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/radix-tree.h>
14 #include <linux/rbtree.h>
15 #include <linux/init.h>
16 #include <linux/pid.h>
17 #include <linux/bug.h>
18 #include <linux/mutex.h>
19 #include <linux/capability.h>
20 #include <linux/semaphore.h>
21 #include <linux/fiemap.h>
22 #include <linux/rculist_bl.h>
23 #include <linux/atomic.h>
24 #include <linux/shrinker.h>
25 #include <linux/migrate_mode.h>
26 #include <linux/uidgid.h>
27 #include <linux/lockdep.h>
28 #include <linux/percpu-rwsem.h>
29 #include <linux/blk_types.h>
31 #include <asm/byteorder.h>
32 #include <uapi/linux/fs.h>
34 struct export_operations
;
40 struct pipe_inode_info
;
41 struct poll_table_struct
;
43 struct vm_area_struct
;
46 struct swap_info_struct
;
49 extern void __init
inode_init(void);
50 extern void __init
inode_init_early(void);
51 extern void __init
files_init(unsigned long);
53 extern struct files_stat_struct files_stat
;
54 extern unsigned long get_max_files(void);
55 extern int sysctl_nr_open
;
56 extern struct inodes_stat_t inodes_stat
;
57 extern int leases_enable
, lease_break_time
;
58 extern int sysctl_protected_symlinks
;
59 extern int sysctl_protected_hardlinks
;
62 typedef int (get_block_t
)(struct inode
*inode
, sector_t iblock
,
63 struct buffer_head
*bh_result
, int create
);
64 typedef void (dio_iodone_t
)(struct kiocb
*iocb
, loff_t offset
,
65 ssize_t bytes
, void *private, int ret
,
68 #define MAY_EXEC 0x00000001
69 #define MAY_WRITE 0x00000002
70 #define MAY_READ 0x00000004
71 #define MAY_APPEND 0x00000008
72 #define MAY_ACCESS 0x00000010
73 #define MAY_OPEN 0x00000020
74 #define MAY_CHDIR 0x00000040
75 /* called from RCU mode, don't block */
76 #define MAY_NOT_BLOCK 0x00000080
79 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
80 * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
83 /* file is open for reading */
84 #define FMODE_READ ((__force fmode_t)0x1)
85 /* file is open for writing */
86 #define FMODE_WRITE ((__force fmode_t)0x2)
87 /* file is seekable */
88 #define FMODE_LSEEK ((__force fmode_t)0x4)
89 /* file can be accessed using pread */
90 #define FMODE_PREAD ((__force fmode_t)0x8)
91 /* file can be accessed using pwrite */
92 #define FMODE_PWRITE ((__force fmode_t)0x10)
93 /* File is opened for execution with sys_execve / sys_uselib */
94 #define FMODE_EXEC ((__force fmode_t)0x20)
95 /* File is opened with O_NDELAY (only set for block devices) */
96 #define FMODE_NDELAY ((__force fmode_t)0x40)
97 /* File is opened with O_EXCL (only set for block devices) */
98 #define FMODE_EXCL ((__force fmode_t)0x80)
99 /* File is opened using open(.., 3, ..) and is writeable only for ioctls
100 (specialy hack for floppy.c) */
101 #define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
102 /* 32bit hashes as llseek() offset (for directories) */
103 #define FMODE_32BITHASH ((__force fmode_t)0x200)
104 /* 64bit hashes as llseek() offset (for directories) */
105 #define FMODE_64BITHASH ((__force fmode_t)0x400)
108 * Don't update ctime and mtime.
110 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
111 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
113 #define FMODE_NOCMTIME ((__force fmode_t)0x800)
115 /* Expect random access pattern */
116 #define FMODE_RANDOM ((__force fmode_t)0x1000)
118 /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
119 #define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
121 /* File is opened with O_PATH; almost nothing can be done with it */
122 #define FMODE_PATH ((__force fmode_t)0x4000)
124 /* File was opened by fanotify and shouldn't generate fanotify events */
125 #define FMODE_NONOTIFY ((__force fmode_t)0x1000000)
128 * Flag for rw_copy_check_uvector and compat_rw_copy_check_uvector
129 * that indicates that they should check the contents of the iovec are
130 * valid, but not check the memory that the iovec elements
133 #define CHECK_IOVEC_ONLY -1
136 * The below are the various read and write types that we support. Some of
137 * them include behavioral modifiers that send information down to the
138 * block layer and IO scheduler. Terminology:
140 * The block layer uses device plugging to defer IO a little bit, in
141 * the hope that we will see more IO very shortly. This increases
142 * coalescing of adjacent IO and thus reduces the number of IOs we
143 * have to send to the device. It also allows for better queuing,
144 * if the IO isn't mergeable. If the caller is going to be waiting
145 * for the IO, then he must ensure that the device is unplugged so
146 * that the IO is dispatched to the driver.
148 * All IO is handled async in Linux. This is fine for background
149 * writes, but for reads or writes that someone waits for completion
150 * on, we want to notify the block layer and IO scheduler so that they
151 * know about it. That allows them to make better scheduling
152 * decisions. So when the below references 'sync' and 'async', it
153 * is referencing this priority hint.
155 * With that in mind, the available types are:
157 * READ A normal read operation. Device will be plugged.
158 * READ_SYNC A synchronous read. Device is not plugged, caller can
159 * immediately wait on this read without caring about
161 * READA Used for read-ahead operations. Lower priority, and the
162 * block layer could (in theory) choose to ignore this
163 * request if it runs into resource problems.
164 * WRITE A normal async write. Device will be plugged.
165 * WRITE_SYNC Synchronous write. Identical to WRITE, but passes down
166 * the hint that someone will be waiting on this IO
167 * shortly. The write equivalent of READ_SYNC.
168 * WRITE_ODIRECT Special case write for O_DIRECT only.
169 * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
170 * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
171 * non-volatile media on completion.
172 * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
173 * by a cache flush and data is guaranteed to be on
174 * non-volatile media on completion.
177 #define RW_MASK REQ_WRITE
178 #define RWA_MASK REQ_RAHEAD
181 #define WRITE RW_MASK
182 #define READA RWA_MASK
183 #define KERNEL_READ (READ|REQ_KERNEL)
184 #define KERNEL_WRITE (WRITE|REQ_KERNEL)
186 #define READ_SYNC (READ | REQ_SYNC)
187 #define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE)
188 #define WRITE_ODIRECT (WRITE | REQ_SYNC)
189 #define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
190 #define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
191 #define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
194 * Attribute flags. These should be or-ed together to figure out what
197 #define ATTR_MODE (1 << 0)
198 #define ATTR_UID (1 << 1)
199 #define ATTR_GID (1 << 2)
200 #define ATTR_SIZE (1 << 3)
201 #define ATTR_ATIME (1 << 4)
202 #define ATTR_MTIME (1 << 5)
203 #define ATTR_CTIME (1 << 6)
204 #define ATTR_ATIME_SET (1 << 7)
205 #define ATTR_MTIME_SET (1 << 8)
206 #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
207 #define ATTR_ATTR_FLAG (1 << 10)
208 #define ATTR_KILL_SUID (1 << 11)
209 #define ATTR_KILL_SGID (1 << 12)
210 #define ATTR_FILE (1 << 13)
211 #define ATTR_KILL_PRIV (1 << 14)
212 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
213 #define ATTR_TIMES_SET (1 << 16)
216 * This is the Inode Attributes structure, used for notify_change(). It
217 * uses the above definitions as flags, to know which values have changed.
