5 #include <linux/linkage.h>
6 #include <linux/wait.h>
7 #include <linux/kdev_t.h>
8 #include <linux/dcache.h>
9 #include <linux/path.h>
10 #include <linux/stat.h>
11 #include <linux/cache.h>
12 #include <linux/list.h>
13 #include <linux/llist.h>
14 #include <linux/radix-tree.h>
15 #include <linux/rbtree.h>
16 #include <linux/init.h>
17 #include <linux/pid.h>
18 #include <linux/bug.h>
19 #include <linux/mutex.h>
20 #include <linux/capability.h>
21 #include <linux/semaphore.h>
22 #include <linux/fiemap.h>
23 #include <linux/rculist_bl.h>
24 #include <linux/atomic.h>
25 #include <linux/shrinker.h>
26 #include <linux/migrate_mode.h>
27 #include <linux/uidgid.h>
28 #include <linux/lockdep.h>
29 #include <linux/percpu-rwsem.h>
30 #include <linux/blk_types.h>
32 #include <asm/byteorder.h>
33 #include <uapi/linux/fs.h>
35 struct export_operations
;
41 struct pipe_inode_info
;
42 struct poll_table_struct
;
44 struct vm_area_struct
;
47 struct swap_info_struct
;
49 struct workqueue_struct
;
51 extern void __init
inode_init(void);
52 extern void __init
inode_init_early(void);
53 extern void __init
files_init(unsigned long);
55 extern struct files_stat_struct files_stat
;
56 extern unsigned long get_max_files(void);
57 extern int sysctl_nr_open
;
58 extern struct inodes_stat_t inodes_stat
;
59 extern int leases_enable
, lease_break_time
;
60 extern int sysctl_protected_symlinks
;
61 extern int sysctl_protected_hardlinks
;
64 typedef int (get_block_t
)(struct inode
*inode
, sector_t iblock
,
65 struct buffer_head
*bh_result
, int create
);
66 typedef void (dio_iodone_t
)(struct kiocb
*iocb
, loff_t offset
,
67 ssize_t bytes
, void *private);
69 #define MAY_EXEC 0x00000001
70 #define MAY_WRITE 0x00000002
71 #define MAY_READ 0x00000004
72 #define MAY_APPEND 0x00000008
73 #define MAY_ACCESS 0x00000010
74 #define MAY_OPEN 0x00000020
75 #define MAY_CHDIR 0x00000040
76 /* called from RCU mode, don't block */
77 #define MAY_NOT_BLOCK 0x00000080
80 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
81 * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
84 /* file is open for reading */
85 #define FMODE_READ ((__force fmode_t)0x1)
86 /* file is open for writing */
87 #define FMODE_WRITE ((__force fmode_t)0x2)
88 /* file is seekable */
89 #define FMODE_LSEEK ((__force fmode_t)0x4)
90 /* file can be accessed using pread */
91 #define FMODE_PREAD ((__force fmode_t)0x8)
92 /* file can be accessed using pwrite */
93 #define FMODE_PWRITE ((__force fmode_t)0x10)
94 /* File is opened for execution with sys_execve / sys_uselib */
95 #define FMODE_EXEC ((__force fmode_t)0x20)
96 /* File is opened with O_NDELAY (only set for block devices) */
97 #define FMODE_NDELAY ((__force fmode_t)0x40)
98 /* File is opened with O_EXCL (only set for block devices) */
99 #define FMODE_EXCL ((__force fmode_t)0x80)
100 /* File is opened using open(.., 3, ..) and is writeable only for ioctls
101 (specialy hack for floppy.c) */
102 #define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
103 /* 32bit hashes as llseek() offset (for directories) */
104 #define FMODE_32BITHASH ((__force fmode_t)0x200)
105 /* 64bit hashes as llseek() offset (for directories) */
106 #define FMODE_64BITHASH ((__force fmode_t)0x400)
109 * Don't update ctime and mtime.
111 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
112 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
114 #define FMODE_NOCMTIME ((__force fmode_t)0x800)
116 /* Expect random access pattern */
117 #define FMODE_RANDOM ((__force fmode_t)0x1000)
119 /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
120 #define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
122 /* File is opened with O_PATH; almost nothing can be done with it */
123 #define FMODE_PATH ((__force fmode_t)0x4000)
125 /* File was opened by fanotify and shouldn't generate fanotify events */
126 #define FMODE_NONOTIFY ((__force fmode_t)0x1000000)
129 * Flag for rw_copy_check_uvector and compat_rw_copy_check_uvector
130 * that indicates that they should check the contents of the iovec are
131 * valid, but not check the memory that the iovec elements
134 #define CHECK_IOVEC_ONLY -1
137 * The below are the various read and write types that we support. Some of
138 * them include behavioral modifiers that send information down to the
139 * block layer and IO scheduler. Terminology:
141 * The block layer uses device plugging to defer IO a little bit, in
142 * the hope that we will see more IO very shortly. This increases
143 * coalescing of adjacent IO and thus reduces the number of IOs we
144 * have to send to the device. It also allows for better queuing,
145 * if the IO isn't mergeable. If the caller is going to be waiting
146 * for the IO, then he must ensure that the device is unplugged so
147 * that the IO is dispatched to the driver.
149 * All IO is handled async in Linux. This is fine for background
150 * writes, but for reads or writes that someone waits for completion
151 * on, we want to notify the block layer and IO scheduler so that they
152 * know about it. That allows them to make better scheduling
153 * decisions. So when the below references 'sync' and 'async', it
154 * is referencing this priority hint.
156 * With that in mind, the available types are:
158 * READ A normal read operation. Device will be plugged.
159 * READ_SYNC A synchronous read. Device is not plugged, caller can
160 * immediately wait on this read without caring about
162 * READA Used for read-ahead operations. Lower priority, and the
163 * block layer could (in theory) choose to ignore this
164 * request if it runs into resource problems.
165 * WRITE A normal async write. Device will be plugged.
166 * WRITE_SYNC Synchronous write. Identical to WRITE, but passes down
167 * the hint that someone will be waiting on this IO
168 * shortly. The write equivalent of READ_SYNC.
169 * WRITE_ODIRECT Special case write for O_DIRECT only.
170 * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
171 * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
172 * non-volatile media on completion.
173 * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
174 * by a cache flush and data is guaranteed to be on
175 * non-volatile media on completion.
178 #define RW_MASK REQ_WRITE
179 #define RWA_MASK REQ_RAHEAD
182 #define WRITE RW_MASK
183 #define READA RWA_MASK
184 #define KERNEL_READ (READ|REQ_KERNEL)
185 #define KERNEL_WRITE (WRITE|REQ_KERNEL)
187 #define READ_SYNC (READ | REQ_SYNC)
188 #define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE)
189 #define WRITE_ODIRECT (WRITE | REQ_SYNC)
190 #define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
191 #define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
192 #define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
195 * Attribute flags. These should be or-ed together to figure out what
198 #define ATTR_MODE (1 << 0)
199 #define ATTR_UID (1 << 1)
200 #define ATTR_GID (1 << 2)
201 #define ATTR_SIZE (1 << 3)
202 #define ATTR_ATIME (1 << 4)
203 #define ATTR_MTIME (1 << 5)
204 #define ATTR_CTIME (1 << 6)
205 #define ATTR_ATIME_SET (1 << 7)
206 #define ATTR_MTIME_SET (1 << 8)
207 #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
208 #define ATTR_ATTR_FLAG (1 << 10)
209 #define ATTR_KILL_SUID (1 << 11)
210 #define ATTR_KILL_SGID (1 << 12)
211 #define ATTR_FILE (1 << 13)
212 #define ATTR_KILL_PRIV (1 << 14)
213 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
214 #define ATTR_TIMES_SET (1 << 16)
217 * This is the Inode Attributes structure, used for notify_change(). It
218 * uses the above definitions as flags, to know which values have changed.
219 * Also, in this manner, a Filesystem can look at only the values it cares
220 * about. Basically, these are the attributes that the VFS layer can
221 * request to change from the FS layer.
223 * Derek Atkins <warlord@MIT.EDU> 94-10-20
226 unsigned int ia_valid
;
231 struct timespec ia_atime
;
232 struct timespec ia_mtime
;
233 struct timespec ia_ctime
;
236 * Not an attribute, but an auxiliary info for filesystems wanting to
237 * implement an ftruncate() like method. NOTE: filesystem should
238 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
240 struct file
*ia_file
;
244 * Includes for diskquotas.
246 #include <linux/quota.h>
249 * enum positive_aop_returns - aop return codes with specific semantics
251 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
252 * completed, that the page is still locked, and
253 * should be considered active. The VM uses this hint
254 * to return the page to the active list -- it won't
255 * be a candidate for writeback again in the near
256 * future. Other callers must be careful to unlock
257 * the page if they get this return. Returned by
260 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
261 * unlocked it and the page might have been truncated.
262 * The caller should back up to acquiring a new page and
263 * trying again. The aop will be taking reasonable
264 * precautions not to livelock. If the caller held a page
265 * reference, it should drop it before retrying. Returned
268 * address_space_operation functions return these large constants to indicate
269 * special semantics to the caller. These are much larger than the bytes in a
270 * page to allow for functions that return the number of bytes operated on in a
274 enum positive_aop_returns
{
275 AOP_WRITEPAGE_ACTIVATE
= 0x80000,
276 AOP_TRUNCATED_PAGE
= 0x80001,
279 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
280 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
281 #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
282 * helper code (eg buffer layer)
283 * to clear GFP_FS from alloc */
286 * oh the beauties of C type declarations.
289 struct address_space
;
290 struct writeback_control
;
293 const struct iovec
*iov
;
294 unsigned long nr_segs
;
299 size_t iov_iter_copy_from_user_atomic(struct page
*page
,
300 struct iov_iter
*i
, unsigned long offset
, size_t bytes
);
301 size_t iov_iter_copy_from_user(struct page
*page
,
302 struct iov_iter
*i
, unsigned long offset
, size_t bytes
);
303 void iov_iter_advance(struct iov_iter
*i
, size_t bytes
);
304 int iov_iter_fault_in_readable(struct iov_iter
*i
, size_t bytes
);
305 size_t iov_iter_single_seg_count(const struct iov_iter
*i
);
307 static inline void iov_iter_init(struct iov_iter
*i
,
308 const struct iovec
*iov
, unsigned long nr_segs
,
309 size_t count
, size_t written
)
312 i
->nr_segs
= nr_segs
;
314 i
->count
= count
+ written
;
316 iov_iter_advance(i
, written
);
319 static inline size_t iov_iter_count(struct iov_iter
*i
)
325 * "descriptor" for what we're up to with a read.
326 * This allows us to use the same read code yet
327 * have multiple different users of the data that
328 * we read from a file.
330 * The simplest case just copies the data to user
343 typedef int (*read_actor_t
)(read_descriptor_t
*, struct page
*,
344 unsigned long, unsigned long);
346 struct address_space_operations
{
347 int (*writepage
)(struct page
*page
, struct writeback_control
*wbc
);
348 int (*readpage
)(struct file
*, 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
);
359 int (*write_begin
)(struct file
*, struct address_space
*mapping
,
360 loff_t pos
, unsigned len
, unsigned flags
,
361 struct page
**pagep
, void **fsdata
);
362 int (*write_end
)(struct file
*, struct address_space
*mapping
,
363 loff_t pos
, unsigned len
, unsigned copied
,
364 struct page
*page
, void *fsdata
);
366 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
367 sector_t (*bmap
)(struct address_space
*, sector_t
);
368 void (*invalidatepage
) (struct page
*, unsigned int, unsigned int);
369 int (*releasepage
) (struct page
*, gfp_t
);
370 void (*freepage
)(struct page
*);
371 ssize_t (*direct_IO
)(int, struct kiocb
*, const struct iovec
*iov
,
372 loff_t offset
, unsigned long nr_segs
);
373 int (*get_xip_mem
)(struct address_space
*, pgoff_t
, int,
374 void **, unsigned long *);
376 * migrate the contents of a page to the specified target. If
377 * migrate_mode is MIGRATE_ASYNC, it must not block.
