_t, void *);
1028 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1029 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1030 int (*check_flags)(int);
1031 int (*dir_notify)(struct file *filp, unsigned long arg);
1032 int (*flock) (struct file *, int, struct file_lock *);
1035 struct inode_operations {
1036 int (*create) (struct inode *,struct dentry *,int, struct nameidata *);
1037 struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
1038 int (*link) (struct dentry *,struct inode *,struct dentry *);
1039 int (*unlink) (struct inode *,struct dentry *);
1040 int (*symlink) (struct inode *,struct dentry *,const char *);
1041 int (*mkdir) (struct inode *,struct dentry *,int);
1042 int (*rmdir) (struct inode *,struct dentry *);
1043 int (*mknod) (struct inode *,struct dentry *,int,dev_t);
1044 int (*rename) (struct inode *, struct dentry *,
1045 struct inode *, struct dentry *);
1046 int (*readlink) (struct dentry *, char __user *,int);
1047 void * (*follow_link) (struct dentry *, struct nameidata *);
1048 void (*put_link) (struct dentry *, struct nameidata *, void *);
1049 void (*truncate) (struct inode *);
1050 int (*permission) (struct inode *, int, struct nameidata *);
1051 int (*setattr) (struct dentry *, struct iattr *);
1052 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1053 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1054 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1055 ssize_t (*listxattr) (struct dentry *, char *, size_t);
1056 int (*removexattr) (struct dentry *, const char *);
1057 void (*truncate_range)(struct inode *, loff_t, loff_t);
1062 extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1063 extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1064 extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1065 unsigned long, loff_t *);
1066 extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1067 unsigned long, loff_t *);
1070 * NOTE: write_inode, delete_inode, clear_inode, put_inode can be called
1071 * without the big kernel lock held in all filesystems.
1073 struct super_operations {
1074 struct inode *(*alloc_inode)(struct super_block *sb);
1075 void (*destroy_inode)(struct inode *);
1077 void (*read_inode) (struct inode *);
1079 void (*dirty_inode) (struct inode *);
1080 int (*write_inode) (struct inode *, int);
1081 void (*put_inode) (struct inode *);
1082 void (*drop_inode) (struct inode *);
1083 void (*delete_inode) (struct inode *);
1084 void (*put_super) (struct super_block *);
1085 void (*write_super) (struct super_block *);
1086 int (*sync_fs)(struct super_block *sb, int wait);
1087 void (*write_super_lockfs) (struct super_block *);
1088 void (*unlockfs) (struct super_block *);
1089 int (*statfs) (struct super_block *, struct kstatfs *);
1090 int (*remount_fs) (struct super_block *, int *, char *);
1091 void (*clear_inode) (struct inode *);
1092 void (*umount_begin) (struct super_block *);
1094 int (*show_options)(struct seq_file *, struct vfsmount *);
1096 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1097 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
1100 /* Inode state bits. Protected by inode_lock. */
1101 #define I_DIRTY_SYNC 1 /* Not dirty enough for O_DATASYNC */
1102 #define I_DIRTY_DATASYNC 2 /* Data-related inode changes pending */
1103 #define I_DIRTY_PAGES 4 /* Data-related inode changes pending */
1105 #define I_LOCK (1 << __I_LOCK)
1106 #define I_FREEING 16
1109 #define I_WILL_FREE 128
1111 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1113 extern void __mark_inode_dirty(struct inode *, int);
1114 static inline void mark_inode_dirty(struct inode *inode)
1116 __mark_inode_dirty(inode, I_DIRTY);
1119 static inline void mark_inode_dirty_sync(struct inode *inode)
1121 __mark_inode_dirty(inode, I_DIRTY_SYNC);
1124 static inline void inode_inc_link_count(struct inode *inode)
1127 mark_inode_dirty(inode);
1130 static inline void inode_dec_link_count(struct inode *inode)
1133 mark_inode_dirty(inode);
1136 extern void touch_atime(struct vfsmount *mnt, struct dentry *dentry);
1137 static inline void file_accessed(struct file *file)
1139 if (!(file->f_flags & O_NOATIME))
1140 touch_atime(file->f_vfsmnt, file->f_dentry);
1143 int sync_inode(struct inode *inode, struct writeback_control *wbc);
1146 * struct export_operations - for nfsd to communicate with file systems
1147 * @decode_fh: decode a file handle fragment and return a &struct dentry
1148 * @encode_fh: encode a file handle fragment from a dentry
1149 * @get_name: find the name for a given inode in a given directory
1150 * @get_parent: find the parent of a given directory
1151 * @get_dentry: find a dentry for the inode given a file handle sub-fragment
1152 * @find_exported_dentry:
1153 * set by the exporting module to a standard helper function.