218 * Also, in this manner, a Filesystem can look at only the values it cares
219 * about. Basically, these are the attributes that the VFS layer can
220 * request to change from the FS layer.
222 * Derek Atkins <warlord@MIT.EDU> 94-10-20
225 unsigned int ia_valid
;
230 struct timespec ia_atime
;
231 struct timespec ia_mtime
;
232 struct timespec ia_ctime
;
235 * Not an attribute, but an auxiliary info for filesystems wanting to
236 * implement an ftruncate() like method. NOTE: filesystem should
237 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
239 struct file
*ia_file
;
243 * Includes for diskquotas.
245 #include <linux/quota.h>
248 * enum positive_aop_returns - aop return codes with specific semantics
250 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
251 * completed, that the page is still locked, and
252 * should be considered active. The VM uses this hint
253 * to return the page to the active list -- it won't
254 * be a candidate for writeback again in the near
255 * future. Other callers must be careful to unlock
256 * the page if they get this return. Returned by
259 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
260 * unlocked it and the page might have been truncated.
261 * The caller should back up to acquiring a new page and
262 * trying again. The aop will be taking reasonable
263 * precautions not to livelock. If the caller held a page
264 * reference, it should drop it before retrying. Returned
267 * address_space_operation functions return these large constants to indicate
268 * special semantics to the caller. These are much larger than the bytes in a
269 * page to allow for functions that return the number of bytes operated on in a
273 enum positive_aop_returns
{
274 AOP_WRITEPAGE_ACTIVATE
= 0x80000,
275 AOP_TRUNCATED_PAGE
= 0x80001,
278 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
279 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
280 #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
281 * helper code (eg buffer layer)
282 * to clear GFP_FS from alloc */
285 * oh the beauties of C type declarations.
288 struct address_space
;
289 struct writeback_control
;
292 const struct iovec
*iov
;
293 unsigned long nr_segs
;
298 size_t iov_iter_copy_from_user_atomic(struct page
*page
,
299 struct iov_iter
*i
, unsigned long offset
, size_t bytes
);
300 size_t iov_iter_copy_from_user(struct page
*page
,
301 struct iov_iter
*i
, unsigned long offset
, size_t bytes
);
302 void iov_iter_advance(struct iov_iter
*i
, size_t bytes
);
303 int iov_iter_fault_in_readable(struct iov_iter
*i
, size_t bytes
);
304 size_t iov_iter_single_seg_count(const struct iov_iter
*i
);
306 static inline void iov_iter_init(struct iov_iter
*i
,
307 const struct iovec
*iov
, unsigned long nr_segs
,
308 size_t count
, size_t written
)
311 i
->nr_segs
= nr_segs
;
313 i
->count
= count
+ written
;
315 iov_iter_advance(i
, written
);
318 static inline size_t iov_iter_count(struct iov_iter
*i
)
324 * "descriptor" for what we're up to with a read.
325 * This allows us to use the same read code yet
326 * have multiple different users of the data that
327 * we read from a file.
329 * The simplest case just copies the data to user
342 typedef int (*read_actor_t
)(read_descriptor_t
*, struct page
*,
343 unsigned long, unsigned long);
345 struct address_space_operations
{
346 int (*writepage
)(struct page
*page
, struct writeback_control
*wbc
);
347 int (*readpage
)(struct file
*, struct page
*);
349 /* Write back some dirty pages from this mapping. */
350 int (*writepages
)(struct address_space
*, struct writeback_control
*);
352 /* Set a page dirty. Return true if this dirtied it */
353 int (*set_page_dirty
)(struct page
*page
);
355 int (*readpages
)(struct file
*filp
, struct address_space
*mapping
,
356 struct list_head
*pages
, unsigned nr_pages
);
358 int (*write_begin
)(struct file
*, struct address_space
*mapping
,
359 loff_t pos
, unsigned len
, unsigned flags
,
360 struct page
**pagep
, void **fsdata
);
361 int (*write_end
)(struct file
*, struct address_space
*mapping
,
362 loff_t pos
, unsigned len
, unsigned copied
,
363 struct page
*page
, void *fsdata
);
365 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
366 sector_t (*bmap
)(struct address_space
*, sector_t
);
367 void (*invalidatepage
) (struct page
*, unsigned int, unsigned int);
368 int (*releasepage
) (struct page
*, gfp_t
);
369 void (*freepage
)(struct page
*);
370 ssize_t (*direct_IO
)(int, struct kiocb
*, const struct iovec
*iov
,
371 loff_t offset
, unsigned long nr_segs
);
372 int (*get_xip_mem
)(struct address_space
*, pgoff_t
, int,
373 void **, unsigned long *);
375 * migrate the contents of a page to the specified target. If
376 * migrate_mode is MIGRATE_ASYNC, it must not block.
378 int (*migratepage
) (struct address_space
*,
379 struct page
*, struct page
*, enum migrate_mode
);
380 int (*launder_page
) (struct page
*);
381 int (*is_partially_uptodate
) (struct page
*, read_descriptor_t
*,
383 void (*is_dirty_writeback
) (struct page
*, bool *, bool *);
384 int (*error_remove_page
)(struct address_space
*, struct page
*);
386 /* swapfile support */
387 int (*swap_activate
)(struct swap_info_struct
*sis
, struct file
*file
,
389 void (*swap_deactivate
)(struct file
*file
);
392 extern const struct address_space_operations empty_aops
;
395 * pagecache_write_begin/pagecache_write_end must be used by general code
396 * to write into the pagecache.
398 int pagecache_write_begin(struct file
*, struct address_space
*mapping
,
399 loff_t pos
, unsigned len
, unsigned flags
,
400 struct page
**pagep
, void **fsdata
);
402 int pagecache_write_end(struct file
*, struct address_space
*mapping
,
403 loff_t pos
, unsigned len
, unsigned copied
,
404 struct page
*page
, void *fsdata
);
406 struct backing_dev_info
;
407 struct address_space
{
408 struct inode
*host
; /* owner: inode, block_device */
409 struct radix_tree_root page_tree
; /* radix tree of all pages */
410 spinlock_t tree_lock
; /* and lock protecting it */
411 unsigned int i_mmap_writable
;/* count VM_SHARED mappings */
412 struct rb_root i_mmap
; /* tree of private and shared mappings */
413 struct list_head i_mmap_nonlinear
;/*list VM_NONLINEAR mappings */
414 struct mutex i_mmap_mutex
; /* protect tree, count, list */
415 /* Protected by tree_lock together with the radix tree */
416 unsigned long nrpages
; /* number of total pages */
417 pgoff_t writeback_index
;/* writeback starts here */
418 const struct address_space_operations
*a_ops
; /* methods */
419 unsigned long flags
; /* error bits/gfp mask */
420 struct backing_dev_info
*backing_dev_info
; /* device readahead, etc */
421 spinlock_t private_lock
; /* for use by the address_space */
422 struct list_head private_list
; /* ditto */
423 void *private_data
; /* ditto */
424 } __attribute__((aligned(sizeof(long))));
426 * On most architectures that alignment is already the case; but
427 * must be enforced here for CRIS, to let the least significant bit
428 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
430 struct request_queue
;
432 struct block_device
{
433 dev_t bd_dev
; /* not a kdev_t - it's a search key */
435 struct inode
* bd_inode
; /* will die */
436 struct super_block
* bd_super
;
437 struct mutex bd_mutex
; /* open/close mutex */
438 struct list_head bd_inodes
;
442 bool bd_write_holder
;
444 struct list_head bd_holder_disks
;
446 struct block_device
* bd_contains
;
447 unsigned bd_block_size
;
448 struct hd_struct
* bd_part
;
449 /* number of times partitions within this device have been opened. */
450 unsigned bd_part_count
;
452 struct gendisk
* bd_disk
;
453 struct request_queue
* bd_queue
;
454 struct list_head bd_list
;
456 * Private data. You must have bd_claim'ed the block_device
457 * to use this. NOTE: bd_claim allows an owner to claim
458 * the same device multiple times, the owner must take special
459 * care to not mess up bd_private for that case.