379 int (*migratepage
) (struct address_space
*,
380 struct page
*, struct page
*, enum migrate_mode
);
381 int (*launder_page
) (struct page
*);
382 int (*is_partially_uptodate
) (struct page
*, read_descriptor_t
*,
384 void (*is_dirty_writeback
) (struct page
*, bool *, bool *);
385 int (*error_remove_page
)(struct address_space
*, struct page
*);
387 /* swapfile support */
388 int (*swap_activate
)(struct swap_info_struct
*sis
, struct file
*file
,
390 void (*swap_deactivate
)(struct file
*file
);
393 extern const struct address_space_operations empty_aops
;
396 * pagecache_write_begin/pagecache_write_end must be used by general code
397 * to write into the pagecache.
399 int pagecache_write_begin(struct file
*, struct address_space
*mapping
,
400 loff_t pos
, unsigned len
, unsigned flags
,
401 struct page
**pagep
, void **fsdata
);
403 int pagecache_write_end(struct file
*, struct address_space
*mapping
,
404 loff_t pos
, unsigned len
, unsigned copied
,
405 struct page
*page
, void *fsdata
);
407 struct backing_dev_info
;
408 struct address_space
{
409 struct inode
*host
; /* owner: inode, block_device */
410 struct radix_tree_root page_tree
; /* radix tree of all pages */
411 spinlock_t tree_lock
; /* and lock protecting it */
412 unsigned int i_mmap_writable
;/* count VM_SHARED mappings */
413 struct rb_root i_mmap
; /* tree of private and shared mappings */
414 struct list_head i_mmap_nonlinear
;/*list VM_NONLINEAR mappings */
415 struct mutex i_mmap_mutex
; /* protect tree, count, list */
416 /* Protected by tree_lock together with the radix tree */
417 unsigned long nrpages
; /* number of total pages */
418 pgoff_t writeback_index
;/* writeback starts here */
419 const struct address_space_operations
*a_ops
; /* methods */
420 unsigned long flags
; /* error bits/gfp mask */
421 struct backing_dev_info
*backing_dev_info
; /* device readahead, etc */
422 spinlock_t private_lock
; /* for use by the address_space */
423 struct list_head private_list
; /* ditto */
424 void *private_data
; /* ditto */
425 } __attribute__((aligned(sizeof(long))));
427 * On most architectures that alignment is already the case; but
428 * must be enforced here for CRIS, to let the least significant bit
429 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
431 struct request_queue
;
433 struct block_device
{
434 dev_t bd_dev
; /* not a kdev_t - it's a search key */
436 struct inode
* bd_inode
; /* will die */
437 struct super_block
* bd_super
;
438 struct mutex bd_mutex
; /* open/close mutex */
439 struct list_head bd_inodes
;
443 bool bd_write_holder
;
445 struct list_head bd_holder_disks
;
447 struct block_device
* bd_contains
;
448 unsigned bd_block_size
;
449 struct hd_struct
* bd_part
;
450 /* number of times partitions within this device have been opened. */
451 unsigned bd_part_count
;
453 struct gendisk
* bd_disk
;
454 struct request_queue
* bd_queue
;
455 struct list_head bd_list
;
457 * Private data. You must have bd_claim'ed the block_device
458 * to use this. NOTE: bd_claim allows an owner to claim
459 * the same device multiple times, the owner must take special
460 * care to not mess up bd_private for that case.
462 unsigned long bd_private
;
464 /* The counter of freeze processes */
465 int bd_fsfreeze_count
;
466 /* Mutex for freeze */
467 struct mutex bd_fsfreeze_mutex
;
471 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
474 #define PAGECACHE_TAG_DIRTY 0
475 #define PAGECACHE_TAG_WRITEBACK 1
476 #define PAGECACHE_TAG_TOWRITE 2
478 int mapping_tagged(struct address_space
*mapping
, int tag
);
481 * Might pages of this file be mapped into userspace?
483 static inline int mapping_mapped(struct address_space
*mapping
)
485 return !RB_EMPTY_ROOT(&mapping
->i_mmap
) ||
486 !list_empty(&mapping
->i_mmap_nonlinear
);
490 * Might pages of this file have been modified in userspace?
491 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
492 * marks vma as VM_SHARED if it is shared, and the file was opened for
493 * writing i.e. vma may be mprotected writable even if now readonly.
495 static inline int mapping_writably_mapped(struct address_space
*mapping
)
497 return mapping
->i_mmap_writable
!= 0;
501 * Use sequence counter to get consistent i_size on 32-bit processors.
503 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
504 #include <linux/seqlock.h>
505 #define __NEED_I_SIZE_ORDERED
506 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
508 #define i_size_ordered_init(inode) do { } while (0)
512 #define ACL_NOT_CACHED ((void *)(-1))
514 #define IOP_FASTPERM 0x0001
515 #define IOP_LOOKUP 0x0002
516 #define IOP_NOFOLLOW 0x0004
519 * Keep mostly read-only and often accessed (especially for
520 * the RCU path lookup and 'stat' data) fields at the beginning
521 * of the 'struct inode'
525 unsigned short i_opflags
;
528 unsigned int i_flags
;
530 #ifdef CONFIG_FS_POSIX_ACL
531 struct posix_acl
*i_acl
;
532 struct posix_acl
*i_default_acl
;
535 const struct inode_operations
*i_op
;
536 struct super_block
*i_sb
;
537 struct address_space
*i_mapping
;
539 #ifdef CONFIG_SECURITY
543 /* Stat data, not accessed from path walking */
546 * Filesystems may only read i_nlink directly. They shall use the
547 * following functions for modification:
549 * (set|clear|inc|drop)_nlink
550 * inode_(inc|dec)_link_count
553 const unsigned int i_nlink
;
554 unsigned int __i_nlink
;
558 struct timespec i_atime
;
559 struct timespec i_mtime
;
560 struct timespec i_ctime
;
561 spinlock_t i_lock
; /* i_blocks, i_bytes, maybe i_size */
562 unsigned short i_bytes
;
563 unsigned int i_blkbits
;
566 #ifdef __NEED_I_SIZE_ORDERED
567 seqcount_t i_size_seqcount
;
571 unsigned long i_state
;
572 struct mutex i_mutex
;
574 unsigned long dirtied_when
; /* jiffies of first dirtying */
576 struct hlist_node i_hash
;
577 struct list_head i_wb_list
; /* backing dev IO list */
578 struct list_head i_lru
; /* inode LRU list */
579 struct list_head i_sb_list
;
581 struct hlist_head i_dentry
;
582 struct rcu_head i_rcu
;
586 atomic_t i_dio_count
;
587 atomic_t i_writecount
;
588 const struct file_operations
*i_fop
; /* former ->i_op->default_file_ops */
589 struct file_lock
*i_flock
;
590 struct address_space i_data
;
592 struct dquot
*i_dquot
[MAXQUOTAS
];
594 struct list_head i_devices
;
596 struct pipe_inode_info
*i_pipe
;
597 struct block_device
*i_bdev
;
603 #ifdef CONFIG_FSNOTIFY
604 __u32 i_fsnotify_mask
; /* all events this inode cares about */
605 struct hlist_head i_fsnotify_marks
;
609 atomic_t i_readcount
; /* struct files open RO */
611 void *i_private
; /* fs or device private pointer */
614 static inline int inode_unhashed(struct inode
*inode
)
616 return hlist_unhashed(&inode
->i_hash
);
620 * inode->i_mutex nesting subclasses for the lock validator:
622 * 0: the object of the current VFS operation
627 * The locking order between these classes is
628 * parent -> child -> normal -> xattr -> quota
630 enum inode_i_mutex_lock_class
640 * NOTE: in a 32bit arch with a preemptable kernel and
641 * an UP compile the i_size_read/write must be atomic
642 * with respect to the local cpu (unlike with preempt disabled),
643 * but they don't need to be atomic with respect to other cpus like in
644 * true SMP (so they need either to either locally disable irq around
645 * the read or for example on x86 they can be still implemented as a
646 * cmpxchg8b without the need of the lock prefix). For SMP compiles
647 * and 64bit archs it makes no difference if preempt is enabled or not.
649 static inline loff_t
i_size_read(const struct inode
*inode
)
651 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
656 seq
= read_seqcount_begin(&inode
->i_size_seqcount
);
657 i_size
= inode
->i_size
;
658 } while (read_seqcount_retry(&inode
->i_size_seqcount
, seq
));
660 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
664 i_size
= inode
->i_size
;
668 return inode
->i_size
;
673 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
674 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
675 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
677 static inline void i_size_write(struct inode
*inode
, loff_t i_size
)
679 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
681 write_seqcount_begin(&inode
->i_size_seqcount
);
682 inode
->i_size
= i_size
;
683 write_seqcount_end(&inode
->i_size_seqcount
);
685 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
687 inode
->i_size
= i_size
;
690 inode
->i_size
= i_size
;
694 /* Helper functions so that in most cases filesystems will
695 * not need to deal directly with kuid_t and kgid_t and can
696 * instead deal with the raw numeric values that are stored
699 static inline uid_t
i_uid_read(const struct inode
*inode
)
701 return from_kuid(&init_user_ns
, inode
->i_uid
);
704 static inline gid_t
i_gid_read(const struct inode
*inode
)
706 return from_kgid(&init_user_ns
, inode
->i_gid
);
709 static inline void i_uid_write(struct inode
*inode
, uid_t uid
)
711 inode
->i_uid
= make_kuid(&init_user_ns
, uid
);
714 static inline void i_gid_write(struct inode
*inode
, gid_t gid
)
716 inode
->i_gid
= make_kgid(&init_user_ns
, gid
);
719 static inline unsigned iminor(const struct inode
*inode
)
721 return MINOR(inode
->i_rdev
);
724 static inline unsigned imajor(const struct inode
*inode
)
726 return MAJOR(inode
->i_rdev
);
729 extern struct block_device
*I_BDEV(struct inode
*inode
);
732 rwlock_t lock
; /* protects pid, uid, euid fields */
733 struct pid
*pid
; /* pid or -pgrp where SIGIO should be sent */
734 enum pid_type pid_type
; /* Kind of process group SIGIO should be sent to */
735 kuid_t uid
, euid
; /* uid/euid of process setting the owner */
736 int signum
; /* posix.1b rt signal to be delivered on IO */
740 * Track a single file's readahead state
742 struct file_ra_state
{
743 pgoff_t start
; /* where readahead started */
744 unsigned int size
; /* # of readahead pages */
745 unsigned int async_size
; /* do asynchronous readahead when
746 there are only # of pages ahead */
748 unsigned int ra_pages
; /* Maximum readahead window */
749 unsigned int mmap_miss
; /* Cache miss stat for mmap accesses */
750 loff_t prev_pos
; /* Cache last read() position */
754 * Check if @index falls in the readahead windows.
756 static inline int ra_has_index(struct file_ra_state
*ra
, pgoff_t index
)
758 return (index
>= ra
->start
&&
759 index
< ra
->start
+ ra
->size
);
762 #define FILE_MNT_WRITE_TAKEN 1
763 #define FILE_MNT_WRITE_RELEASED 2
767 * fu_list becomes invalid after file_free is called and queued via
768 * fu_rcuhead for RCU freeing
771 struct list_head fu_list
;
772 struct llist_node fu_llist
;
773 struct rcu_head fu_rcuhead
;
776 #define f_dentry f_path.dentry
777 struct inode
*f_inode
; /* cached value */
778 const struct file_operations
*f_op
;
781 * Protects f_ep_links, f_flags, f_pos vs i_size in lseek SEEK_CUR.
782 * Must not be taken from IRQ context.