1156 * The export_operations structure provides a means for nfsd to communicate
1157 * with a particular exported file system - particularly enabling nfsd and
1158 * the filesystem to co-operate when dealing with file handles.
1160 * export_operations contains two basic operation for dealing with file
1161 * handles, decode_fh() and encode_fh(), and allows for some other
1162 * operations to be defined which standard helper routines use to get
1163 * specific information from the filesystem.
1165 * nfsd encodes information use to determine which filesystem a filehandle
1166 * applies to in the initial part of the file handle. The remainder, termed
1167 * a file handle fragment, is controlled completely by the filesystem. The
1168 * standard helper routines assume that this fragment will contain one or
1169 * two sub-fragments, one which identifies the file, and one which may be
1170 * used to identify the (a) directory containing the file.
1172 * In some situations, nfsd needs to get a dentry which is connected into a
1173 * specific part of the file tree. To allow for this, it passes the
1174 * function acceptable() together with a @context which can be used to see
1175 * if the dentry is acceptable. As there can be multiple dentrys for a
1176 * given file, the filesystem should check each one for acceptability before
1177 * looking for the next. As soon as an acceptable one is found, it should
1181 * @decode_fh is given a &struct super_block (@sb), a file handle fragment
1182 * (@fh, @fh_len) and an acceptability testing function (@acceptable,
1183 * @context). It should return a &struct dentry which refers to the same
1184 * file that the file handle fragment refers to, and which passes the
1185 * acceptability test. If it cannot, it should return a %NULL pointer if
1186 * the file was found but no acceptable &dentries were available, or a
1187 * %ERR_PTR error code indicating why it couldn't be found (e.g. %ENOENT or
1191 * @encode_fh should store in the file handle fragment @fh (using at most
1192 * @max_len bytes) information that can be used by @decode_fh to recover the
1193 * file refered to by the &struct dentry @de. If the @connectable flag is
1194 * set, the encode_fh() should store sufficient information so that a good
1195 * attempt can be made to find not only the file but also it's place in the
1196 * filesystem. This typically means storing a reference to de->d_parent in
1197 * the filehandle fragment. encode_fh() should return the number of bytes
1198 * stored or a negative error code such as %-ENOSPC
1201 * @get_name should find a name for the given @child in the given @parent
1202 * directory. The name should be stored in the @name (with the
1203 * understanding that it is already pointing to a a %NAME_MAX+1 sized
1204 * buffer. get_name() should return %0 on success, a negative error code
1205 * or error. @get_name will be called without @parent->i_mutex held.
1208 * @get_parent should find the parent directory for the given @child which
1209 * is also a directory. In the event that it cannot be found, or storage
1210 * space cannot be allocated, a %ERR_PTR should be returned.
1213 * Given a &super_block (@sb) and a pointer to a file-system specific inode
1214 * identifier, possibly an inode number, (@inump) get_dentry() should find
1215 * the identified inode and return a dentry for that inode. Any suitable
1216 * dentry can be returned including, if necessary, a new dentry created with
1217 * d_alloc_root. The caller can then find any other extant dentrys by
1218 * following the d_alias links. If a new dentry was created using
1219 * d_alloc_root, DCACHE_NFSD_DISCONNECTED should be set, and the dentry
1220 * should be d_rehash()ed.
1222 * If the inode cannot be found, either a %NULL pointer or an %ERR_PTR code
1223 * can be returned. The @inump will be whatever was passed to
1224 * nfsd_find_fh_dentry() in either the @obj or @parent parameters.