461 unsigned long bd_private
;
463 /* The counter of freeze processes */
464 int bd_fsfreeze_count
;
465 /* Mutex for freeze */
466 struct mutex bd_fsfreeze_mutex
;
470 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
473 #define PAGECACHE_TAG_DIRTY 0
474 #define PAGECACHE_TAG_WRITEBACK 1
475 #define PAGECACHE_TAG_TOWRITE 2
477 int mapping_tagged(struct address_space
*mapping
, int tag
);
480 * Might pages of this file be mapped into userspace?
482 static inline int mapping_mapped(struct address_space
*mapping
)
484 return !RB_EMPTY_ROOT(&mapping
->i_mmap
) ||
485 !list_empty(&mapping
->i_mmap_nonlinear
);
489 * Might pages of this file have been modified in userspace?
490 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
491 * marks vma as VM_SHARED if it is shared, and the file was opened for
492 * writing i.e. vma may be mprotected writable even if now readonly.
494 static inline int mapping_writably_mapped(struct address_space
*mapping
)
496 return mapping
->i_mmap_writable
!= 0;
500 * Use sequence counter to get consistent i_size on 32-bit processors.
502 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
503 #include <linux/seqlock.h>
504 #define __NEED_I_SIZE_ORDERED
505 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
507 #define i_size_ordered_init(inode) do { } while (0)
511 #define ACL_NOT_CACHED ((void *)(-1))
513 #define IOP_FASTPERM 0x0001
514 #define IOP_LOOKUP 0x0002
515 #define IOP_NOFOLLOW 0x0004
518 * Keep mostly read-only and often accessed (especially for
519 * the RCU path lookup and 'stat' data) fields at the beginning
520 * of the 'struct inode'
524 unsigned short i_opflags
;
527 unsigned int i_flags
;
529 #ifdef CONFIG_FS_POSIX_ACL
530 struct posix_acl
*i_acl
;
531 struct posix_acl
*i_default_acl
;
534 const struct inode_operations
*i_op
;
535 struct super_block
*i_sb
;
536 struct address_space
*i_mapping
;
538 #ifdef CONFIG_SECURITY
542 /* Stat data, not accessed from path walking */
545 * Filesystems may only read i_nlink directly. They shall use the
546 * following functions for modification:
548 * (set|clear|inc|drop)_nlink
549 * inode_(inc|dec)_link_count
552 const unsigned int i_nlink
;
553 unsigned int __i_nlink
;
557 struct timespec i_atime
;
558 struct timespec i_mtime
;
559 struct timespec i_ctime
;
560 spinlock_t i_lock
; /* i_blocks, i_bytes, maybe i_size */
561 unsigned short i_bytes
;
562 unsigned int i_blkbits
;
565 #ifdef __NEED_I_SIZE_ORDERED
566 seqcount_t i_size_seqcount
;
570 unsigned long i_state
;
571 struct mutex i_mutex
;
573 unsigned long dirtied_when
; /* jiffies of first dirtying */
575 struct hlist_node i_hash
;
576 struct list_head i_wb_list
; /* backing dev IO list */
577 struct list_head i_lru
; /* inode LRU list */
578 struct list_head i_sb_list
;
580 struct hlist_head i_dentry
;
581 struct rcu_head i_rcu
;
585 atomic_t i_dio_count
;
586 atomic_t i_writecount
;
587 const struct file_operations
*i_fop
; /* former ->i_op->default_file_ops */
588 struct file_lock
*i_flock
;
589 struct address_space i_data
;
591 struct dquot
*i_dquot
[MAXQUOTAS
];
593 struct list_head i_devices
;
595 struct pipe_inode_info
*i_pipe
;
596 struct block_device
*i_bdev
;
602 #ifdef CONFIG_FSNOTIFY
603 __u32 i_fsnotify_mask
; /* all events this inode cares about */
604 struct hlist_head i_fsnotify_marks
;
608 atomic_t i_readcount
; /* struct files open RO */
610 void *i_private
; /* fs or device private pointer */
613 static inline int inode_unhashed(struct inode
*inode
)
615 return hlist_unhashed(&inode
->i_hash
);
619 * inode->i_mutex nesting subclasses for the lock validator:
621 * 0: the object of the current VFS operation
626 * The locking order between these classes is
627 * parent -> child -> normal -> xattr -> quota
629 enum inode_i_mutex_lock_class
639 * NOTE: in a 32bit arch with a preemptable kernel and
640 * an UP compile the i_size_read/write must be atomic
641 * with respect to the local cpu (unlike with preempt disabled),
642 * but they don't need to be atomic with respect to other cpus like in
643 * true SMP (so they need either to either locally disable irq around
644 * the read or for example on x86 they can be still implemented as a
645 * cmpxchg8b without the need of the lock prefix). For SMP compiles
646 * and 64bit archs it makes no difference if preempt is enabled or not.
648 static inline loff_t
i_size_read(const struct inode
*inode
)
650 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
655 seq
= read_seqcount_begin(&inode
->i_size_seqcount
);
656 i_size
= inode
->i_size
;
657 } while (read_seqcount_retry(&inode
->i_size_seqcount
, seq
));
659 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
663 i_size
= inode
->i_size
;
667 return inode
->i_size
;
672 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
673 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
674 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
676 static inline void i_size_write(struct inode
*inode
, loff_t i_size
)
678 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
680 write_seqcount_begin(&inode
->i_size_seqcount
);
681 inode
->i_size
= i_size
;
682 write_seqcount_end(&inode
->i_size_seqcount
);
684 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
686 inode
->i_size
= i_size
;
689 inode
->i_size
= i_size
;
693 /* Helper functions so that in most cases filesystems will
694 * not need to deal directly with kuid_t and kgid_t and can
695 * instead deal with the raw numeric values that are stored
698 static inline uid_t
i_uid_read(const struct inode
*inode
)
700 return from_kuid(&init_user_ns
, inode
->i_uid
);
703 static inline gid_t
i_gid_read(const struct inode
*inode
)
705 return from_kgid(&init_user_ns
, inode
->i_gid
);
708 static inline void i_uid_write(struct inode
*inode
, uid_t uid
)
710 inode
->i_uid
= make_kuid(&init_user_ns
, uid
);
713 static inline void i_gid_write(struct inode
*inode
, gid_t gid
)
715 inode
->i_gid
= make_kgid(&init_user_ns
, gid
);
718 static inline unsigned iminor(const struct inode
*inode
)
720 return MINOR(inode
->i_rdev
);
723 static inline unsigned imajor(const struct inode
*inode
)
725 return MAJOR(inode
->i_rdev
);
728 extern struct block_device
*I_BDEV(struct inode
*inode
);
731 rwlock_t lock
; /* protects pid, uid, euid fields */
732 struct pid
*pid
; /* pid or -pgrp where SIGIO should be sent */
733 enum pid_type pid_type
; /* Kind of process group SIGIO should be sent to */
734 kuid_t uid
, euid
; /* uid/euid of process setting the owner */
735 int signum
; /* posix.1b rt signal to be delivered on IO */
739 * Track a single file's readahead state
741 struct file_ra_state
{
742 pgoff_t start
; /* where readahead started */
743 unsigned int size
; /* # of readahead pages */
744 unsigned int async_size
; /* do asynchronous readahead when
745 there are only # of pages ahead */
747 unsigned int ra_pages
; /* Maximum readahead window */
748 unsigned int mmap_miss
; /* Cache miss stat for mmap accesses */
749 loff_t prev_pos
; /* Cache last read() position */
753 * Check if @index falls in the readahead windows.