788 atomic_long_t f_count
;
789 unsigned int f_flags
;
792 struct fown_struct f_owner
;
793 const struct cred
*f_cred
;
794 struct file_ra_state f_ra
;
797 #ifdef CONFIG_SECURITY
800 /* needed for tty driver, and maybe others */
804 /* Used by fs/eventpoll.c to link all the hooks to this file */
805 struct list_head f_ep_links
;
806 struct list_head f_tfile_llink
;
807 #endif /* #ifdef CONFIG_EPOLL */
808 struct address_space
*f_mapping
;
809 #ifdef CONFIG_DEBUG_WRITECOUNT
810 unsigned long f_mnt_write_state
;
817 /* file identifier */
818 unsigned char f_handle
[0];
821 static inline struct file
*get_file(struct file
*f
)
823 atomic_long_inc(&f
->f_count
);
826 #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
827 #define file_count(x) atomic_long_read(&(x)->f_count)
829 #ifdef CONFIG_DEBUG_WRITECOUNT
830 static inline void file_take_write(struct file
*f
)
832 WARN_ON(f
->f_mnt_write_state
!= 0);
833 f
->f_mnt_write_state
= FILE_MNT_WRITE_TAKEN
;
835 static inline void file_release_write(struct file
*f
)
837 f
->f_mnt_write_state
|= FILE_MNT_WRITE_RELEASED
;
839 static inline void file_reset_write(struct file
*f
)
841 f
->f_mnt_write_state
= 0;
843 static inline void file_check_state(struct file
*f
)
846 * At this point, either both or neither of these bits
849 WARN_ON(f
->f_mnt_write_state
== FILE_MNT_WRITE_TAKEN
);
850 WARN_ON(f
->f_mnt_write_state
== FILE_MNT_WRITE_RELEASED
);
852 static inline int file_check_writeable(struct file
*f
)
854 if (f
->f_mnt_write_state
== FILE_MNT_WRITE_TAKEN
)
856 printk(KERN_WARNING
"writeable file with no "
857 "mnt_want_write()\n");
861 #else /* !CONFIG_DEBUG_WRITECOUNT */
862 static inline void file_take_write(struct file
*filp
) {}
863 static inline void file_release_write(struct file
*filp
) {}
864 static inline void file_reset_write(struct file
*filp
) {}
865 static inline void file_check_state(struct file
*filp
) {}
866 static inline int file_check_writeable(struct file
*filp
)
870 #endif /* CONFIG_DEBUG_WRITECOUNT */
872 #define MAX_NON_LFS ((1UL<<31) - 1)
874 /* Page cache limit. The filesystems should put that into their s_maxbytes
875 limits, otherwise bad things can happen in VM. */
876 #if BITS_PER_LONG==32
877 #define MAX_LFS_FILESIZE (((loff_t)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
878 #elif BITS_PER_LONG==64
879 #define MAX_LFS_FILESIZE ((loff_t)0x7fffffffffffffffLL)
884 #define FL_ACCESS 8 /* not trying to lock, just looking */
885 #define FL_EXISTS 16 /* when unlocking, test for existence */
886 #define FL_LEASE 32 /* lease held on this file */
887 #define FL_CLOSE 64 /* unlock on close */
888 #define FL_SLEEP 128 /* A blocking lock */
889 #define FL_DOWNGRADE_PENDING 256 /* Lease is being downgraded */
890 #define FL_UNLOCK_PENDING 512 /* Lease is being broken */
893 * Special return value from posix_lock_file() and vfs_lock_file() for
894 * asynchronous locking.
896 #define FILE_LOCK_DEFERRED 1
899 * The POSIX file lock owner is determined by
900 * the "struct files_struct" in the thread group
901 * (or NULL for no owner - BSD locks).
903 * Lockd stuffs a "host" pointer into this.
905 typedef struct files_struct
*fl_owner_t
;
907 struct file_lock_operations
{
908 void (*fl_copy_lock
)(struct file_lock
*, struct file_lock
*);
909 void (*fl_release_private
)(struct file_lock
*);
912 struct lock_manager_operations
{
913 int (*lm_compare_owner
)(struct file_lock
*, struct file_lock
*);
914 unsigned long (*lm_owner_key
)(struct file_lock
*);
915 void (*lm_notify
)(struct file_lock
*); /* unblock callback */
916 int (*lm_grant
)(struct file_lock
*, struct file_lock
*, int);
917 void (*lm_break
)(struct file_lock
*);
918 int (*lm_change
)(struct file_lock
**, int);
921 struct lock_manager
{
922 struct list_head list
;
926 void locks_start_grace(struct net
*, struct lock_manager
*);
927 void locks_end_grace(struct lock_manager
*);
928 int locks_in_grace(struct net
*);
930 /* that will die - we need it for nfs_lock_info */
931 #include <linux/nfs_fs_i.h>
934 * struct file_lock represents a generic "file lock". It's used to represent
935 * POSIX byte range locks, BSD (flock) locks, and leases. It's important to
936 * note that the same struct is used to represent both a request for a lock and
937 * the lock itself, but the same object is never used for both.
939 * FIXME: should we create a separate "struct lock_request" to help distinguish
942 * The i_flock list is ordered by:
944 * 1) lock type -- FL_LEASEs first, then FL_FLOCK, and finally FL_POSIX
946 * 3) lock range start
949 * Obviously, the last two criteria only matter for POSIX locks.
952 struct file_lock
*fl_next
; /* singly linked list for this inode */
953 struct hlist_node fl_link
; /* node in global lists */
954 struct list_head fl_block
; /* circular list of blocked processes */
956 unsigned int fl_flags
;
957 unsigned char fl_type
;
959 int fl_link_cpu
; /* what cpu's list is this on? */
960 struct pid
*fl_nspid
;
961 wait_queue_head_t fl_wait
;
962 struct file
*fl_file
;
966 struct fasync_struct
* fl_fasync
; /* for lease break notifications */
967 /* for lease breaks: */
968 unsigned long fl_break_time
;
969 unsigned long fl_downgrade_time
;
971 const struct file_lock_operations
*fl_ops
; /* Callbacks for filesystems */
972 const struct lock_manager_operations
*fl_lmops
; /* Callbacks for lockmanagers */
974 struct nfs_lock_info nfs_fl
;
975 struct nfs4_lock_info nfs4_fl
;
977 struct list_head link
; /* link in AFS vnode's pending_locks list */
978 int state
; /* state of grant or error if -ve */
983 /* The following constant reflects the upper bound of the file/locking space */
985 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
986 #define OFFSET_MAX INT_LIMIT(loff_t)
987 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
990 #include <linux/fcntl.h>
992 extern void send_sigio(struct fown_struct
*fown
, int fd
, int band
);
994 #ifdef CONFIG_FILE_LOCKING
995 extern int fcntl_getlk(struct file
*, struct flock __user
*);
996 extern int fcntl_setlk(unsigned int, struct file
*, unsigned int,
997 struct flock __user
*);
999 #if BITS_PER_LONG == 32
1000 extern int fcntl_getlk64(struct file
*, struct flock64 __user
*);
1001 extern int fcntl_setlk64(unsigned int, struct file
*, unsigned int,
1002 struct flock64 __user
*);
1005 extern int fcntl_setlease(unsigned int fd
, struct file
*filp
, long arg
);
1006 extern int fcntl_getlease(struct file
*filp
);
1009 void locks_free_lock(struct file_lock
*fl
);
1010 extern void locks_init_lock(struct file_lock
*);
1011 extern struct file_lock
* locks_alloc_lock(void);
1012 extern void locks_copy_lock(struct file_lock
*, struct file_lock
*);
1013 extern void __locks_copy_lock(struct file_lock
*, const struct file_lock
*);
1014 extern void locks_remove_posix(struct file
*, fl_owner_t
);
1015 extern void locks_remove_flock(struct file
*);
1016 extern void locks_release_private(struct file_lock
*);
1017 extern void posix_test_lock(struct file
*, struct file_lock
*);
1018 extern int posix_lock_file(struct file
*, struct file_lock
*, struct file_lock
*);
1019 extern int posix_lock_file_wait(struct file
*, struct file_lock
*);
1020 extern int posix_unblock_lock(struct file_lock
*);
1021 extern int vfs_test_lock(struct file
*, struct file_lock
*);
1022 extern int vfs_lock_file(struct file
*, unsigned int, struct file_lock
*, struct file_lock
*);
1023 extern int vfs_cancel_lock(struct file
*filp
, struct file_lock
*fl
);
1024 extern int flock_lock_file_wait(struct file
*filp
, struct file_lock
*fl
);
1025 extern int __break_lease(struct inode
*inode
, unsigned int flags
);
1026 extern void lease_get_mtime(struct inode
*, struct timespec
*time
);
1027 extern int generic_setlease(struct file
*, long, struct file_lock
**);
1028 extern int vfs_setlease(struct file
*, long, struct file_lock
**);
1029 extern int lease_modify(struct file_lock
**, int);
1030 extern int lock_may_read(struct inode
*, loff_t start
, unsigned long count
);
1031 extern int lock_may_write(struct inode
*, loff_t start
, unsigned long count
);
1032 #else /* !CONFIG_FILE_LOCKING */
1033 static inline int fcntl_getlk(struct file
*file
, struct flock __user
*user
)
1038 static inline int fcntl_setlk(unsigned int fd
, struct file
*file
,
1039 unsigned int cmd
, struct flock __user
*user
)
1044 #if BITS_PER_LONG == 32
1045 static inline int fcntl_getlk64(struct file
*file
, struct flock64 __user
*user
)
1050 static inline int fcntl_setlk64(unsigned int fd
, struct file
*file
,
1051 unsigned int cmd
, struct flock64 __user
*user
)
1056 static inline int fcntl_setlease(unsigned int fd
, struct file
*filp
, long arg
)
1061 static inline int fcntl_getlease(struct file
*filp
)
1066 static inline void locks_init_lock(struct file_lock
*fl
)
1071 static inline void __locks_copy_lock(struct file_lock
*new, struct file_lock
*fl
)
1076 static inline void locks_copy_lock(struct file_lock
*new, struct file_lock
*fl
)
1081 static inline void locks_remove_posix(struct file
*filp
, fl_owner_t owner
)
1086 static inline void locks_remove_flock(struct file
*filp
)
1091 static inline void posix_test_lock(struct file
*filp
, struct file_lock
*fl
)
1096 static inline int posix_lock_file(struct file
*filp
, struct file_lock
*fl
,
1097 struct file_lock
*conflock
)
1102 static inline int posix_lock_file_wait(struct file
*filp
, struct file_lock
*fl
)
1107 static inline int posix_unblock_lock(struct file_lock
*waiter
)
1112 static inline int vfs_test_lock(struct file
*filp
, struct file_lock
*fl
)
1117 static inline int vfs_lock_file(struct file
*filp
, unsigned int cmd
,
1118 struct file_lock
*fl
, struct file_lock
*conf
)
1123 static inline int vfs_cancel_lock(struct file
*filp
, struct file_lock
*fl
)
1128 static inline int flock_lock_file_wait(struct file
*filp
,
1129 struct file_lock
*request
)
1134 static inline int __break_lease(struct inode
*inode
, unsigned int mode
)
1139 static inline void lease_get_mtime(struct inode
*inode
, struct timespec
*time
)
1144 static inline int generic_setlease(struct file
*filp
, long arg
,
1145 struct file_lock
**flp
)
1150 static inline int vfs_setlease(struct file
*filp
, long arg
,
1151 struct file_lock
**lease
)
1156 static inline int lease_modify(struct file_lock
**before
, int arg
)
1161 static inline int lock_may_read(struct inode
*inode
, loff_t start
,
1167 static inline int lock_may_write(struct inode
*inode
, loff_t start
,
1172 #endif /* !CONFIG_FILE_LOCKING */
1175 struct fasync_struct
{
1179 struct fasync_struct
*fa_next
; /* singly linked list */
1180 struct file
*fa_file
;
1181 struct rcu_head fa_rcu
;
1184 #define FASYNC_MAGIC 0x4601
1186 /* SMP safe fasync helpers: */
1187 extern int fasync_helper(int, struct file
*, int, struct fasync_struct
**);
1188 extern struct fasync_struct
*fasync_insert_entry(int, struct file
*, struct fasync_struct
**, struct fasync_struct
*);
1189 extern int fasync_remove_entry(struct file
*, struct fasync_struct
**);
1190 extern struct fasync_struct
*fasync_alloc(void);
1191 extern void fasync_free(struct fasync_struct
*);
1193 /* can be called from interrupts */
1194 extern void kill_fasync(struct fasync_struct
**, int, int);
1196 extern int __f_setown(struct file
*filp
, struct pid
*, enum pid_type
, int force
);
1197 extern int f_setown(struct file
*filp
, unsigned long arg
, int force
);
1198 extern void f_delown(struct file
*filp
);
1199 extern pid_t
f_getown(struct file
*filp
);
1200 extern int send_sigurg(struct fown_struct
*fown
);
1208 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1209 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1210 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1211 #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1212 #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
1214 extern struct list_head super_blocks
;
1215 extern spinlock_t sb_lock
;
1217 /* Possible states of 'frozen' field */
1219 SB_UNFROZEN
= 0, /* FS is unfrozen */
1220 SB_FREEZE_WRITE
= 1, /* Writes, dir ops, ioctls frozen */
1221 SB_FREEZE_PAGEFAULT
= 2, /* Page faults stopped as well */
1222 SB_FREEZE_FS
= 3, /* For internal FS use (e.g. to stop
1223 * internal threads if needed) */
1224 SB_FREEZE_COMPLETE
= 4, /* ->freeze_fs finished successfully */
1227 #define SB_FREEZE_LEVELS (SB_FREEZE_COMPLETE - 1)
1230 /* Counters for counting writers at each level */
1231 struct percpu_counter counter
[SB_FREEZE_LEVELS
];
1232 wait_queue_head_t wait
; /* queue for waiting for
1233 writers / faults to finish */
1234 int frozen
; /* Is sb frozen? */
1235 wait_queue_head_t wait_unfrozen
; /* queue for waiting for
1237 #ifdef CONFIG_DEBUG_LOCK_ALLOC
1238 struct lockdep_map lock_map
[SB_FREEZE_LEVELS
];
1242 struct super_block
{
1243 struct list_head s_list
; /* Keep this first */
1244 dev_t s_dev
; /* search index; _not_ kdev_t */
1245 unsigned char s_blocksize_bits
;
1246 unsigned long s_blocksize
;
1247 loff_t s_maxbytes
; /* Max file size */
1248 struct file_system_type
*s_type
;
1249 const struct super_operations
*s_op
;
1250 const struct dquot_operations
*dq_op
;
1251 const struct quotactl_ops
*s_qcop
;
1252 const struct export_operations
*s_export_op
;
1253 unsigned long s_flags
;
1254 unsigned long s_magic
;
1255 struct dentry
*s_root
;
1256 struct rw_semaphore s_umount
;
1259 #ifdef CONFIG_SECURITY
1262 const struct xattr_handler
**s_xattr
;
1264 struct list_head s_inodes
; /* all inodes */
1265 struct hlist_bl_head s_anon
; /* anonymous dentries for (nfs) exporting */
1267 struct list_head __percpu
*s_files
;
1269 struct list_head s_files
;
1271 struct list_head s_mounts
; /* list of mounts; _not_ for fs use */
1273 /* s_dentry_lru_lock protects s_dentry_lru and s_nr_dentry_unused */
1274 spinlock_t s_dentry_lru_lock ____cacheline_aligned_in_smp
;
1275 struct list_head s_dentry_lru
; /* unused dentry lru */
1276 long s_nr_dentry_unused
; /* # of dentry on lru */
1278 /* s_inode_lru_lock protects s_inode_lru and s_nr_inodes_unused */
1279 spinlock_t s_inode_lru_lock ____cacheline_aligned_in_smp
;
1280 struct list_head s_inode_lru
; /* unused inode lru */
1281 long s_nr_inodes_unused
; /* # of inodes on lru */
1283 struct block_device
*s_bdev
;
1284 struct backing_dev_info
*s_bdi
;
1285 struct mtd_info
*s_mtd
;
1286 struct hlist_node s_instances
;
1287 struct quota_info s_dquot
; /* Diskquota specific options */
1289 struct sb_writers s_writers
;
1291 char s_id
[32]; /* Informational name */
1292 u8 s_uuid
[16]; /* UUID */
1294 void *s_fs_info
; /* Filesystem private info */
1295 unsigned int s_max_links
;
1298 /* Granularity of c/m/atime in ns.