1227 * get_parent is called with child->d_inode->i_mutex down
1228 * get_name is not (which is possibly inconsistent)
1231 struct export_operations {
1232 struct dentry *(*decode_fh)(struct super_block *sb, __u32 *fh, int fh_len, int fh_type,
1233 int (*acceptable)(void *context, struct dentry *de),
1235 int (*encode_fh)(struct dentry *de, __u32 *fh, int *max_len,
1238 /* the following are only called from the filesystem itself */
1239 int (*get_name)(struct dentry *parent, char *name,
1240 struct dentry *child);
1241 struct dentry * (*get_parent)(struct dentry *child);
1242 struct dentry * (*get_dentry)(struct super_block *sb, void *inump);
1244 /* This is set by the exporting module to a standard helper */
1245 struct dentry * (*find_exported_dentry)(
1246 struct super_block *sb, void *obj, void *parent,
1247 int (*acceptable)(void *context, struct dentry *de),
1253 extern struct dentry *
1254 find_exported_dentry(struct super_block *sb, void *obj, void *parent,
1255 int (*acceptable)(void *context, struct dentry *de),
1258 struct file_system_type {
1261 struct super_block *(*get_sb) (struct file_system_type *, int,
1262 const char *, void *);
1263 void (*kill_sb) (struct super_block *);
1264 struct module *owner;
1265 struct file_system_type * next;
1266 struct list_head fs_supers;
1269 struct super_block *get_sb_bdev(struct file_system_type *fs_type,
1270 int flags, const char *dev_name, void *data,
1271 int (*fill_super)(struct super_block *, void *, int));
1272 struct super_block *get_sb_single(struct file_system_type *fs_type,
1273 int flags, void *data,
1274 int (*fill_super)(struct super_block *, void *, int));
1275 struct super_block *get_sb_nodev(struct file_system_type *fs_type,
1276 int flags, void *data,
1277 int (*fill_super)(struct super_block *, void *, int));
1278 void generic_shutdown_super(struct super_block *sb);
1279 void kill_block_super(struct super_block *sb);
1280 void kill_anon_super(struct super_block *sb);
1281 void kill_litter_super(struct super_block *sb);
1282 void deactivate_super(struct super_block *sb);
1283 int set_anon_super(struct super_block *s, void *data);
1284 struct super_block *sget(struct file_system_type *type,
1285 int (*test)(struct super_block *,void *),
1286 int (*set)(struct super_block *,void *),
1288 struct super_block *get_sb_pseudo(struct file_system_type *, char *,
1289 struct super_operations *ops, unsigned long);
1290 int __put_super(struct super_block *sb);
1291 int __put_super_and_need_restart(struct super_block *sb);
1292 void unnamed_dev_init(void);
1294 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1295 #define fops_get(fops) \
1296 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1297 #define fops_put(fops) \
1298 do { if (fops) module_put((fops)->owner); } while(0)
1300 extern int register_filesystem(struct file_system_type *);
1301 extern int unregister_filesystem(struct file_system_type *);
1302 extern struct vfsmount *kern_mount(struct file_system_type *);
1303 extern int may_umount_tree(struct vfsmount *);
1304 extern int may_umount(struct vfsmount *);
1305 extern void umount_tree(struct vfsmount *, int, struct list_head *);
1306 extern void release_mounts(struct list_head *);
1307 extern long do_mount(char *, char *, char *, unsigned long, void *);
1308 extern struct vfsmount *copy_tree(struct vfsmount *, struct dentry *, int);
1309 extern void mnt_set_mountpoint(struct vfsmount *, struct dentry *,
1312 extern int vfs_statfs(struct super_block *, struct kstatfs *);
1315 extern struct subsystem fs_subsys;
1317 #define FLOCK_VERIFY_READ 1
1318 #define FLOCK_VERIFY_WRITE 2
1320 extern int locks_mandatory_locked(struct inode *);
1321 extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
1324 * Candidates for mandatory locking have the setgid bit set
1325 * but no group execute bit - an otherwise meaningless combination.