755 static inline int ra_has_index(struct file_ra_state
*ra
, pgoff_t index
)
757 return (index
>= ra
->start
&&
758 index
< ra
->start
+ ra
->size
);
761 #define FILE_MNT_WRITE_TAKEN 1
762 #define FILE_MNT_WRITE_RELEASED 2
766 * fu_list becomes invalid after file_free is called and queued via
767 * fu_rcuhead for RCU freeing
770 struct list_head fu_list
;
771 struct rcu_head fu_rcuhead
;
774 #define f_dentry f_path.dentry
775 struct inode
*f_inode
; /* cached value */
776 const struct file_operations
*f_op
;
779 * Protects f_ep_links, f_flags, f_pos vs i_size in lseek SEEK_CUR.
780 * Must not be taken from IRQ context.
786 atomic_long_t f_count
;
787 unsigned int f_flags
;
790 struct fown_struct f_owner
;
791 const struct cred
*f_cred
;
792 struct file_ra_state f_ra
;
795 #ifdef CONFIG_SECURITY
798 /* needed for tty driver, and maybe others */
802 /* Used by fs/eventpoll.c to link all the hooks to this file */
803 struct list_head f_ep_links
;
804 struct list_head f_tfile_llink
;
805 #endif /* #ifdef CONFIG_EPOLL */
806 struct address_space
*f_mapping
;
807 #ifdef CONFIG_DEBUG_WRITECOUNT
808 unsigned long f_mnt_write_state
;
815 /* file identifier */
816 unsigned char f_handle
[0];
819 static inline struct file
*get_file(struct file
*f
)
821 atomic_long_inc(&f
->f_count
);
824 #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
825 #define file_count(x) atomic_long_read(&(x)->f_count)
827 #ifdef CONFIG_DEBUG_WRITECOUNT
828 static inline void file_take_write(struct file
*f
)
830 WARN_ON(f
->f_mnt_write_state
!= 0);
831 f
->f_mnt_write_state
= FILE_MNT_WRITE_TAKEN
;
833 static inline void file_release_write(struct file
*f
)
835 f
->f_mnt_write_state
|= FILE_MNT_WRITE_RELEASED
;
837 static inline void file_reset_write(struct file
*f
)
839 f
->f_mnt_write_state
= 0;
841 static inline void file_check_state(struct file
*f
)
844 * At this point, either both or neither of these bits
847 WARN_ON(f
->f_mnt_write_state
== FILE_MNT_WRITE_TAKEN
);
848 WARN_ON(f
->f_mnt_write_state
== FILE_MNT_WRITE_RELEASED
);
850 static inline int file_check_writeable(struct file
*f
)
852 if (f
->f_mnt_write_state
== FILE_MNT_WRITE_TAKEN
)
854 printk(KERN_WARNING
"writeable file with no "
855 "mnt_want_write()\n");
859 #else /* !CONFIG_DEBUG_WRITECOUNT */
860 static inline void file_take_write(struct file
*filp
) {}
861 static inline void file_release_write(struct file
*filp
) {}
862 static inline void file_reset_write(struct file
*filp
) {}
863 static inline void file_check_state(struct file
*filp
) {}
864 static inline int file_check_writeable(struct file
*filp
)
868 #endif /* CONFIG_DEBUG_WRITECOUNT */
870 #define MAX_NON_LFS ((1UL<<31) - 1)
872 /* Page cache limit. The filesystems should put that into their s_maxbytes
873 limits, otherwise bad things can happen in VM. */
874 #if BITS_PER_LONG==32
875 #define MAX_LFS_FILESIZE (((loff_t)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
876 #elif BITS_PER_LONG==64
877 #define MAX_LFS_FILESIZE ((loff_t)0x7fffffffffffffffLL)
882 #define FL_ACCESS 8 /* not trying to lock, just looking */
883 #define FL_EXISTS 16 /* when unlocking, test for existence */
884 #define FL_LEASE 32 /* lease held on this file */
885 #define FL_CLOSE 64 /* unlock on close */
886 #define FL_SLEEP 128 /* A blocking lock */
887 #define FL_DOWNGRADE_PENDING 256 /* Lease is being downgraded */
888 #define FL_UNLOCK_PENDING 512 /* Lease is being broken */
891 * Special return value from posix_lock_file() and vfs_lock_file() for
892 * asynchronous locking.
894 #define FILE_LOCK_DEFERRED 1
897 * The POSIX file lock owner is determined by
898 * the "struct files_struct" in the thread group
899 * (or NULL for no owner - BSD locks).
901 * Lockd stuffs a "host" pointer into this.
903 typedef struct files_struct
*fl_owner_t
;
905 struct file_lock_operations
{
906 void (*fl_copy_lock
)(struct file_lock
*, struct file_lock
*);
907 void (*fl_release_private
)(struct file_lock
*);
910 struct lock_manager_operations
{
911 int (*lm_compare_owner
)(struct file_lock
*, struct file_lock
*);
912 unsigned long (*lm_owner_key
)(struct file_lock
*);
913 void (*lm_notify
)(struct file_lock
*); /* unblock callback */
914 int (*lm_grant
)(struct file_lock
*, struct file_lock
*, int);
915 void (*lm_break
)(struct file_lock
*);
916 int (*lm_change
)(struct file_lock
**, int);
919 struct lock_manager
{
920 struct list_head list
;
924 void locks_start_grace(struct net
*, struct lock_manager
*);
925 void locks_end_grace(struct lock_manager
*);
926 int locks_in_grace(struct net
*);
928 /* that will die - we need it for nfs_lock_info */
929 #include <linux/nfs_fs_i.h>
932 * struct file_lock represents a generic "file lock". It's used to represent
933 * POSIX byte range locks, BSD (flock) locks, and leases. It's important to
934 * note that the same struct is used to represent both a request for a lock and
935 * the lock itself, but the same object is never used for both.
937 * FIXME: should we create a separate "struct lock_request" to help distinguish
940 * The i_flock list is ordered by:
942 * 1) lock type -- FL_LEASEs first, then FL_FLOCK, and finally FL_POSIX
944 * 3) lock range start
947 * Obviously, the last two criteria only matter for POSIX locks.