1299 Cannot be worse than a second */
1303 * The next field is for VFS *only*. No filesystems have any business
1304 * even looking at it. You had been warned.
1306 struct mutex s_vfs_rename_mutex
; /* Kludge */
1309 * Filesystem subtype. If non-empty the filesystem type field
1310 * in /proc/mounts will be "type.subtype"
1315 * Saved mount options for lazy filesystems using
1316 * generic_show_options()
1318 char __rcu
*s_options
;
1319 const struct dentry_operations
*s_d_op
; /* default d_op for dentries */
1322 * Saved pool identifier for cleancache (-1 means none)
1324 int cleancache_poolid
;
1326 struct shrinker s_shrink
; /* per-sb shrinker handle */
1328 /* Number of inodes with nlink == 0 but still referenced */
1329 atomic_long_t s_remove_count
;
1331 /* Being remounted read-only */
1332 int s_readonly_remount
;
1334 /* AIO completions deferred from interrupt context */
1335 struct workqueue_struct
*s_dio_done_wq
;
1338 extern struct timespec
current_fs_time(struct super_block
*sb
);
1341 * Snapshotting support.
1344 void __sb_end_write(struct super_block
*sb
, int level
);
1345 int __sb_start_write(struct super_block
*sb
, int level
, bool wait
);
1348 * sb_end_write - drop write access to a superblock
1349 * @sb: the super we wrote to
1351 * Decrement number of writers to the filesystem. Wake up possible waiters
1352 * wanting to freeze the filesystem.
1354 static inline void sb_end_write(struct super_block
*sb
)
1356 __sb_end_write(sb
, SB_FREEZE_WRITE
);
1360 * sb_end_pagefault - drop write access to a superblock from a page fault
1361 * @sb: the super we wrote to
1363 * Decrement number of processes handling write page fault to the filesystem.
1364 * Wake up possible waiters wanting to freeze the filesystem.
1366 static inline void sb_end_pagefault(struct super_block
*sb
)
1368 __sb_end_write(sb
, SB_FREEZE_PAGEFAULT
);
1372 * sb_end_intwrite - drop write access to a superblock for internal fs purposes
1373 * @sb: the super we wrote to
1375 * Decrement fs-internal number of writers to the filesystem. Wake up possible
1376 * waiters wanting to freeze the filesystem.
1378 static inline void sb_end_intwrite(struct super_block
*sb
)
1380 __sb_end_write(sb
, SB_FREEZE_FS
);
1384 * sb_start_write - get write access to a superblock
1385 * @sb: the super we write to
1387 * When a process wants to write data or metadata to a file system (i.e. dirty
1388 * a page or an inode), it should embed the operation in a sb_start_write() -
1389 * sb_end_write() pair to get exclusion against file system freezing. This
1390 * function increments number of writers preventing freezing. If the file
1391 * system is already frozen, the function waits until the file system is
1394 * Since freeze protection behaves as a lock, users have to preserve
1395 * ordering of freeze protection and other filesystem locks. Generally,
1396 * freeze protection should be the outermost lock. In particular, we have:
1399 * -> i_mutex (write path, truncate, directory ops, ...)
1400 * -> s_umount (freeze_super, thaw_super)
1402 static inline void sb_start_write(struct super_block
*sb
)
1404 __sb_start_write(sb
, SB_FREEZE_WRITE
, true);
1407 static inline int sb_start_write_trylock(struct super_block
*sb
)
1409 return __sb_start_write(sb
, SB_FREEZE_WRITE
, false);
1413 * sb_start_pagefault - get write access to a superblock from a page fault
1414 * @sb: the super we write to
1416 * When a process starts handling write page fault, it should embed the
1417 * operation into sb_start_pagefault() - sb_end_pagefault() pair to get
1418 * exclusion against file system freezing. This is needed since the page fault
1419 * is going to dirty a page. This function increments number of running page
1420 * faults preventing freezing. If the file system is already frozen, the
1421 * function waits until the file system is thawed.
1423 * Since page fault freeze protection behaves as a lock, users have to preserve
1424 * ordering of freeze protection and other filesystem locks. It is advised to
1425 * put sb_start_pagefault() close to mmap_sem in lock ordering. Page fault
1426 * handling code implies lock dependency:
1429 * -> sb_start_pagefault
1431 static inline void sb_start_pagefault(struct super_block
*sb
)
1433 __sb_start_write(sb
, SB_FREEZE_PAGEFAULT
, true);
1437 * sb_start_intwrite - get write access to a superblock for internal fs purposes
1438 * @sb: the super we write to
1440 * This is the third level of protection against filesystem freezing. It is
1441 * free for use by a filesystem. The only requirement is that it must rank
1442 * below sb_start_pagefault.
1444 * For example filesystem can call sb_start_intwrite() when starting a
1445 * transaction which somewhat eases handling of freezing for internal sources
1446 * of filesystem changes (internal fs threads, discarding preallocation on file
1449 static inline void sb_start_intwrite(struct super_block
*sb
)
1451 __sb_start_write(sb
, SB_FREEZE_FS
, true);
1455 extern bool inode_owner_or_capable(const struct inode
*inode
);
1458 * VFS helper functions..
1460 extern int vfs_create(struct inode
*, struct dentry
*, umode_t
, bool);
1461 extern int vfs_mkdir(struct inode
*, struct dentry
*, umode_t
);
1462 extern int vfs_mknod(struct inode
*, struct dentry
*, umode_t
, dev_t
);
1463 extern int vfs_symlink(struct inode
*, struct dentry
*, const char *);
1464 extern int vfs_link(struct dentry
*, struct inode
*, struct dentry
*);
1465 extern int vfs_rmdir(struct inode
*, struct dentry
*);
1466 extern int vfs_unlink(struct inode
*, struct dentry
*);
1467 extern int vfs_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
1470 * VFS dentry helper functions.
1472 extern void dentry_unhash(struct dentry
*dentry
);
1475 * VFS file helper functions.
1477 extern void inode_init_owner(struct inode
*inode
, const struct inode
*dir
,
1480 * VFS FS_IOC_FIEMAP helper definitions.
1482 struct fiemap_extent_info
{
1483 unsigned int fi_flags
; /* Flags as passed from user */
1484 unsigned int fi_extents_mapped
; /* Number of mapped extents */
1485 unsigned int fi_extents_max
; /* Size of fiemap_extent array */
1486 struct fiemap_extent __user
*fi_extents_start
; /* Start of
1487 fiemap_extent array */
1489 int fiemap_fill_next_extent(struct fiemap_extent_info
*info
, u64 logical
,
1490 u64 phys
, u64 len
, u32 flags
);
1491 int fiemap_check_flags(struct fiemap_extent_info
*fieinfo
, u32 fs_flags
);
1496 * NOTE! These match bits 12..15 of stat.st_mode
1497 * (ie "(i_mode >> 12) & 15").
1499 #define DT_UNKNOWN 0
1510 * This is the "filldir" function type, used by readdir() to let
1511 * the kernel specify what kind of dirent layout it wants to have.
1512 * This allows the kernel to read directories into kernel space or
1513 * to have different dirent layouts depending on the binary type.