1327 #define MANDATORY_LOCK(inode) \
1328 (IS_MANDLOCK(inode) && ((inode)->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID)
1330 static inline int locks_verify_locked(struct inode *inode)
1332 if (MANDATORY_LOCK(inode))
1333 return locks_mandatory_locked(inode);
1337 extern int rw_verify_area(int, struct file *, loff_t *, size_t);
1339 static inline int locks_verify_truncate(struct inode *inode,
1343 if (inode->i_flock && MANDATORY_LOCK(inode))
1344 return locks_mandatory_area(
1345 FLOCK_VERIFY_WRITE, inode, filp,
1346 size < inode->i_size ? size : inode->i_size,
1347 (size < inode->i_size ? inode->i_size - size
1348 : size - inode->i_size)
1353 static inline int break_lease(struct inode *inode, unsigned int mode)
1356 return __break_lease(inode, mode);
1362 extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
1364 extern long do_sys_open(int fdf, const char __user *filename, int flags,
1366 extern struct file *filp_open(const char *, int, int);
1367 extern struct file * dentry_open(struct dentry *, struct vfsmount *, int);
1368 extern int filp_close(struct file *, fl_owner_t id);
1369 extern char * getname(const char __user *);
1372 extern void __init vfs_caches_init_early(void);
1373 extern void __init vfs_caches_init(unsigned long);
1375 #define __getname() kmem_cache_alloc(names_cachep, SLAB_KERNEL)
1376 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
1377 #ifndef CONFIG_AUDITSYSCALL
1378 #define putname(name) __putname(name)
1380 extern void putname(const char *name);
1383 extern int register_blkdev(unsigned int, const char *);
1384 extern int unregister_blkdev(unsigned int, const char *);
1385 extern struct block_device *bdget(dev_t);
1386 extern void bd_set_size(struct block_device *, loff_t size);
1387 extern void bd_forget(struct inode *inode);
1388 extern void bdput(struct block_device *);
1389 extern struct block_device *open_by_devnum(dev_t, unsigned);
1390 extern struct file_operations def_blk_fops;
1391 extern struct address_space_operations def_blk_aops;
1392 extern struct file_operations def_chr_fops;
1393 extern struct file_operations bad_sock_fops;
1394 extern struct file_operations def_fifo_fops;
1395 extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
1396 extern int blkdev_ioctl(struct inode *, struct file *, unsigned, unsigned long);
1397 extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
1398 extern int blkdev_get(struct block_device *, mode_t, unsigned);
1399 extern int blkdev_put(struct block_device *);
1400 extern int bd_claim(struct block_device *, void *);
1401 extern void bd_release(struct block_device *);
1404 extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
1405 extern int register_chrdev_region(dev_t, unsigned, const char *);
1406 extern int register_chrdev(unsigned int, const char *,
1407 struct file_operations *);
1408 extern int unregister_chrdev(unsigned int, const char *);
1409 extern void unregister_chrdev_region(dev_t, unsigned);
1410 extern int chrdev_open(struct inode *, struct file *);
1411 extern int get_chrdev_list(char *);
1412 extern void *acquire_chrdev_list(void);
1413 extern int count_chrdev_list(void);
1414 extern void *get_next_chrdev(void *);
1415 extern int get_chrdev_info(void *, int *, char **);
1416 extern void release_chrdev_list(void *);
1418 /* fs/block_dev.c */
1419 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
1420 extern const char *__bdevname(dev_t, char *buffer);
1421 extern const char *bdevname(struct block_device *bdev, char *buffer);
1422 extern struct block_device *lookup_bdev(const char *);
1423 extern struct block_device *open_bdev_excl(const char *, int, void *);
1424 extern void close_bdev_excl(struct block_device *);
1425 extern void *acquire_blkdev_list(void);
1426 extern int count_blkdev_list(void);
1427 extern void *get_next_blkdev(void *);
1428 extern int get_blkdev_info(void *, int *, char **);
1429 extern void release_blkdev_list(void *);
1431 extern void init_special_inode(struct inode *, umode_t, dev_t);
1433 /* Invalid inode operations -- fs/bad_inode.c */
1434 extern void make_bad_inode(struct inode *);
1435 extern int is_bad_inode(struct inode *);
1437 extern struct file_operations read_fifo_fops;
1438 extern struct file_operations write_fifo_fops;
1439 extern struct file_operations rdwr_fifo_fops;