950 struct file_lock
*fl_next
; /* singly linked list for this inode */
951 struct hlist_node fl_link
; /* node in global lists */
952 struct list_head fl_block
; /* circular list of blocked processes */
954 unsigned int fl_flags
;
955 unsigned char fl_type
;
957 int fl_link_cpu
; /* what cpu's list is this on? */
958 struct pid
*fl_nspid
;
959 wait_queue_head_t fl_wait
;
960 struct file
*fl_file
;
964 struct fasync_struct
* fl_fasync
; /* for lease break notifications */
965 /* for lease breaks: */
966 unsigned long fl_break_time
;
967 unsigned long fl_downgrade_time
;
969 const struct file_lock_operations
*fl_ops
; /* Callbacks for filesystems */
970 const struct lock_manager_operations
*fl_lmops
; /* Callbacks for lockmanagers */
972 struct nfs_lock_info nfs_fl
;
973 struct nfs4_lock_info nfs4_fl
;
975 struct list_head link
; /* link in AFS vnode's pending_locks list */
976 int state
; /* state of grant or error if -ve */
981 /* The following constant reflects the upper bound of the file/locking space */
983 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
984 #define OFFSET_MAX INT_LIMIT(loff_t)
985 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
988 #include <linux/fcntl.h>
990 extern void send_sigio(struct fown_struct
*fown
, int fd
, int band
);
992 #ifdef CONFIG_FILE_LOCKING
993 extern int fcntl_getlk(struct file
*, struct flock __user
*);
994 extern int fcntl_setlk(unsigned int, struct file
*, unsigned int,
995 struct flock __user
*);
997 #if BITS_PER_LONG == 32
998 extern int fcntl_getlk64(struct file
*, struct flock64 __user
*);
999 extern int fcntl_setlk64(unsigned int, struct file
*, unsigned int,
1000 struct flock64 __user
*);
1003 extern int fcntl_setlease(unsigned int fd
, struct file
*filp
, long arg
);
1004 extern int fcntl_getlease(struct file
*filp
);
1007 void locks_free_lock(struct file_lock
*fl
);
1008 extern void locks_init_lock(struct file_lock
*);
1009 extern struct file_lock
* locks_alloc_lock(void);
1010 extern void locks_copy_lock(struct file_lock
*, struct file_lock
*);
1011 extern void __locks_copy_lock(struct file_lock
*, const struct file_lock
*);
1012 extern void locks_remove_posix(struct file
*, fl_owner_t
);
1013 extern void locks_remove_flock(struct file
*);
1014 extern void locks_release_private(struct file_lock
*);
1015 extern void posix_test_lock(struct file
*, struct file_lock
*);
1016 extern int posix_lock_file(struct file
*, struct file_lock
*, struct file_lock
*);
1017 extern int posix_lock_file_wait(struct file
*, struct file_lock
*);
1018 extern int posix_unblock_lock(struct file_lock
*);
1019 extern int vfs_test_lock(struct file
*, struct file_lock
*);
1020 extern int vfs_lock_file(struct file
*, unsigned int, struct file_lock
*, struct file_lock
*);
1021 extern int vfs_cancel_lock(struct file
*filp
, struct file_lock
*fl
);
1022 extern int flock_lock_file_wait(struct file
*filp
, struct file_lock
*fl
);
1023 extern int __break_lease(struct inode
*inode
, unsigned int flags
);
1024 extern void lease_get_mtime(struct inode
*, struct timespec
*time
);
1025 extern int generic_setlease(struct file
*, long, struct file_lock
**);
1026 extern int vfs_setlease(struct file
*, long, struct file_lock
**);
1027 extern int lease_modify(struct file_lock
**, int);
1028 extern int lock_may_read(struct inode
*, loff_t start
, unsigned long count
);
1029 extern int lock_may_write(struct inode
*, loff_t start
, unsigned long count
);
1030 #else /* !CONFIG_FILE_LOCKING */
1031 static inline int fcntl_getlk(struct file
*file
, struct flock __user
*user
)
1036 static inline int fcntl_setlk(unsigned int fd
, struct file
*file
,
1037 unsigned int cmd
, struct flock __user
*user
)
1042 #if BITS_PER_LONG == 32
1043 static inline int fcntl_getlk64(struct file
*file
, struct flock64 __user
*user
)
1048 static inline int fcntl_setlk64(unsigned int fd
, struct file
*file
,
1049 unsigned int cmd
, struct flock64 __user
*user
)
1054 static inline int fcntl_setlease(unsigned int fd
, struct file
*filp
, long arg
)
1059 static inline int fcntl_getlease(struct file
*filp
)
1064 static inline void locks_init_lock(struct file_lock
*fl
)
1069 static inline void __locks_copy_lock(struct file_lock
*new, struct file_lock
*fl
)
1074 static inline void locks_copy_lock(struct file_lock
*new, struct file_lock
*fl
)
1079 static inline void locks_remove_posix(struct file
*filp
, fl_owner_t owner
)
1084 static inline void locks_remove_flock(struct file
*filp
)
1089 static inline void posix_test_lock(struct file
*filp
, struct file_lock
*fl
)
1094 static inline int posix_lock_file(struct file
*filp
, struct file_lock
*fl
,
1095 struct file_lock
*conflock
)
1100 static inline int posix_lock_file_wait(struct file
*filp
, struct file_lock
*fl
)
1105 static inline int posix_unblock_lock(struct file_lock
*waiter
)
1110 static inline int vfs_test_lock(struct file
*filp
, struct file_lock
*fl
)
1115 static inline int vfs_lock_file(struct file
*filp
, unsigned int cmd
,
1116 struct file_lock
*fl
, struct file_lock
*conf
)
1121 static inline int vfs_cancel_lock(struct file
*filp
, struct file_lock
*fl
)
1126 static inline int flock_lock_file_wait(struct file
*filp
,
1127 struct file_lock
*request
)
1132 static inline int __break_lease(struct inode
*inode
, unsigned int mode
)
1137 static inline void lease_get_mtime(struct inode
*inode
, struct timespec
*time
)
1142 static inline int generic_setlease(struct file
*filp
, long arg
,
1143 struct file_lock
**flp
)
1148 static inline int vfs_setlease(struct file
*filp
, long arg
,
1149 struct file_lock
**lease
)
1154 static inline int lease_modify(struct file_lock
**before
, int arg
)
1159 static inline int lock_may_read(struct inode
*inode
, loff_t start
,
1165 static inline int lock_may_write(struct inode
*inode
, loff_t start
,
1170 #endif /* !CONFIG_FILE_LOCKING */
1173 struct fasync_struct
{
1177 struct fasync_struct
*fa_next
; /* singly linked list */
1178 struct file
*fa_file
;
1179 struct rcu_head fa_rcu
;
1182 #define FASYNC_MAGIC 0x4601
1184 /* SMP safe fasync helpers: */
1185 extern int fasync_helper(int, struct file
*, int, struct fasync_struct
**);
1186 extern struct fasync_struct
*fasync_insert_entry(int, struct file
*, struct fasync_struct
**, struct fasync_struct
*);
1187 extern int fasync_remove_entry(struct file
*, struct fasync_struct
**);
1188 extern struct fasync_struct
*fasync_alloc(void);
1189 extern void fasync_free(struct fasync_struct
*);
1191 /* can be called from interrupts */