1515 typedef int (*filldir_t
)(void *, const char *, int, loff_t
, u64
, unsigned);
1516 struct dir_context
{
1517 const filldir_t actor
;
1521 struct block_device_operations
;
1523 /* These macros are for out of kernel modules to test that
1524 * the kernel supports the unlocked_ioctl and compat_ioctl
1525 * fields in struct file_operations. */
1526 #define HAVE_COMPAT_IOCTL 1
1527 #define HAVE_UNLOCKED_IOCTL 1
1529 struct file_operations
{
1530 struct module
*owner
;
1531 loff_t (*llseek
) (struct file
*, loff_t
, int);
1532 ssize_t (*read
) (struct file
*, char __user
*, size_t, loff_t
*);
1533 ssize_t (*write
) (struct file
*, const char __user
*, size_t, loff_t
*);
1534 ssize_t (*aio_read
) (struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1535 ssize_t (*aio_write
) (struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1536 int (*iterate
) (struct file
*, struct dir_context
*);
1537 unsigned int (*poll
) (struct file
*, struct poll_table_struct
*);
1538 long (*unlocked_ioctl
) (struct file
*, unsigned int, unsigned long);
1539 long (*compat_ioctl
) (struct file
*, unsigned int, unsigned long);
1540 int (*mmap
) (struct file
*, struct vm_area_struct
*);
1541 int (*open
) (struct inode
*, struct file
*);
1542 int (*flush
) (struct file
*, fl_owner_t id
);
1543 int (*release
) (struct inode
*, struct file
*);
1544 int (*fsync
) (struct file
*, loff_t
, loff_t
, int datasync
);
1545 int (*aio_fsync
) (struct kiocb
*, int datasync
);
1546 int (*fasync
) (int, struct file
*, int);
1547 int (*lock
) (struct file
*, int, struct file_lock
*);
1548 ssize_t (*sendpage
) (struct file
*, struct page
*, int, size_t, loff_t
*, int);
1549 unsigned long (*get_unmapped_area
)(struct file
*, unsigned long, unsigned long, unsigned long, unsigned long);
1550 int (*check_flags
)(int);
1551 int (*flock
) (struct file
*, int, struct file_lock
*);
1552 ssize_t (*splice_write
)(struct pipe_inode_info
*, struct file
*, loff_t
*, size_t, unsigned int);
1553 ssize_t (*splice_read
)(struct file
*, loff_t
*, struct pipe_inode_info
*, size_t, unsigned int);
1554 int (*setlease
)(struct file
*, long, struct file_lock
**);
1555 long (*fallocate
)(struct file
*file
, int mode
, loff_t offset
,
1557 int (*show_fdinfo
)(struct seq_file
*m
, struct file
*f
);
1560 struct inode_operations
{
1561 struct dentry
* (*lookup
) (struct inode
*,struct dentry
*, unsigned int);
1562 void * (*follow_link
) (struct dentry
*, struct nameidata
*);
1563 int (*permission
) (struct inode
*, int);
1564 struct posix_acl
* (*get_acl
)(struct inode
*, int);
1566 int (*readlink
) (struct dentry
*, char __user
*,int);
1567 void (*put_link
) (struct dentry
*, struct nameidata
*, void *);
1569 int (*create
) (struct inode
*,struct dentry
*, umode_t
, bool);
1570 int (*link
) (struct dentry
*,struct inode
*,struct dentry
*);
1571 int (*unlink
) (struct inode
*,struct dentry
*);
1572 int (*symlink
) (struct inode
*,struct dentry
*,const char *);
1573 int (*mkdir
) (struct inode
*,struct dentry
*,umode_t
);
1574 int (*rmdir
) (struct inode
*,struct dentry
*);
1575 int (*mknod
) (struct inode
*,struct dentry
*,umode_t
,dev_t
);
1576 int (*rename
) (struct inode
*, struct dentry
*,
1577 struct inode
*, struct dentry
*);
1578 int (*setattr
) (struct dentry
*, struct iattr
*);
1579 int (*getattr
) (struct vfsmount
*mnt
, struct dentry
*, struct kstat
*);
1580 int (*setxattr
) (struct dentry
*, const char *,const void *,size_t,int);
1581 ssize_t (*getxattr
) (struct dentry
*, const char *, void *, size_t);
1582 ssize_t (*listxattr
) (struct dentry
*, char *, size_t);
1583 int (*removexattr
) (struct dentry
*, const char *);
1584 int (*fiemap
)(struct inode
*, struct fiemap_extent_info
*, u64 start
,
1586 int (*update_time
)(struct inode
*, struct timespec
*, int);
1587 int (*atomic_open
)(struct inode
*, struct dentry
*,
1588 struct file
*, unsigned open_flag
,
1589 umode_t create_mode
, int *opened
);
1590 int (*tmpfile
) (struct inode
*, struct dentry
*, umode_t
);
1591 } ____cacheline_aligned
;
1593 ssize_t
rw_copy_check_uvector(int type
, const struct iovec __user
* uvector
,
1594 unsigned long nr_segs
, unsigned long fast_segs
,
1595 struct iovec
*fast_pointer
,
1596 struct iovec
**ret_pointer
);
1598 extern ssize_t
vfs_read(struct file
*, char __user
*, size_t, loff_t
*);
1599 extern ssize_t
vfs_write(struct file
*, const char __user
*, size_t, loff_t
*);
1600 extern ssize_t
vfs_readv(struct file
*, const struct iovec __user
*,
1601 unsigned long, loff_t
*);
1602 extern ssize_t
vfs_writev(struct file
*, const struct iovec __user
*,
1603 unsigned long, loff_t
*);
1605 struct super_operations
{
1606 struct inode
*(*alloc_inode
)(struct super_block
*sb
);
1607 void (*destroy_inode
)(struct inode
*);
1609 void (*dirty_inode
) (struct inode
*, int flags
);
1610 int (*write_inode
) (struct inode
*, struct writeback_control
*wbc
);
1611 int (*drop_inode
) (struct inode
*);
1612 void (*evict_inode
) (struct inode
*);
1613 void (*put_super
) (struct super_block
*);
1614 int (*sync_fs
)(struct super_block
*sb
, int wait
);
1615 int (*freeze_fs
) (struct super_block
*);
1616 int (*unfreeze_fs
) (struct super_block
*);
1617 int (*statfs
) (struct dentry
*, struct kstatfs
*);
1618 int (*remount_fs
) (struct super_block
*, int *, char *);
1619 void (*umount_begin
) (struct super_block
*);
1621 int (*show_options
)(struct seq_file
*, struct dentry
*);
1622 int (*show_devname
)(struct seq_file
*, struct dentry
*);
1623 int (*show_path
)(struct seq_file
*, struct dentry
*);
1624 int (*show_stats
)(struct seq_file
*, struct dentry
*);
1626 ssize_t (*quota_read
)(struct super_block
*, int, char *, size_t, loff_t
);
1627 ssize_t (*quota_write
)(struct super_block
*, int, const char *, size_t, loff_t
);
1629 int (*bdev_try_to_free_page
)(struct super_block
*, struct page
*, gfp_t
);
1630 long (*nr_cached_objects
)(struct super_block
*);
1631 long (*free_cached_objects
)(struct super_block
*, long);
1635 * Inode flags - they have no relation to superblock flags now
1637 #define S_SYNC 1 /* Writes are synced at once */
1638 #define S_NOATIME 2 /* Do not update access times */
1639 #define S_APPEND 4 /* Append-only file */
1640 #define S_IMMUTABLE 8 /* Immutable file */
1641 #define S_DEAD 16 /* removed, but still open directory */
1642 #define S_NOQUOTA 32 /* Inode is not counted to quota */
1643 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
1644 #define S_NOCMTIME 128 /* Do not update file c/mtime */
1645 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
1646 #define S_PRIVATE 512 /* Inode is fs-internal */
1647 #define S_IMA 1024 /* Inode has an associated IMA struct */
1648 #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
1649 #define S_NOSEC 4096 /* no suid or xattr security attributes */
1652 * Note that nosuid etc flags are inode-specific: setting some file-system
1653 * flags just means all the inodes inherit those flags by default. It might be
1654 * possible to override it selectively if you really wanted to with some
1655 * ioctl() that is not currently implemented.
1657 * Exception: MS_RDONLY is always applied to the entire file system.
1659 * Unfortunately, it is possible to change a filesystems flags with it mounted
1660 * with files in use. This means that all of the inodes will not have their
1661 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
1662 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
1664 #define __IS_FLG(inode, flg) ((inode)->i_sb->s_flags & (flg))
1666 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
1667 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
1668 ((inode)->i_flags & S_SYNC))
1669 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
1670 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
1671 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
1672 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
1673 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
1675 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
1676 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
1677 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
1678 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
1680 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
1681 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
1682 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
1683 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
1684 #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
1685 #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
1686 #define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
1689 * Inode state bits. Protected by inode->i_lock
1691 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1692 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1694 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1695 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1696 * various stages of removing an inode.
1698 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1700 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1701 * fdatasync(). i_atime is the usual cause.
1702 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1703 * these changes separately from I_DIRTY_SYNC so that we
1704 * don't have to write inode on fdatasync() when only
1705 * mtime has changed in it.
1706 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1707 * I_NEW Serves as both a mutex and completion notification.
1708 * New inodes set I_NEW. If two processes both create
1709 * the same inode, one of them will release its inode and
1710 * wait for I_NEW to be released before returning.
1711 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1712 * also cause waiting on I_NEW, without I_NEW actually
1713 * being set. find_inode() uses this to prevent returning
1714 * nearly-dead inodes.
1715 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1716 * is zero. I_FREEING must be set when I_WILL_FREE is
1718 * I_FREEING Set when inode is about to be freed but still has dirty
1719 * pages or buffers attached or the inode itself is still
1721 * I_CLEAR Added by clear_inode(). In this state the inode is
1722 * clean and can be destroyed. Inode keeps I_FREEING.
1724 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1725 * prohibited for many purposes. iget() must wait for
1726 * the inode to be completely released, then create it
1727 * anew. Other functions will just ignore such inodes,
1728 * if appropriate. I_NEW is used for waiting.
1730 * I_SYNC Writeback of inode is running. The bit is set during
1731 * data writeback, and cleared with a wakeup on the bit
1732 * address once it is done. The bit is also used to pin
1733 * the inode in memory for flusher thread.
1735 * I_REFERENCED Marks the inode as recently references on the LRU list.
1737 * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit().
1739 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1741 #define I_DIRTY_SYNC (1 << 0)
1742 #define I_DIRTY_DATASYNC (1 << 1)
1743 #define I_DIRTY_PAGES (1 << 2)
1745 #define I_NEW (1 << __I_NEW)
1746 #define I_WILL_FREE (1 << 4)
1747 #define I_FREEING (1 << 5)
1748 #define I_CLEAR (1 << 6)
1750 #define I_SYNC (1 << __I_SYNC)
1751 #define I_REFERENCED (1 << 8)
1752 #define __I_DIO_WAKEUP 9
1753 #define I_DIO_WAKEUP (1 << I_DIO_WAKEUP)
1754 #define I_LINKABLE (1 << 10)
1756 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1758 extern void __mark_inode_dirty(struct inode
*, int);
1759 static inline void mark_inode_dirty(struct inode
*inode
)
1761 __mark_inode_dirty(inode
, I_DIRTY
);
1764 static inline void mark_inode_dirty_sync(struct inode
*inode
)
1766 __mark_inode_dirty(inode
, I_DIRTY_SYNC
);
1769 extern void inc_nlink(struct inode
*inode
);
1770 extern void drop_nlink(struct inode
*inode
);
1771 extern void clear_nlink(struct inode
*inode
);
1772 extern void set_nlink(struct inode
*inode
, unsigned int nlink
);
1774 static inline void inode_inc_link_count(struct inode
*inode
)
1777 mark_inode_dirty(inode
);
1780 static inline void inode_dec_link_count(struct inode
*inode
)
1783 mark_inode_dirty(inode
);
1787 * inode_inc_iversion - increments i_version
1788 * @inode: inode that need to be updated
1790 * Every time the inode is modified, the i_version field will be incremented.