1441 extern int fs_may_remount_ro(struct super_block *);
1444 * return READ, READA, or WRITE
1446 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
1449 * return data direction, READ or WRITE
1451 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
1453 extern int check_disk_change(struct block_device *);
1454 extern int invalidate_inodes(struct super_block *);
1455 extern int __invalidate_device(struct block_device *);
1456 extern int invalidate_partition(struct gendisk *, int);
1457 unsigned long invalidate_mapping_pages(struct address_space *mapping,
1458 pgoff_t start, pgoff_t end);
1459 unsigned long invalidate_inode_pages(struct address_space *mapping);
1460 static inline void invalidate_remote_inode(struct inode *inode)
1462 if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
1463 S_ISLNK(inode->i_mode))
1464 invalidate_inode_pages(inode->i_mapping);
1466 extern int invalidate_inode_pages2(struct address_space *mapping);
1467 extern int invalidate_inode_pages2_range(struct address_space *mapping,
1468 pgoff_t start, pgoff_t end);
1469 extern int write_inode_now(struct inode *, int);
1470 extern int filemap_fdatawrite(struct address_space *);
1471 extern int filemap_flush(struct address_space *);
1472 extern int filemap_fdatawait(struct address_space *);
1473 extern int filemap_write_and_wait(struct address_space *mapping);
1474 extern int filemap_write_and_wait_range(struct address_space *mapping,
1475 loff_t lstart, loff_t lend);
1476 extern int wait_on_page_writeback_range(struct address_space *mapping,
1477 pgoff_t start, pgoff_t end);
1478 extern int __filemap_fdatawrite_range(struct address_space *mapping,
1479 loff_t start, loff_t end, int sync_mode);
1481 extern long do_fsync(struct file *file, int datasync);
1482 extern void sync_supers(void);
1483 extern void sync_filesystems(int wait);
1484 extern void emergency_sync(void);
1485 extern void emergency_remount(void);
1486 extern int do_remount_sb(struct super_block *sb, int flags,
1487 void *data, int force);
1488 extern sector_t bmap(struct inode *, sector_t);
1489 extern int notify_change(struct dentry *, struct iattr *);
1490 extern int permission(struct inode *, int, struct nameidata *);
1491 extern int generic_permission(struct inode *, int,
1492 int (*check_acl)(struct inode *, int));
1494 extern int get_write_access(struct inode *);
1495 extern int deny_write_access(struct file *);
1496 static inline void put_write_access(struct inode * inode)
1498 atomic_dec(&inode->i_writecount);
1500 static inline void allow_write_access(struct file *file)
1503 atomic_inc(&file->f_dentry->d_inode->i_writecount);
1505 extern int do_pipe(int *);
1507 extern int open_namei(int dfd, const char *, int, int, struct nameidata *);
1508 extern int may_open(struct nameidata *, int, int);
1510 extern int kernel_read(struct file *, unsigned long, char *, unsigned long);
1511 extern struct file * open_exec(const char *);
1513 /* fs/dcache.c -- generic fs support functions */
1514 extern int is_subdir(struct dentry *, struct dentry *);
1515 extern ino_t find_inode_number(struct dentry *, struct qstr *);
1517 #include <linux/err.h>
1519 /* needed for stackable file system support */
1520 extern loff_t default_llseek(struct file *file, loff_t offset, int origin);
1522 extern loff_t vfs_llseek(struct file *file, loff_t offset, int origin);
1524 extern void inode_init_once(struct inode *);
1525 extern void iput(struct inode *);
1526 extern struct inode * igrab(struct inode *);
1527 extern ino_t iunique(struct super_block *, ino_t);
1528 extern int inode_needs_sync(struct inode *inode);
1529 extern void generic_delete_inode(struct inode *inode);
1530 extern void generic_drop_inode(struct inode *inode);
1532 extern struct inode *ilookup5_nowait(struct super_block *sb,
1533 unsigned long hashval, int (*test)(struct inode *, void *),
1535 extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
1536 int (*test)(struct inode *, void *), void *data);
1537 extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
1539 extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
1540 extern struct inode * iget_locked(struct super_block *, unsigned long);
1541 extern void unlock_new_inode(struct inode *);
1543 static inline struct inode *iget(struct super_block *sb, unsigned long ino)