1192 extern void kill_fasync(struct fasync_struct
**, int, int);
1194 extern int __f_setown(struct file
*filp
, struct pid
*, enum pid_type
, int force
);
1195 extern int f_setown(struct file
*filp
, unsigned long arg
, int force
);
1196 extern void f_delown(struct file
*filp
);
1197 extern pid_t
f_getown(struct file
*filp
);
1198 extern int send_sigurg(struct fown_struct
*fown
);
1206 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1207 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1208 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1209 #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1210 #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
1212 extern struct list_head super_blocks
;
1213 extern spinlock_t sb_lock
;
1215 /* Possible states of 'frozen' field */
1217 SB_UNFROZEN
= 0, /* FS is unfrozen */
1218 SB_FREEZE_WRITE
= 1, /* Writes, dir ops, ioctls frozen */
1219 SB_FREEZE_PAGEFAULT
= 2, /* Page faults stopped as well */
1220 SB_FREEZE_FS
= 3, /* For internal FS use (e.g. to stop
1221 * internal threads if needed) */
1222 SB_FREEZE_COMPLETE
= 4, /* ->freeze_fs finished successfully */
1225 #define SB_FREEZE_LEVELS (SB_FREEZE_COMPLETE - 1)
1228 /* Counters for counting writers at each level */
1229 struct percpu_counter counter
[SB_FREEZE_LEVELS
];
1230 wait_queue_head_t wait
; /* queue for waiting for
1231 writers / faults to finish */
1232 int frozen
; /* Is sb frozen? */
1233 wait_queue_head_t wait_unfrozen
; /* queue for waiting for
1235 #ifdef CONFIG_DEBUG_LOCK_ALLOC
1236 struct lockdep_map lock_map
[SB_FREEZE_LEVELS
];
1240 struct super_block
{
1241 struct list_head s_list
; /* Keep this first */
1242 dev_t s_dev
; /* search index; _not_ kdev_t */
1243 unsigned char s_blocksize_bits
;
1244 unsigned long s_blocksize
;
1245 loff_t s_maxbytes
; /* Max file size */
1246 struct file_system_type
*s_type
;
1247 const struct super_operations
*s_op
;
1248 const struct dquot_operations
*dq_op
;
1249 const struct quotactl_ops
*s_qcop
;
1250 const struct export_operations
*s_export_op
;
1251 unsigned long s_flags
;
1252 unsigned long s_magic
;
1253 struct dentry
*s_root
;
1254 struct rw_semaphore s_umount
;
1257 #ifdef CONFIG_SECURITY
1260 const struct xattr_handler
**s_xattr
;
1262 struct list_head s_inodes
; /* all inodes */
1263 struct hlist_bl_head s_anon
; /* anonymous dentries for (nfs) exporting */
1265 struct list_head __percpu
*s_files
;
1267 struct list_head s_files
;
1269 struct list_head s_mounts
; /* list of mounts; _not_ for fs use */
1270 /* s_dentry_lru, s_nr_dentry_unused protected by dcache.c lru locks */
1271 struct list_head s_dentry_lru
; /* unused dentry lru */
1272 int s_nr_dentry_unused
; /* # of dentry on lru */
1274 /* s_inode_lru_lock protects s_inode_lru and s_nr_inodes_unused */
1275 spinlock_t s_inode_lru_lock ____cacheline_aligned_in_smp
;
1276 struct list_head s_inode_lru
; /* unused inode lru */
1277 int s_nr_inodes_unused
; /* # of inodes on lru */
1279 struct block_device
*s_bdev
;
1280 struct backing_dev_info
*s_bdi
;
1281 struct mtd_info
*s_mtd
;
1282 struct hlist_node s_instances
;
1283 struct quota_info s_dquot
; /* Diskquota specific options */
1285 struct sb_writers s_writers
;
1287 char s_id
[32]; /* Informational name */
1288 u8 s_uuid
[16]; /* UUID */
1290 void *s_fs_info
; /* Filesystem private info */
1291 unsigned int s_max_links
;
1294 /* Granularity of c/m/atime in ns.
1295 Cannot be worse than a second */
1299 * The next field is for VFS *only*. No filesystems have any business
1300 * even looking at it. You had been warned.
1302 struct mutex s_vfs_rename_mutex
; /* Kludge */
1305 * Filesystem subtype. If non-empty the filesystem type field
1306 * in /proc/mounts will be "type.subtype"
1311 * Saved mount options for lazy filesystems using
1312 * generic_show_options()
1314 char __rcu
*s_options
;
1315 const struct dentry_operations
*s_d_op
; /* default d_op for dentries */
1318 * Saved pool identifier for cleancache (-1 means none)
1320 int cleancache_poolid
;
1322 struct shrinker s_shrink
; /* per-sb shrinker handle */
1324 /* Number of inodes with nlink == 0 but still referenced */
1325 atomic_long_t s_remove_count
;
1327 /* Being remounted read-only */
1328 int s_readonly_remount
;
1331 /* superblock cache pruning functions */
1332 extern void prune_icache_sb(struct super_block
*sb
, int nr_to_scan
);
1333 extern void prune_dcache_sb(struct super_block
*sb
, int nr_to_scan
);
1335 extern struct timespec
current_fs_time(struct super_block
*sb
);
1338 * Snapshotting support.
1341 void __sb_end_write(struct super_block
*sb
, int level
);
1342 int __sb_start_write(struct super_block
*sb
, int level
, bool wait
);
1345 * sb_end_write - drop write access to a superblock
1346 * @sb: the super we wrote to
1348 * Decrement number of writers to the filesystem. Wake up possible waiters
1349 * wanting to freeze the filesystem.
1351 static inline void sb_end_write(struct super_block
*sb
)
1353 __sb_end_write(sb
, SB_FREEZE_WRITE
);
1357 * sb_end_pagefault - drop write access to a superblock from a page fault
1358 * @sb: the super we wrote to
1360 * Decrement number of processes handling write page fault to the filesystem.
1361 * Wake up possible waiters wanting to freeze the filesystem.
1363 static inline void sb_end_pagefault(struct super_block
*sb
)
1365 __sb_end_write(sb
, SB_FREEZE_PAGEFAULT
);
1369 * sb_end_intwrite - drop write access to a superblock for internal fs purposes
1370 * @sb: the super we wrote to
1372 * Decrement fs-internal number of writers to the filesystem. Wake up possible
1373 * waiters wanting to freeze the filesystem.
1375 static inline void sb_end_intwrite(struct super_block
*sb
)
1377 __sb_end_write(sb
, SB_FREEZE_FS
);
1381 * sb_start_write - get write access to a superblock
1382 * @sb: the super we write to
1384 * When a process wants to write data or metadata to a file system (i.e. dirty
1385 * a page or an inode), it should embed the operation in a sb_start_write() -
1386 * sb_end_write() pair to get exclusion against file system freezing. This
1387 * function increments number of writers preventing freezing. If the file
1388 * system is already frozen, the function waits until the file system is
1391 * Since freeze protection behaves as a lock, users have to preserve
1392 * ordering of freeze protection and other filesystem locks. Generally,
1393 * freeze protection should be the outermost lock. In particular, we have:
1396 * -> i_mutex (write path, truncate, directory ops, ...)