1791 * The filesystem has to be mounted with i_version flag
1794 static inline void inode_inc_iversion(struct inode
*inode
)
1796 spin_lock(&inode
->i_lock
);
1798 spin_unlock(&inode
->i_lock
);
1801 enum file_time_flags
{
1808 extern void touch_atime(const struct path
*);
1809 static inline void file_accessed(struct file
*file
)
1811 if (!(file
->f_flags
& O_NOATIME
))
1812 touch_atime(&file
->f_path
);
1815 int sync_inode(struct inode
*inode
, struct writeback_control
*wbc
);
1816 int sync_inode_metadata(struct inode
*inode
, int wait
);
1818 struct file_system_type
{
1821 #define FS_REQUIRES_DEV 1
1822 #define FS_BINARY_MOUNTDATA 2
1823 #define FS_HAS_SUBTYPE 4
1824 #define FS_USERNS_MOUNT 8 /* Can be mounted by userns root */
1825 #define FS_USERNS_DEV_MOUNT 16 /* A userns mount does not imply MNT_NODEV */
1826 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() during rename() internally. */
1827 struct dentry
*(*mount
) (struct file_system_type
*, int,
1828 const char *, void *);
1829 void (*kill_sb
) (struct super_block
*);
1830 struct module
*owner
;
1831 struct file_system_type
* next
;
1832 struct hlist_head fs_supers
;
1834 struct lock_class_key s_lock_key
;
1835 struct lock_class_key s_umount_key
;
1836 struct lock_class_key s_vfs_rename_key
;
1837 struct lock_class_key s_writers_key
[SB_FREEZE_LEVELS
];
1839 struct lock_class_key i_lock_key
;
1840 struct lock_class_key i_mutex_key
;
1841 struct lock_class_key i_mutex_dir_key
;
1844 #define MODULE_ALIAS_FS(NAME) MODULE_ALIAS("fs-" NAME)
1846 extern struct dentry
*mount_ns(struct file_system_type
*fs_type
, int flags
,
1847 void *data
, int (*fill_super
)(struct super_block
*, void *, int));
1848 extern struct dentry
*mount_bdev(struct file_system_type
*fs_type
,
1849 int flags
, const char *dev_name
, void *data
,
1850 int (*fill_super
)(struct super_block
*, void *, int));
1851 extern struct dentry
*mount_single(struct file_system_type
*fs_type
,
1852 int flags
, void *data
,
1853 int (*fill_super
)(struct super_block
*, void *, int));
1854 extern struct dentry
*mount_nodev(struct file_system_type
*fs_type
,
1855 int flags
, void *data
,
1856 int (*fill_super
)(struct super_block
*, void *, int));
1857 extern struct dentry
*mount_subtree(struct vfsmount
*mnt
, const char *path
);
1858 void generic_shutdown_super(struct super_block
*sb
);
1859 void kill_block_super(struct super_block
*sb
);
1860 void kill_anon_super(struct super_block
*sb
);
1861 void kill_litter_super(struct super_block
*sb
);
1862 void deactivate_super(struct super_block
*sb
);
1863 void deactivate_locked_super(struct super_block
*sb
);
1864 int set_anon_super(struct super_block
*s
, void *data
);
1865 int get_anon_bdev(dev_t
*);
1866 void free_anon_bdev(dev_t
);
1867 struct super_block
*sget(struct file_system_type
*type
,
1868 int (*test
)(struct super_block
*,void *),
1869 int (*set
)(struct super_block
*,void *),
1870 int flags
, void *data
);
1871 extern struct dentry
*mount_pseudo(struct file_system_type
*, char *,
1872 const struct super_operations
*ops
,
1873 const struct dentry_operations
*dops
,
1876 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1877 #define fops_get(fops) \
1878 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1879 #define fops_put(fops) \
1880 do { if (fops) module_put((fops)->owner); } while(0)
1882 extern int register_filesystem(struct file_system_type
*);
1883 extern int unregister_filesystem(struct file_system_type
*);
1884 extern struct vfsmount
*kern_mount_data(struct file_system_type
*, void *data
);
1885 #define kern_mount(type) kern_mount_data(type, NULL)
1886 extern void kern_unmount(struct vfsmount
*mnt
);
1887 extern int may_umount_tree(struct vfsmount
*);
1888 extern int may_umount(struct vfsmount
*);
1889 extern long do_mount(const char *, const char *, const char *, unsigned long, void *);
1890 extern struct vfsmount
*collect_mounts(struct path
*);
1891 extern void drop_collected_mounts(struct vfsmount
*);
1892 extern int iterate_mounts(int (*)(struct vfsmount
*, void *), void *,
1894 extern int vfs_statfs(struct path
*, struct kstatfs
*);
1895 extern int user_statfs(const char __user
*, struct kstatfs
*);
1896 extern int fd_statfs(int, struct kstatfs
*);
1897 extern int vfs_ustat(dev_t
, struct kstatfs
*);
1898 extern int freeze_super(struct super_block
*super
);
1899 extern int thaw_super(struct super_block
*super
);
1900 extern bool our_mnt(struct vfsmount
*mnt
);
1901 extern bool fs_fully_visible(struct file_system_type
*);
1903 extern int current_umask(void);
1906 extern struct kobject
*fs_kobj
;
1908 #define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
1910 #define FLOCK_VERIFY_READ 1
1911 #define FLOCK_VERIFY_WRITE 2
1913 #ifdef CONFIG_FILE_LOCKING
1914 extern int locks_mandatory_locked(struct inode
*);
1915 extern int locks_mandatory_area(int, struct inode
*, struct file
*, loff_t
, size_t);
1918 * Candidates for mandatory locking have the setgid bit set
1919 * but no group execute bit - an otherwise meaningless combination.
1922 static inline int __mandatory_lock(struct inode
*ino
)
1924 return (ino
->i_mode
& (S_ISGID
| S_IXGRP
)) == S_ISGID
;
1928 * ... and these candidates should be on MS_MANDLOCK mounted fs,
1929 * otherwise these will be advisory locks
1932 static inline int mandatory_lock(struct inode
*ino
)
1934 return IS_MANDLOCK(ino
) && __mandatory_lock(ino
);
1937 static inline int locks_verify_locked(struct inode
*inode
)
1939 if (mandatory_lock(inode
))
1940 return locks_mandatory_locked(inode
);
1944 static inline int locks_verify_truncate(struct inode
*inode
,
1948 if (inode
->i_flock
&& mandatory_lock(inode
))
1949 return locks_mandatory_area(
1950 FLOCK_VERIFY_WRITE
, inode
, filp
,
1951 size
< inode
->i_size
? size
: inode
->i_size
,
1952 (size
< inode
->i_size
? inode
->i_size
- size
1953 : size
- inode
->i_size
)
1958 static inline int break_lease(struct inode
*inode
, unsigned int mode
)
1961 return __break_lease(inode
, mode
);
1964 #else /* !CONFIG_FILE_LOCKING */
1965 static inline int locks_mandatory_locked(struct inode
*inode
)
1970 static inline int locks_mandatory_area(int rw
, struct inode
*inode
,
1971 struct file
*filp
, loff_t offset
,
1977 static inline int __mandatory_lock(struct inode
*inode
)
1982 static inline int mandatory_lock(struct inode
*inode
)
1987 static inline int locks_verify_locked(struct inode
*inode
)
1992 static inline int locks_verify_truncate(struct inode
*inode
, struct file
*filp
,
1998 static inline int break_lease(struct inode
*inode
, unsigned int mode
)
2003 #endif /* CONFIG_FILE_LOCKING */
2008 const char *name
; /* pointer to actual string */
2009 const __user
char *uptr
; /* original userland pointer */
2010 struct audit_names
*aname
;
2011 bool separate
; /* should "name" be freed? */
2014 extern long vfs_truncate(struct path
*, loff_t
);
2015 extern int do_truncate(struct dentry
*, loff_t start
, unsigned int time_attrs
,
2017 extern int do_fallocate(struct file
*file
, int mode
, loff_t offset
,
2019 extern long do_sys_open(int dfd
, const char __user
*filename
, int flags
,
2021 extern struct file
*file_open_name(struct filename
*, int, umode_t
);
2022 extern struct file
*filp_open(const char *, int, umode_t
);
2023 extern struct file
*file_open_root(struct dentry
*, struct vfsmount
*,
2025 extern struct file
* dentry_open(const struct path
*, int, const struct cred
*);
2026 extern int filp_close(struct file
*, fl_owner_t id
);
2028 extern struct filename
*getname(const char __user
*);
2034 extern int finish_open(struct file
*file
, struct dentry
*dentry
,
2035 int (*open
)(struct inode
*, struct file
*),
2037 extern int finish_no_open(struct file
*file
, struct dentry
*dentry
);
2041 extern int ioctl_preallocate(struct file
*filp
, void __user
*argp
);
2044 extern void __init
vfs_caches_init_early(void);
2045 extern void __init
vfs_caches_init(unsigned long);
2047 extern struct kmem_cache
*names_cachep
;
2049 extern void final_putname(struct filename
*name
);
2051 #define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL)
2052 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
2053 #ifndef CONFIG_AUDITSYSCALL
2054 #define putname(name) final_putname(name)
2056 extern void putname(struct filename
*name
);
2060 extern int register_blkdev(unsigned int, const char *);
2061 extern void unregister_blkdev(unsigned int, const char *);
2062 extern struct block_device
*bdget(dev_t
);
2063 extern struct block_device
*bdgrab(struct block_device
*bdev
);
2064 extern void bd_set_size(struct block_device
*, loff_t size
);
2065 extern void bd_forget(struct inode
*inode
);
2066 extern void bdput(struct block_device
*);
2067 extern void invalidate_bdev(struct block_device
*);
2068 extern void iterate_bdevs(void (*)(struct block_device
*, void *), void *);
2069 extern int sync_blockdev(struct block_device
*bdev
);
2070 extern void kill_bdev(struct block_device
*);
2071 extern struct super_block
*freeze_bdev(struct block_device
*);
2072 extern void emergency_thaw_all(void);
2073 extern int thaw_bdev(struct block_device
*bdev
, struct super_block
*sb
);
2074 extern int fsync_bdev(struct block_device
*);
2076 static inline void bd_forget(struct inode
*inode
) {}
2077 static inline int sync_blockdev(struct block_device
*bdev
) { return 0; }
2078 static inline void kill_bdev(struct block_device
*bdev
) {}
2079 static inline void invalidate_bdev(struct block_device
*bdev
) {}
2081 static inline struct super_block
*freeze_bdev(struct block_device
*sb
)
2086 static inline int thaw_bdev(struct block_device
*bdev
, struct super_block
*sb
)
2091 static inline void iterate_bdevs(void (*f
)(struct block_device
*, void *), void *arg
)
2095 extern int sync_filesystem(struct super_block
*);
2096 extern const struct file_operations def_blk_fops
;
2097 extern const struct file_operations def_chr_fops
;
2098 extern const struct file_operations bad_sock_fops
;
2100 extern int ioctl_by_bdev(struct block_device
*, unsigned, unsigned long);
2101 extern int blkdev_ioctl(struct block_device
*, fmode_t
, unsigned, unsigned long);
2102 extern long compat_blkdev_ioctl(struct file
*, unsigned, unsigned long);
2103 extern int blkdev_get(struct block_device
*bdev
, fmode_t mode
, void *holder
);
2104 extern struct block_device
*blkdev_get_by_path(const char *path
, fmode_t mode
,
2106 extern struct block_device
*blkdev_get_by_dev(dev_t dev
, fmode_t mode
,
2108 extern void blkdev_put(struct block_device
*bdev
, fmode_t mode
);
2110 extern int bd_link_disk_holder(struct block_device
*bdev
, struct gendisk
*disk
);
2111 extern void bd_unlink_disk_holder(struct block_device
*bdev
,
2112 struct gendisk
*disk
);
2114 static inline int bd_link_disk_holder(struct block_device
*bdev
,
2115 struct gendisk
*disk
)
2119 static inline void bd_unlink_disk_holder(struct block_device
*bdev
,
2120 struct gendisk
*disk
)
2127 #define CHRDEV_MAJOR_HASH_SIZE 255
2128 extern int alloc_chrdev_region(dev_t
*, unsigned, unsigned, const char *);
2129 extern int register_chrdev_region(dev_t
, unsigned, const char *);
2130 extern int __register_chrdev(unsigned int major
, unsigned int baseminor
,
2131 unsigned int count
, const char *name
,
2132 const struct file_operations
*fops
);
2133 extern void __unregister_chrdev(unsigned int major
, unsigned int baseminor
,
2134 unsigned int count
, const char *name
);
2135 extern void unregister_chrdev_region(dev_t
, unsigned);
2136 extern void chrdev_show(struct seq_file
*,off_t
);
2138 static inline int register_chrdev(unsigned int major
, const char *name
,
2139 const struct file_operations
*fops
)
2141 return __register_chrdev(major
, 0, 256, name
, fops
);
2144 static inline void unregister_chrdev(unsigned int major
, const char *name
)
2146 __unregister_chrdev(major
, 0, 256, name
);
2149 /* fs/block_dev.c */
2150 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
2151 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
2154 #define BLKDEV_MAJOR_HASH_SIZE 255
2155 extern const char *__bdevname(dev_t
, char *buffer
);
2156 extern const char *bdevname(struct block_device
*bdev
, char *buffer
);
2157 extern struct block_device
*lookup_bdev(const char *);
2158 extern void blkdev_show(struct seq_file
*,off_t
);
2161 #define BLKDEV_MAJOR_HASH_SIZE 0
2164 extern void init_special_inode(struct inode
*, umode_t
, dev_t
);
2166 /* Invalid inode operations -- fs/bad_inode.c */
2167 extern void make_bad_inode(struct inode
*);
2168 extern int is_bad_inode(struct inode
*);
2172 * return READ, READA, or WRITE
2174 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2177 * return data direction, READ or WRITE
2179 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
2181 extern void check_disk_size_change(struct gendisk
*disk
,
2182 struct block_device
*bdev
);
2183 extern int revalidate_disk(struct gendisk
*);
2184 extern int check_disk_change(struct block_device
*);
2185 extern int __invalidate_device(struct block_device
*, bool);
2186 extern int invalidate_partition(struct gendisk
*, int);
2188 unsigned long invalidate_mapping_pages(struct address_space
*mapping
,
2189 pgoff_t start
, pgoff_t end
);
2191 static inline void invalidate_remote_inode(struct inode
*inode
)
2193 if (S_ISREG(inode
->i_mode
) || S_ISDIR(inode
->i_mode
) ||
2194 S_ISLNK(inode
->i_mode
))
2195 invalidate_mapping_pages(inode
->i_mapping
, 0, -1);
2197 extern int invalidate_inode_pages2(struct address_space
*mapping
);
2198 extern int invalidate_inode_pages2_range(struct address_space
*mapping
,
2199 pgoff_t start
, pgoff_t end
);
2200 extern int write_inode_now(struct inode
*, int);
2201 extern int filemap_fdatawrite(struct address_space
*);
2202 extern int filemap_flush(struct address_space
*);
2203 extern int filemap_fdatawait(struct address_space
*);
2204 extern int filemap_fdatawait_range(struct address_space
*, loff_t lstart
,
2206 extern int filemap_write_and_wait(struct address_space
*mapping
);
2207 extern int filemap_write_and_wait_range(struct address_space
*mapping
,
2208 loff_t lstart
, loff_t lend
);
2209 extern int __filemap_fdatawrite_range(struct address_space
*mapping
,
2210 loff_t start
, loff_t end
, int sync_mode
);
2211 extern int filemap_fdatawrite_range(struct address_space
*mapping
,
2212 loff_t start
, loff_t end
);
2214 extern int vfs_fsync_range(struct file
*file
, loff_t start
, loff_t end
,
2216 extern int vfs_fsync(struct file
*file
, int datasync
);
2217 extern int generic_write_sync(struct file
*file
, loff_t pos
, loff_t count
);
2218 extern void emergency_sync(void);
2219 extern void emergency_remount(void);
2221 extern sector_t
bmap(struct inode
*, sector_t
);
2223 extern int notify_change(struct dentry
*, struct iattr
*);
2224 extern int inode_permission(struct inode
*, int);
2225 extern int generic_permission(struct inode
*, int);
2227 static inline bool execute_ok(struct inode
*inode
)
2229 return (inode
->i_mode
& S_IXUGO
) || S_ISDIR(inode
->i_mode
);
2232 static inline struct inode
*file_inode(struct file
*f
)
2237 static inline void file_start_write(struct file
*file
)
2239 if (!S_ISREG(file_inode(file
)->i_mode
))
2241 __sb_start_write(file_inode(file
)->i_sb
, SB_FREEZE_WRITE
, true);
2244 static inline bool file_start_write_trylock(struct file
*file
)
2246 if (!S_ISREG(file_inode(file
)->i_mode
))
2248 return __sb_start_write(file_inode(file
)->i_sb
, SB_FREEZE_WRITE
, false);
2251 static inline void file_end_write(struct file
*file
)
2253 if (!S_ISREG(file_inode(file
)->i_mode
))
2255 __sb_end_write(file_inode(file
)->i_sb
, SB_FREEZE_WRITE
);
2259 * get_write_access() gets write permission for a file.