1545 struct inode *inode = iget_locked(sb, ino);
1547 if (inode && (inode->i_state & I_NEW)) {
1548 sb->s_op->read_inode(inode);
1549 unlock_new_inode(inode);
1555 extern void __iget(struct inode * inode);
1556 extern void clear_inode(struct inode *);
1557 extern void destroy_inode(struct inode *);
1558 extern struct inode *new_inode(struct super_block *);
1559 extern int remove_suid(struct dentry *);
1560 extern void remove_dquot_ref(struct super_block *, int, struct list_head *);
1561 extern struct mutex iprune_mutex;
1563 extern void __insert_inode_hash(struct inode *, unsigned long hashval);
1564 extern void remove_inode_hash(struct inode *);
1565 static inline void insert_inode_hash(struct inode *inode) {
1566 __insert_inode_hash(inode, inode->i_ino);
1569 extern struct file * get_empty_filp(void);
1570 extern void file_move(struct file *f, struct list_head *list);
1571 extern void file_kill(struct file *f);
1573 extern void submit_bio(int, struct bio *);
1574 extern int bdev_read_only(struct block_device *);
1575 extern int set_blocksize(struct block_device *, int);
1576 extern int sb_set_blocksize(struct super_block *, int);
1577 extern int sb_min_blocksize(struct super_block *, int);
1579 extern int generic_file_mmap(struct file *, struct vm_area_struct *);
1580 extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
1581 extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
1582 extern int file_send_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
1583 extern ssize_t generic_file_read(struct file *, char __user *, size_t, loff_t *);
1584 int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
1585 extern ssize_t generic_file_write(struct file *, const char __user *, size_t, loff_t *);
1586 extern ssize_t generic_file_aio_read(struct kiocb *, char __user *, size_t, loff_t);
1587 extern ssize_t __generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t *);
1588 extern ssize_t generic_file_aio_write(struct kiocb *, const char __user *, size_t, loff_t);
1589 extern ssize_t generic_file_aio_write_nolock(struct kiocb *, const struct iovec *,
1590 unsigned long, loff_t *);
1591 extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
1592 unsigned long *, loff_t, loff_t *, size_t, size_t);
1593 extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
1594 unsigned long, loff_t, loff_t *, size_t, ssize_t);
1595 extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
1596 extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
1597 ssize_t generic_file_write_nolock(struct file *file, const struct iovec *iov,
1598 unsigned long nr_segs, loff_t *ppos);
1599 extern ssize_t generic_file_sendfile(struct file *, loff_t *, size_t, read_actor_t, void *);
1600 extern void do_generic_mapping_read(struct address_space *mapping,
1601 struct file_ra_state *, struct file *,
1602 loff_t *, read_descriptor_t *, read_actor_t);
1604 file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
1605 extern ssize_t generic_file_readv(struct file *filp, const struct iovec *iov,
1606 unsigned long nr_segs, loff_t *ppos);
1607 ssize_t generic_file_writev(struct file *filp, const struct iovec *iov,
1608 unsigned long nr_segs, loff_t *ppos);
1609 extern loff_t no_llseek(struct file *file, loff_t offset, int origin);
1610 extern loff_t generic_file_llseek(struct file *file, loff_t offset, int origin);
1611 extern loff_t remote_llseek(struct file *file, loff_t offset, int origin);
1612 extern int generic_file_open(struct inode * inode, struct file * filp);
1613 extern int nonseekable_open(struct inode * inode, struct file * filp);
1615 #ifdef CONFIG_FS_XIP
1616 extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
1618 extern ssize_t xip_file_sendfile(struct file *in_file, loff_t *ppos,
1619 size_t count, read_actor_t actor,
1621 extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
1622 extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
1623 size_t len, loff_t *ppos);
1624 extern int xip_truncate_page(struct address_space *mapping, loff_t from);
1626 static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
1632 static inline void do_generic_file_read(struct file * filp, loff_t *ppos,
1633 read_descriptor_t * desc,
1636 do_generic_mapping_read(filp->f_mapping,