1397 * -> s_umount (freeze_super, thaw_super)
1399 static inline void sb_start_write(struct super_block
*sb
)
1401 __sb_start_write(sb
, SB_FREEZE_WRITE
, true);
1404 static inline int sb_start_write_trylock(struct super_block
*sb
)
1406 return __sb_start_write(sb
, SB_FREEZE_WRITE
, false);
1410 * sb_start_pagefault - get write access to a superblock from a page fault
1411 * @sb: the super we write to
1413 * When a process starts handling write page fault, it should embed the
1414 * operation into sb_start_pagefault() - sb_end_pagefault() pair to get
1415 * exclusion against file system freezing. This is needed since the page fault
1416 * is going to dirty a page. This function increments number of running page
1417 * faults preventing freezing. If the file system is already frozen, the
1418 * function waits until the file system is thawed.
1420 * Since page fault freeze protection behaves as a lock, users have to preserve
1421 * ordering of freeze protection and other filesystem locks. It is advised to
1422 * put sb_start_pagefault() close to mmap_sem in lock ordering. Page fault
1423 * handling code implies lock dependency:
1426 * -> sb_start_pagefault
1428 static inline void sb_start_pagefault(struct super_block
*sb
)
1430 __sb_start_write(sb
, SB_FREEZE_PAGEFAULT
, true);
1434 * sb_start_intwrite - get write access to a superblock for internal fs purposes
1435 * @sb: the super we write to
1437 * This is the third level of protection against filesystem freezing. It is
1438 * free for use by a filesystem. The only requirement is that it must rank
1439 * below sb_start_pagefault.
1441 * For example filesystem can call sb_start_intwrite() when starting a
1442 * transaction which somewhat eases handling of freezing for internal sources
1443 * of filesystem changes (internal fs threads, discarding preallocation on file
1446 static inline void sb_start_intwrite(struct super_block
*sb
)
1448 __sb_start_write(sb
, SB_FREEZE_FS
, true);
1452 extern bool inode_owner_or_capable(const struct inode
*inode
);
1455 * VFS helper functions..
1457 extern int vfs_create(struct inode
*, struct dentry
*, umode_t
, bool);
1458 extern int vfs_mkdir(struct inode
*, struct dentry
*, umode_t
);
1459 extern int vfs_mknod(struct inode
*, struct dentry
*, umode_t
, dev_t
);
1460 extern int vfs_symlink(struct inode
*, struct dentry
*, const char *);
1461 extern int vfs_link(struct dentry
*, struct inode
*, struct dentry
*);
1462 extern int vfs_rmdir(struct inode
*, struct dentry
*);
1463 extern int vfs_unlink(struct inode
*, struct dentry
*);
1464 extern int vfs_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
1467 * VFS dentry helper functions.
1469 extern void dentry_unhash(struct dentry
*dentry
);
1472 * VFS file helper functions.
1474 extern void inode_init_owner(struct inode
*inode
, const struct inode
*dir
,
1477 * VFS FS_IOC_FIEMAP helper definitions.
1479 struct fiemap_extent_info
{
1480 unsigned int fi_flags
; /* Flags as passed from user */
1481 unsigned int fi_extents_mapped
; /* Number of mapped extents */
1482 unsigned int fi_extents_max
; /* Size of fiemap_extent array */
1483 struct fiemap_extent __user
*fi_extents_start
; /* Start of
1484 fiemap_extent array */
1486 int fiemap_fill_next_extent(struct fiemap_extent_info
*info
, u64 logical
,
1487 u64 phys
, u64 len
, u32 flags
);
1488 int fiemap_check_flags(struct fiemap_extent_info
*fieinfo
, u32 fs_flags
);
1493 * NOTE! These match bits 12..15 of stat.st_mode
1494 * (ie "(i_mode >> 12) & 15").
1496 #define DT_UNKNOWN 0
1507 * This is the "filldir" function type, used by readdir() to let
1508 * the kernel specify what kind of dirent layout it wants to have.
1509 * This allows the kernel to read directories into kernel space or
1510 * to have different dirent layouts depending on the binary type.
1512 typedef int (*filldir_t
)(void *, const char *, int, loff_t
, u64
, unsigned);
1513 struct dir_context
{
1514 const filldir_t actor
;
1518 struct block_device_operations
;
1520 /* These macros are for out of kernel modules to test that
1521 * the kernel supports the unlocked_ioctl and compat_ioctl
1522 * fields in struct file_operations. */
1523 #define HAVE_COMPAT_IOCTL 1
1524 #define HAVE_UNLOCKED_IOCTL 1
1526 struct file_operations
{
1527 struct module
*owner
;
1528 loff_t (*llseek
) (struct file
*, loff_t
, int);
1529 ssize_t (*read
) (struct file
*, char __user
*, size_t, loff_t
*);
1530 ssize_t (*write
) (struct file
*, const char __user
*, size_t, loff_t
*);
1531 ssize_t (*aio_read
) (struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1532 ssize_t (*aio_write
) (struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1533 int (*iterate
) (struct file
*, struct dir_context
*);
1534 unsigned int (*poll
) (struct file
*, struct poll_table_struct
*);
1535 long (*unlocked_ioctl
) (struct file
*, unsigned int, unsigned long);
1536 long (*compat_ioctl
) (struct file
*, unsigned int, unsigned long);
1537 int (*mmap
) (struct file
*, struct vm_area_struct
*);
1538 int (*open
) (struct inode
*, struct file
*);
1539 int (*flush
) (struct file
*, fl_owner_t id
);
1540 int (*release
) (struct inode
*, struct file
*);
1541 int (*fsync
) (struct file
*, loff_t
, loff_t
, int datasync
);
1542 int (*aio_fsync
) (struct kiocb
*, int datasync
);
1543 int (*fasync
) (int, struct file
*, int);
1544 int (*lock
) (struct file
*, int, struct file_lock
*);
1545 ssize_t (*sendpage
) (struct file
*, struct page
*, int, size_t, loff_t
*, int);
1546 unsigned long (*get_unmapped_area
)(struct file
*, unsigned long, unsigned long, unsigned long, unsigned long);
1547 int (*check_flags
)(int);
1548 int (*flock
) (struct file
*, int, struct file_lock
*);
1549 ssize_t (*splice_write
)(struct pipe_inode_info
*, struct file
*, loff_t
*, size_t, unsigned int);
1550 ssize_t (*splice_read
)(struct file
*, loff_t
*, struct pipe_inode_info
*, size_t, unsigned int);
1551 int (*setlease
)(struct file
*, long, struct file_lock
**);
1552 long (*fallocate
)(struct file
*file
, int mode
, loff_t offset
,
1554 int (*show_fdinfo
)(struct seq_file
*m
, struct file
*f
);
1557 struct inode_operations
{
1558 struct dentry
* (*lookup
) (struct inode
*,struct dentry
*, unsigned int);
1559 void * (*follow_link
) (struct dentry
*, struct nameidata
*);
1560 int (*permission
) (struct inode
*, int);
1561 struct posix_acl
* (*get_acl
)(struct inode
*, int);
1563 int (*readlink
) (struct dentry
*, char __user
*,int);
1564 void (*put_link
) (struct dentry
*, struct nameidata
*, void *);
1566 int (*create
) (struct inode
*,struct dentry
*, umode_t
, bool);
1567 int (*link
) (struct dentry
*,struct inode
*,struct dentry
*);
1568 int (*unlink
) (struct inode
*,struct dentry
*);
1569 int (*symlink
) (struct inode
*,struct dentry
*,const char *);
1570 int (*mkdir
) (struct inode
*,struct dentry
*,umode_t
);
1571 int (*rmdir
) (struct inode
*,struct dentry
*);
1572 int (*mknod
) (struct inode
*,struct dentry
*,umode_t
,dev_t
);
1573 int (*rename
) (struct inode
*, struct dentry
*,
1574 struct inode
*, struct dentry
*);
1575 int (*setattr
) (struct dentry
*, struct iattr
*);
1576 int (*getattr
) (struct vfsmount
*mnt
, struct dentry
*, struct kstat
*);
1577 int (*setxattr
) (struct dentry
*, const char *,const void *,size_t,int);
1578 ssize_t (*getxattr
) (struct dentry
*, const char *, void *, size_t);
1579 ssize_t (*listxattr
) (struct dentry
*, char *, size_t);
1580 int (*removexattr