2260 * put_write_access() releases this write permission.
2261 * This is used for regular files.
2262 * We cannot support write (and maybe mmap read-write shared) accesses and
2263 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2264 * can have the following values:
2265 * 0: no writers, no VM_DENYWRITE mappings
2266 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2267 * > 0: (i_writecount) users are writing to the file.
2269 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2270 * except for the cases where we don't hold i_writecount yet. Then we need to
2271 * use {get,deny}_write_access() - these functions check the sign and refuse
2272 * to do the change if sign is wrong.
2274 static inline int get_write_access(struct inode
*inode
)
2276 return atomic_inc_unless_negative(&inode
->i_writecount
) ? 0 : -ETXTBSY
;
2278 static inline int deny_write_access(struct file
*file
)
2280 struct inode
*inode
= file_inode(file
);
2281 return atomic_dec_unless_positive(&inode
->i_writecount
) ? 0 : -ETXTBSY
;
2283 static inline void put_write_access(struct inode
* inode
)
2285 atomic_dec(&inode
->i_writecount
);
2287 static inline void allow_write_access(struct file
*file
)
2290 atomic_inc(&file_inode(file
)->i_writecount
);
2293 static inline void i_readcount_dec(struct inode
*inode
)
2295 BUG_ON(!atomic_read(&inode
->i_readcount
));
2296 atomic_dec(&inode
->i_readcount
);
2298 static inline void i_readcount_inc(struct inode
*inode
)
2300 atomic_inc(&inode
->i_readcount
);
2303 static inline void i_readcount_dec(struct inode
*inode
)
2307 static inline void i_readcount_inc(struct inode
*inode
)
2312 extern int do_pipe_flags(int *, int);
2314 extern int kernel_read(struct file
*, loff_t
, char *, unsigned long);
2315 extern ssize_t
kernel_write(struct file
*, const char *, size_t, loff_t
);
2316 extern struct file
* open_exec(const char *);
2318 /* fs/dcache.c -- generic fs support functions */
2319 extern int is_subdir(struct dentry
*, struct dentry
*);
2320 extern int path_is_under(struct path
*, struct path
*);
2322 #include <linux/err.h>
2324 /* needed for stackable file system support */
2325 extern loff_t
default_llseek(struct file
*file
, loff_t offset
, int whence
);
2327 extern loff_t
vfs_llseek(struct file
*file
, loff_t offset
, int whence
);
2329 extern int inode_init_always(struct super_block
*, struct inode
*);
2330 extern void inode_init_once(struct inode
*);
2331 extern void address_space_init_once(struct address_space
*mapping
);
2332 extern void ihold(struct inode
* inode
);
2333 extern void iput(struct inode
*);
2334 extern struct inode
* igrab(struct inode
*);
2335 extern ino_t
iunique(struct super_block
*, ino_t
);
2336 extern int inode_needs_sync(struct inode
*inode
);
2337 extern int generic_delete_inode(struct inode
*inode
);
2338 static inline int generic_drop_inode(struct inode
*inode
)
2340 return !inode
->i_nlink
|| inode_unhashed(inode
);
2343 extern struct inode
*ilookup5_nowait(struct super_block
*sb
,
2344 unsigned long hashval
, int (*test
)(struct inode
*, void *),
2346 extern struct inode
*ilookup5(struct super_block
*sb
, unsigned long hashval
,
2347 int (*test
)(struct inode
*, void *), void *data
);
2348 extern struct inode
*ilookup(struct super_block
*sb
, unsigned long ino
);
2350 extern struct inode
* iget5_locked(struct super_block
*, unsigned long, int (*test
)(struct inode
*, void *), int (*set
)(struct inode
*, void *), void *);
2351 extern struct inode
* iget_locked(struct super_block
*, unsigned long);
2352 extern int insert_inode_locked4(struct inode
*, unsigned long, int (*test
)(struct inode
*, void *), void *);
2353 extern int insert_inode_locked(struct inode
*);
2354 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2355 extern void lockdep_annotate_inode_mutex_key(struct inode
*inode
);
2357 static inline void lockdep_annotate_inode_mutex_key(struct inode
*inode
) { };
2359 extern void unlock_new_inode(struct inode
*);
2360 extern unsigned int get_next_ino(void);
2362 extern void __iget(struct inode
* inode
);
2363 extern void iget_failed(struct inode
*);
2364 extern void clear_inode(struct inode
*);
2365 extern void __destroy_inode(struct inode
*);
2366 extern struct inode
*new_inode_pseudo(struct super_block
*sb
);
2367 extern struct inode
*new_inode(struct super_block
*sb
);
2368 extern void free_inode_nonrcu(struct inode
*inode
);
2369 extern int should_remove_suid(struct dentry
*);
2370 extern int file_remove_suid(struct file
*);
2372 extern void __insert_inode_hash(struct inode
*, unsigned long hashval
);
2373 static inline void insert_inode_hash(struct inode
*inode
)
2375 __insert_inode_hash(inode
, inode
->i_ino
);
2378 extern void __remove_inode_hash(struct inode
*);
2379 static inline void remove_inode_hash(struct inode
*inode
)
2381 if (!inode_unhashed(inode
))
2382 __remove_inode_hash(inode
);
2385 extern void inode_sb_list_add(struct inode
*inode
);
2388 extern void submit_bio(int, struct bio
*);
2389 extern int bdev_read_only(struct block_device
*);
2391 extern int set_blocksize(struct block_device
*, int);
2392 extern int sb_set_blocksize(struct super_block
*, int);
2393 extern int sb_min_blocksize(struct super_block
*, int);
2395 extern int generic_file_mmap(struct file
*, struct vm_area_struct
*);
2396 extern int generic_file_readonly_mmap(struct file
*, struct vm_area_struct
*);
2397 extern int generic_file_remap_pages(struct vm_area_struct
*, unsigned long addr
,
2398 unsigned long size
, pgoff_t pgoff
);
2399 extern int file_read_actor(read_descriptor_t
* desc
, struct page
*page
, unsigned long offset
, unsigned long size
);
2400 int generic_write_checks(struct file
*file
, loff_t
*pos
, size_t *count
, int isblk
);
2401 extern ssize_t
generic_file_aio_read(struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
2402 extern ssize_t
__generic_file_aio_write(struct kiocb
*, const struct iovec
*, unsigned long,
2404 extern ssize_t
generic_file_aio_write(struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
2405 extern ssize_t
generic_file_direct_write(struct kiocb
*, const struct iovec
*,
2406 unsigned long *, loff_t
, loff_t
*, size_t, size_t);
2407 extern ssize_t
generic_file_buffered_write(struct kiocb
*, const struct iovec
*,
2408 unsigned long, loff_t
, loff_t
*, size_t, ssize_t
);
2409 extern ssize_t
do_sync_read(struct file
*filp
, char __user
*buf
, size_t len
, loff_t
*ppos
);
2410 extern ssize_t
do_sync_write(struct file
*filp
, const char __user
*buf
, size_t len
, loff_t
*ppos
);
2411 extern int generic_segment_checks(const struct iovec
*iov
,
2412 unsigned long *nr_segs
, size_t *count
, int access_flags
);
2414 /* fs/block_dev.c */
2415 extern ssize_t
blkdev_aio_write(struct kiocb
*iocb
, const struct iovec
*iov
,
2416 unsigned long nr_segs
, loff_t pos
);
2417 extern int blkdev_fsync(struct file
*filp
, loff_t start
, loff_t end
,
2419 extern void block_sync_page(struct page
*page
);
2422 extern ssize_t
generic_file_splice_read(struct file
*, loff_t
*,
2423 struct pipe_inode_info
*, size_t, unsigned int);
2424 extern ssize_t
default_file_splice_read(struct file
*, loff_t
*,
2425 struct pipe_inode_info
*, size_t, unsigned int);
2426 extern ssize_t
generic_file_splice_write(struct pipe_inode_info
*,
2427 struct file
*, loff_t
*, size_t, unsigned int);
2428 extern ssize_t
generic_splice_sendpage(struct pipe_inode_info
*pipe
,
2429 struct file
*out
, loff_t
*, size_t len
, unsigned int flags
);
2432 file_ra_state_init(struct file_ra_state
*ra
, struct address_space
*mapping
);
2433 extern loff_t
noop_llseek(struct file
*file
, loff_t offset
, int whence
);
2434 extern loff_t
no_llseek(struct file
*file
, loff_t offset
, int whence
);
2435 extern loff_t
vfs_setpos(struct file
*file
, loff_t offset
, loff_t maxsize
);
2436 extern loff_t
generic_file_llseek(struct file
*file
, loff_t offset
, int whence
);
2437 extern loff_t
generic_file_llseek_size(struct file
*file
, loff_t offset
,
2438 int whence
, loff_t maxsize
, loff_t eof
);
2439 extern loff_t
fixed_size_llseek(struct file
*file
, loff_t offset
,
2440 int whence
, loff_t size
);
2441 extern int generic_file_open(struct inode
* inode
, struct file
* filp
);
2442 extern int nonseekable_open(struct inode
* inode
, struct file
* filp
);
2444 #ifdef CONFIG_FS_XIP
2445 extern ssize_t
xip_file_read(struct file
*filp
, char __user
*buf
, size_t len
,
2447 extern int xip_file_mmap(struct file
* file
, struct vm_area_struct
* vma
);
2448 extern ssize_t
xip_file_write(struct file
*filp
, const char __user
*buf
,
2449 size_t len
, loff_t
*ppos
);
2450 extern int xip_truncate_page(struct address_space
*mapping
, loff_t from
);
2452 static inline int xip_truncate_page(struct address_space
*mapping
, loff_t from
)
2459 typedef void (dio_submit_t
)(int rw
, struct bio
*bio
, struct inode
*inode
,
2460 loff_t file_offset
);
2463 /* need locking between buffered and direct access */
2466 /* filesystem does not support filling holes */
2467 DIO_SKIP_HOLES
= 0x02,
2470 void dio_end_io(struct bio
*bio
, int error
);
2472 ssize_t
__blockdev_direct_IO(int rw
, struct kiocb
*iocb
, struct inode
*inode
,
2473 struct block_device
*bdev
, const struct iovec
*iov
, loff_t offset
,
2474 unsigned long nr_segs
, get_block_t get_block
, dio_iodone_t end_io
,
2475 dio_submit_t submit_io
, int flags
);
2477 static inline ssize_t
blockdev_direct_IO(int rw
, struct kiocb
*iocb
,
2478 struct inode
*inode
, const struct iovec
*iov
, loff_t offset
,
2479 unsigned long nr_segs
, get_block_t get_block
)
2481 return __blockdev_direct_IO(rw
, iocb
, inode
, inode
->i_sb
->s_bdev
, iov
,
2482 offset
, nr_segs
, get_block
, NULL
, NULL
,
2483 DIO_LOCKING
| DIO_SKIP_HOLES
);
2487 void inode_dio_wait(struct inode
*inode
);
2488 void inode_dio_done(struct inode
*inode
);
2490 extern const struct file_operations generic_ro_fops
;
2492 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2494 extern int vfs_readlink(struct dentry
*, char __user
*, int, const char *);
2495 extern int page_readlink(struct dentry
*, char __user
*, int);
2496 extern void *page_follow_link_light(struct dentry
*, struct nameidata