1644 ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
1645 struct block_device *bdev, const struct iovec *iov, loff_t offset,
1646 unsigned long nr_segs, get_blocks_t get_blocks, dio_iodone_t end_io,
1650 DIO_LOCKING = 1, /* need locking between buffered and direct access */
1651 DIO_NO_LOCKING, /* bdev; no locking at all between buffered/direct */
1652 DIO_OWN_LOCKING, /* filesystem locks buffered and direct internally */
1655 static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
1656 struct inode *inode, struct block_device *bdev, const struct iovec *iov,
1657 loff_t offset, unsigned long nr_segs, get_blocks_t get_blocks,
1658 dio_iodone_t end_io)
1660 return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset,
1661 nr_segs, get_blocks, end_io, DIO_LOCKING);
1664 static inline ssize_t blockdev_direct_IO_no_locking(int rw, struct kiocb *iocb,
1665 struct inode *inode, struct block_device *bdev, const struct iovec *iov,
1666 loff_t offset, unsigned long nr_segs, get_blocks_t get_blocks,
1667 dio_iodone_t end_io)
1669 return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset,
1670 nr_segs, get_blocks, end_io, DIO_NO_LOCKING);
1673 static inline ssize_t blockdev_direct_IO_own_locking(int rw, struct kiocb *iocb,
1674 struct inode *inode, struct block_device *bdev, const struct iovec *iov,
1675 loff_t offset, unsigned long nr_segs, get_blocks_t get_blocks,
1676 dio_iodone_t end_io)
1678 return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset,
1679 nr_segs, get_blocks, end_io, DIO_OWN_LOCKING);
1682 extern struct file_operations generic_ro_fops;
1684 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
1686 extern int vfs_readlink(struct dentry *, char __user *, int, const char *);
1687 extern int vfs_follow_link(struct nameidata *, const char *);
1688 extern int page_readlink(struct dentry *, char __user *, int);
1689 extern void *page_follow_link_light(struct dentry *, struct nameidata *);
1690 extern void page_put_link(struct dentry *, struct nameidata *, void *);
1691 extern int __page_symlink(struct inode *inode, const char *symname, int len,
1693 extern int page_symlink(struct inode *inode, const char *symname, int len);
1694 extern struct inode_operations page_symlink_inode_operations;
1695 extern int generic_readlink(struct dentry *, char __user *, int);
1696 extern void generic_fillattr(struct inode *, struct kstat *);
1697 extern int vfs_getattr(struct vfsmount *, struct dentry *, struct kstat *);
1698 void inode_add_bytes(struct inode *inode, loff_t bytes);
1699 void inode_sub_bytes(struct inode *inode, loff_t bytes);
1700 loff_t inode_get_bytes(struct inode *inode);
1701 void inode_set_bytes(struct inode *inode, loff_t bytes);
1703 extern int vfs_readdir(struct file *, filldir_t, void *);
1705 extern int vfs_stat(char __user *, struct kstat *);
1706 extern int vfs_lstat(char __user *, struct kstat *);
1707 extern int vfs_stat_fd(int dfd, char __user *, struct kstat *);
1708 extern int vfs_lstat_fd(int dfd, char __user *, struct kstat *);
1709 extern int vfs_fstat(unsigned int, struct kstat *);
1711 extern int vfs_ioctl(struct file *, unsigned int, unsigned int, unsigned long);
1713 extern struct file_system_type *get_fs_type(const char *name);
1714 extern struct super_block *get_super(struct block_device *);
1715 extern struct super_block *user_get_super(dev_t);
1716 extern void drop_super(struct super_block *sb);
1718 extern int dcache_dir_open(struct inode *, struct file *);
1719 extern int dcache_dir_close(struct inode *, struct file *);
1720 extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
1721 extern int dcache_readdir(struct file *, void *, filldir_t);
1722 extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
1723 extern int simple_statfs(struct super_block *, struct kstatfs *);
1724 extern int simple_link(struct dentry *, struct inode *, struct dentry *);
1725 extern int simple_unlink(struct inode *, struct dentry *);
1726 extern int simple_rmdir(struct inode *, struct dentry *);
1727 extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
1728 extern int simple_sync_file(struct file *, struct dentry *, int);
1729 extern int simple_empty(struct dentry *);
1730 extern int simple_readpage(struct file *file, struct page *page);
1731 extern int simple_prepare_write(struct file *file, struct page *page,
1732 unsigned offset, unsigned to);
1733 extern int simple_commit_write(struct file *file, struct page *page,
1734 unsigned offset, unsigned to);
1736 extern struct dentry *simple_lookup(struct inode *, struct dentry *, struct nameidata *);
1737 extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
1738 extern struct file_operations simple_dir_operations;
1739 extern struct inode_operations simple_dir_inode_operations;
1740 struct tree_descr { char *name; struct file_operations *ops; int mode; };
1741 struct dentry *d_alloc_name(struct dentry *, const char *);
1742 extern int simple_fill_super(struct super_block *, int, struct tree_descr *);
1743 extern int simple_pin_fs(char *name, struct vfsmount **mount, int *count);
1744 extern void simple_release_fs(struct vfsmount **mount, int *count);
1746 extern ssize_t simple_read_from_buffer(void __user *, size_t, loff_t *, const void *, size_t);
1748 #ifdef CONFIG_MIGRATION
1749 extern int buffer_migrate_page(struct page *, struct page *);
1751 #define buffer_migrate_page NULL
1754 extern int inode_change_ok(struct inode *, struct iattr *);
1755 extern int __must_check inode_setattr(struct inode *, struct iattr *);
1757 extern void file_update_time(struct file *file);
1759 static inline ino_t parent_ino(struct dentry *dentry)
1763 spin_lock(&dentry->d_lock);
1764 res = dentry->d_parent->d_inode->i_ino;
1765 spin_unlock(&dentry->d_lock);
1770 extern int unshare_files(void);
1772 /* Transaction based IO helpers */
1775 * An argresp is stored in an allocated page and holds the
1776 * size of the argument or response, along with its content
1778 struct simple_transaction_argresp {
1783 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
1785 char *simple_transaction_get(struct file *file, const char __user *buf,
1787 ssize_t simple_transaction_read(struct file *file, char __user *buf,
1788 size_t size, loff_t *pos);
1789 int simple_transaction_release(struct inode *inode, struct file *file);
1791 static inline void simple_transaction_set(struct file *file, size_t n)
1793 struct simple_transaction_argresp *ar = file->private_data;
1795 BUG_ON(n > SIMPLE_TRANSACTION_LIMIT);
1798 * The barrier ensures that ar->size will really remain zero until
1799 * ar->data is ready for reading.
1806 * simple attribute files
1808 * These attributes behave similar to those in sysfs:
1810 * Writing to an attribute immediately sets a value, an open file can be
1811 * written to multiple times.
1813 * Reading from an attribute creates a buffer from the value that might get
1814 * read with multiple read calls. When the attribute has been read
1815 * completely, no further read calls are possible until the file is opened
1818 * All attributes contain a text representation of a numeric value
1819 * that are accessed with the get() and set() functions.
1821 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
1822 static int __fops ## _open(struct inode *inode, struct file *file) \
1824 __simple_attr_check_format(__fmt, 0ull); \
1825 return simple_attr_open(inode, file, __get, __set, __fmt); \
1827 static struct file_operations __fops = { \
1828 .owner = THIS_MODULE, \
1829 .open = __fops ## _open, \
1830 .release = simple_attr_close, \
1831 .read = simple_attr_read, \
1832 .write = simple_attr_write, \
1835 static inline void __attribute__((format(printf, 1, 2)))
1836 __simple_attr_check_format(const char *fmt, ...)
1838 /* don't do anything, just let the compiler check the arguments; */
1841 int simple_attr_open(struct inode *inode, struct file *file,
1842 u64 (*get)(void *), void (*set)(void *, u64),
1844 int simple_attr_close(struct inode *inode, struct file *file);
1845 ssize_t simple_attr_read(struct file *file, char __user *buf,
1846 size_t len, loff_t *ppos);
1847 ssize_t simple_attr_write(struct file *file, const char __user *buf,
1848 size_t len, loff_t *ppos);
1851 #ifdef CONFIG_SECURITY
1852 static inline char *alloc_secdata(void)
1854 return (char *)get_zeroed_page(GFP_KERNEL);
1857 static inline void free_secdata(void *secdata)
1859 free_page((unsigned long)secdata);
1862 static inline char *alloc_secdata(void)
1867 static inline void free_secdata(void *secdata)
1869 #endif /* CONFIG_SECURITY */
1871 #endif /* __KERNEL__ */
1872 #endif /* _LINUX_FS_H */