*);
2497 extern void page_put_link(struct dentry
*, struct nameidata
*, void *);
2498 extern int __page_symlink(struct inode
*inode
, const char *symname
, int len
,
2500 extern int page_symlink(struct inode
*inode
, const char *symname
, int len
);
2501 extern const struct inode_operations page_symlink_inode_operations
;
2502 extern int generic_readlink(struct dentry
*, char __user
*, int);
2503 extern void generic_fillattr(struct inode
*, struct kstat
*);
2504 extern int vfs_getattr(struct path
*, struct kstat
*);
2505 void __inode_add_bytes(struct inode
*inode
, loff_t bytes
);
2506 void inode_add_bytes(struct inode
*inode
, loff_t bytes
);
2507 void __inode_sub_bytes(struct inode
*inode
, loff_t bytes
);
2508 void inode_sub_bytes(struct inode
*inode
, loff_t bytes
);
2509 loff_t
inode_get_bytes(struct inode
*inode
);
2510 void inode_set_bytes(struct inode
*inode
, loff_t bytes
);
2512 extern int vfs_readdir(struct file
*, filldir_t
, void *);
2513 extern int iterate_dir(struct file
*, struct dir_context
*);
2515 extern int vfs_stat(const char __user
*, struct kstat
*);
2516 extern int vfs_lstat(const char __user
*, struct kstat
*);
2517 extern int vfs_fstat(unsigned int, struct kstat
*);
2518 extern int vfs_fstatat(int , const char __user
*, struct kstat
*, int);
2520 extern int do_vfs_ioctl(struct file
*filp
, unsigned int fd
, unsigned int cmd
,
2522 extern int __generic_block_fiemap(struct inode
*inode
,
2523 struct fiemap_extent_info
*fieinfo
,
2524 loff_t start
, loff_t len
,
2525 get_block_t
*get_block
);
2526 extern int generic_block_fiemap(struct inode
*inode
,
2527 struct fiemap_extent_info
*fieinfo
, u64 start
,
2528 u64 len
, get_block_t
*get_block
);
2530 extern void get_filesystem(struct file_system_type
*fs
);
2531 extern void put_filesystem(struct file_system_type
*fs
);
2532 extern struct file_system_type
*get_fs_type(const char *name
);
2533 extern struct super_block
*get_super(struct block_device
*);
2534 extern struct super_block
*get_super_thawed(struct block_device
*);
2535 extern struct super_block
*get_active_super(struct block_device
*bdev
);
2536 extern void drop_super(struct super_block
*sb
);
2537 extern void iterate_supers(void (*)(struct super_block
*, void *), void *);
2538 extern void iterate_supers_type(struct file_system_type
*,
2539 void (*)(struct super_block
*, void *), void *);
2541 extern int dcache_dir_open(struct inode
*, struct file
*);
2542 extern int dcache_dir_close(struct inode
*, struct file
*);
2543 extern loff_t
dcache_dir_lseek(struct file
*, loff_t
, int);
2544 extern int dcache_readdir(struct file
*, struct dir_context
*);
2545 extern int simple_setattr(struct dentry
*, struct iattr
*);
2546 extern int simple_getattr(struct vfsmount
*, struct dentry
*, struct kstat
*);
2547 extern int simple_statfs(struct dentry
*, struct kstatfs
*);
2548 extern int simple_open(struct inode
*inode
, struct file
*file
);
2549 extern int simple_link(struct dentry
*, struct inode
*, struct dentry
*);
2550 extern int simple_unlink(struct inode
*, struct dentry
*);
2551 extern int simple_rmdir(struct inode
*, struct dentry
*);
2552 extern int simple_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
2553 extern int noop_fsync(struct file
*, loff_t
, loff_t
, int);
2554 extern int simple_empty(struct dentry
*);
2555 extern int simple_readpage(struct file
*file
, struct page
*page
);
2556 extern int simple_write_begin(struct file
*file
, struct address_space
*mapping
,
2557 loff_t pos
, unsigned len
, unsigned flags
,
2558 struct page
**pagep
, void **fsdata
);
2559 extern int simple_write_end(struct file
*file
, struct address_space
*mapping
,
2560 loff_t pos
, unsigned len
, unsigned copied
,
2561 struct page
*page
, void *fsdata
);
2563 extern struct dentry
*simple_lookup(struct inode
*, struct dentry
*, unsigned int flags
);
2564 extern ssize_t
generic_read_dir(struct file
*, char __user
*, size_t, loff_t
*);
2565 extern const struct file_operations simple_dir_operations
;
2566 extern const struct inode_operations simple_dir_inode_operations
;
2567 struct tree_descr
{ char *name
; const struct file_operations
*ops
; int mode
; };
2568 struct dentry
*d_alloc_name(struct dentry
*, const char *);
2569 extern int simple_fill_super(struct super_block
*, unsigned long, struct tree_descr
*);
2570 extern int simple_pin_fs(struct file_system_type
*, struct vfsmount
**mount
, int *count
);
2571 extern void simple_release_fs(struct vfsmount
**mount
, int *count
);
2573 extern ssize_t
simple_read_from_buffer(void __user
*to
, size_t count
,
2574 loff_t
*ppos
, const void *from
, size_t available
);
2575 extern ssize_t
simple_write_to_buffer(void *to
, size_t available
, loff_t
*ppos
,
2576 const void __user
*from
, size_t count
);
2578 extern int generic_file_fsync(struct file
*, loff_t
, loff_t
, int);
2580 extern int generic_check_addressable(unsigned, u64
);
2582 #ifdef CONFIG_MIGRATION
2583 extern int buffer_migrate_page(struct address_space
*,
2584 struct page
*, struct page
*,
2587 #define buffer_migrate_page NULL
2590 extern int inode_change_ok(const struct inode
*, struct iattr
*);
2591 extern int inode_newsize_ok(const struct inode
*, loff_t offset
);
2592 extern void setattr_copy(struct inode
*inode
, const struct iattr
*attr
);
2594 extern int file_update_time(struct file
*file
);
2596 extern int generic_show_options(struct seq_file
*m
, struct dentry
*root
);
2597 extern void save_mount_options(struct super_block
*sb
, char *options
);
2598 extern void replace_mount_options(struct super_block
*sb
, char *options
);
2600 static inline ino_t
parent_ino(struct dentry
*dentry
)
2605 * Don't strictly need d_lock here? If the parent ino could change
2606 * then surely we'd have a deeper race in the caller?
2608 spin_lock(&dentry
->d_lock
);
2609 res
= dentry
->d_parent
->d_inode
->i_ino
;
2610 spin_unlock(&dentry
->d_lock
);
2614 /* Transaction based IO helpers */
2617 * An argresp is stored in an allocated page and holds the
2618 * size of the argument or response, along with its content
2620 struct simple_transaction_argresp
{
2625 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2627 char *simple_transaction_get(struct file
*file
, const char __user
*buf
,
2629 ssize_t
simple_transaction_read(struct file
*file
, char __user
*buf
,
2630 size_t size
, loff_t
*pos
);
2631 int simple_transaction_release(struct inode
*inode
, struct file
*file
);
2633 void simple_transaction_set(struct file
*file
, size_t n
);
2636 * simple attribute files
2638 * These attributes behave similar to those in sysfs:
2640 * Writing to an attribute immediately sets a value, an open file can be
2641 * written to multiple times.
2643 * Reading from an attribute creates a buffer from the value that might get
2644 * read with multiple read calls. When the attribute has been read
2645 * completely, no further read calls are possible until the file is opened
2648 * All attributes contain a text representation of a numeric value
2649 * that are accessed with the get() and set() functions.
2651 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2652 static int __fops ## _open(struct inode *inode, struct file *file) \
2654 __simple_attr_check_format(__fmt, 0ull); \
2655 return simple_attr_open(inode, file, __get, __set, __fmt); \
2657 static const struct file_operations __fops = { \
2658 .owner = THIS_MODULE, \
2659 .open = __fops ## _open, \
2660 .release = simple_attr_release, \
2661 .read = simple_attr_read, \
2662 .write = simple_attr_write, \
2663 .llseek = generic_file_llseek, \
2666 static inline __printf(1, 2)
2667 void __simple_attr_check_format(const char *fmt
, ...)
2669 /* don't do anything, just let the compiler check the arguments; */
2672 int simple_attr_open(struct inode
*inode
, struct file
*file
,
2673 int (*get
)(void *, u64
*), int (*set
)(void *, u64
),
2675 int simple_attr_release(struct inode
*inode
, struct file
*file
);
2676 ssize_t
simple_attr_read(struct file
*file
, char __user
*buf
,
2677 size_t len
, loff_t
*ppos
);
2678 ssize_t
simple_attr_write(struct file
*file
, const char __user
*buf
,
2679 size_t len
, loff_t
*ppos
);
2682 int proc_nr_files(struct ctl_table
*table
, int write
,
2683 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2684 int proc_nr_dentry(struct ctl_table
*table
, int write
,
2685 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2686 int proc_nr_inodes(struct ctl_table
*table
, int write
,
2687 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2688 int __init
get_filesystem_list(char *buf
);
2690 #define __FMODE_EXEC ((__force int) FMODE_EXEC)
2691 #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
2693 #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
2694 #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
2695 (flag & __FMODE_NONOTIFY)))
2697 static inline int is_sxid(umode_t mode
)
2699 return (mode
& S_ISUID
) || ((mode
& S_ISGID
) && (mode
& S_IXGRP
));
2702 static inline void inode_has_no_xattr(struct inode
*inode
)
2704 if (!is_sxid(inode
->i_mode
) && (inode
->i_sb
->s_flags
& MS_NOSEC
))
2705 inode
->i_flags
|= S_NOSEC
;
2708 static inline bool dir_emit(struct dir_context
*ctx
,
2709 const char *name
, int namelen
,
2710 u64 ino
, unsigned type
)
2712 return ctx
->actor(ctx
, name
, namelen
, ctx
->pos
, ino
, type
) == 0;
2714 static inline bool dir_emit_dot(struct file
*file
, struct dir_context
*ctx
)
2716 return ctx
->actor(ctx
, ".", 1, ctx
->pos
,
2717 file
->f_path
.dentry
->d_inode
->i_ino
, DT_DIR
) == 0;
2719 static inline bool dir_emit_dotdot(struct file
*file
, struct dir_context
*ctx
)
2721 return ctx
->actor(ctx
, "..", 2, ctx
->pos
,
2722 parent_ino(file
->f_path
.dentry
), DT_DIR
) == 0;
2724 static inline bool dir_emit_dots(struct file
*file
, struct dir_context
*ctx
)
2726 if (ctx
->pos
== 0) {
2727 if (!dir_emit_dot(file
, ctx
))
2731 if (ctx
->pos
== 1) {
2732 if (!dir_emit_dotdot(file
, ctx
))
2738 static inline bool dir_relax(struct inode
*inode
)
2740 mutex_unlock(&inode
->i_mutex
);
2741 mutex_lock(&inode
->i_mutex
);
2742 return !IS_DEADDIR(inode
);
2745 #endif /* _LINUX_FS_H */