) (struct dentry
*, const char *);
1581 int (*fiemap
)(struct inode
*, struct fiemap_extent_info
*, u64 start
,
1583 int (*update_time
)(struct inode
*, struct timespec
*, int);
1584 int (*atomic_open
)(struct inode
*, struct dentry
*,
1585 struct file
*, unsigned open_flag
,
1586 umode_t create_mode
, int *opened
);
1587 int (*tmpfile
) (struct inode
*, struct dentry
*, umode_t
);
1588 } ____cacheline_aligned
;
1590 ssize_t
rw_copy_check_uvector(int type
, const struct iovec __user
* uvector
,
1591 unsigned long nr_segs
, unsigned long fast_segs
,
1592 struct iovec
*fast_pointer
,
1593 struct iovec
**ret_pointer
);
1595 extern ssize_t
vfs_read(struct file
*, char __user
*, size_t, loff_t
*);
1596 extern ssize_t
vfs_write(struct file
*, const char __user
*, size_t, loff_t
*);
1597 extern ssize_t
vfs_readv(struct file
*, const struct iovec __user
*,
1598 unsigned long, loff_t
*);
1599 extern ssize_t
vfs_writev(struct file
*, const struct iovec __user
*,
1600 unsigned long, loff_t
*);
1602 struct super_operations
{
1603 struct inode
*(*alloc_inode
)(struct super_block
*sb
);
1604 void (*destroy_inode
)(struct inode
*);
1606 void (*dirty_inode
) (struct inode
*, int flags
);
1607 int (*write_inode
) (struct inode
*, struct writeback_control
*wbc
);
1608 int (*drop_inode
) (struct inode
*);
1609 void (*evict_inode
) (struct inode
*);
1610 void (*put_super
) (struct super_block
*);
1611 int (*sync_fs
)(struct super_block
*sb
, int wait
);
1612 int (*freeze_fs
) (struct super_block
*);
1613 int (*unfreeze_fs
) (struct super_block
*);
1614 int (*statfs
) (struct dentry
*, struct kstatfs
*);
1615 int (*remount_fs
) (struct super_block
*, int *, char *);
1616 void (*umount_begin
) (struct super_block
*);
1618 int (*show_options
)(struct seq_file
*, struct dentry
*);
1619 int (*show_devname
)(struct seq_file
*, struct dentry
*);
1620 int (*show_path
)(struct seq_file
*, struct dentry
*);
1621 int (*show_stats
)(struct seq_file
*, struct dentry
*);
1623 ssize_t (*quota_read
)(struct super_block
*, int, char *, size_t, loff_t
);
1624 ssize_t (*quota_write
)(struct super_block
*, int, const char *, size_t, loff_t
);
1626 int (*bdev_try_to_free_page
)(struct super_block
*, struct page
*, gfp_t
);
1627 int (*nr_cached_objects
)(struct super_block
*);
1628 void (*free_cached_objects
)(struct super_block
*, int);
1632 * Inode flags - they have no relation to superblock flags now
1634 #define S_SYNC 1 /* Writes are synced at once */
1635 #define S_NOATIME 2 /* Do not update access times */
1636 #define S_APPEND 4 /* Append-only file */
1637 #define S_IMMUTABLE 8 /* Immutable file */
1638 #define S_DEAD 16 /* removed, but still open directory */
1639 #define S_NOQUOTA 32 /* Inode is not counted to quota */
1640 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
1641 #define S_NOCMTIME 128 /* Do not update file c/mtime */
1642 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
1643 #define S_PRIVATE 512 /* Inode is fs-internal */
1644 #define S_IMA 1024 /* Inode has an associated IMA struct */
1645 #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
1646 #define S_NOSEC 4096 /* no suid or xattr security attributes */
1649 * Note that nosuid etc flags are inode-specific: setting some file-system
1650 * flags just means all the inodes inherit those flags by default. It might be
1651 * possible to override it selectively if you really wanted to with some
1652 * ioctl() that is not currently implemented.
1654 * Exception: MS_RDONLY is always applied to the entire file system.
1656 * Unfortunately, it is possible to change a filesystems flags with it mounted
1657 * with files in use. This means that all of the inodes will not have their
1658 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
1659 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
1661 #define __IS_FLG(inode, flg) ((inode)->i_sb->s_flags & (flg))
1663 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
1664 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
1665 ((inode)->i_flags & S_SYNC))
1666 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
1667 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
1668 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
1669 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
1670 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
1672 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
1673 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
1674 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
1675 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
1677 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
1678 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
1679 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
1680 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
1681 #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
1682 #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
1683 #define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
1686 * Inode state bits. Protected by inode->i_lock
1688 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1689 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1691 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1692 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1693 * various stages of removing an inode.
1695 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1697 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1698 * fdatasync(). i_atime is the usual cause.
1699 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1700 * these changes separately from I_DIRTY_SYNC so that we
1701 * don't have to write inode on fdatasync() when only
1702 * mtime has changed in it.
1703 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1704 * I_NEW Serves as both a mutex and completion notification.
1705 * New inodes set I_NEW. If two processes both create
1706 * the same inode, one of them will release its inode and
1707 * wait for I_NEW to be released before returning.
1708 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1709 * also cause waiting on I_NEW, without I_NEW actually
1710 * being set. find_inode() uses this to prevent returning
1711 * nearly-dead inodes.
1712 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1713 * is zero. I_FREEING must be set when I_WILL_FREE is
1715 * I_FREEING Set when inode is about to be freed but still has dirty
1716 * pages or buffers attached or the inode itself is still
1718 * I_CLEAR Added by clear_inode(). In this state the inode is
1719 * clean and can be destroyed. Inode keeps I_FREEING.
1721 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1722 * prohibited for many purposes. iget() must wait for
1723 * the inode to be completely released, then create it
1724 * anew. Other functions will just ignore such inodes,
1725 * if appropriate. I_NEW is used for waiting.
1727 * I_SYNC Writeback of inode is running. The bit is set during
1728 * data writeback, and cleared with a wakeup on the bit
1729 * address once it is done. The bit is also used to pin
1730 * the inode in memory for flusher thread.
1732 * I_REFERENCED Marks the inode as recently references on the LRU list.
1734 * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit().
1736 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1738 #define I_DIRTY_SYNC (1 << 0)
1739 #define I_DIRTY_DATASYNC (1 << 1)
1740 #define I_DIRTY_PAGES (1 << 2)
1742 #define I_NEW (1 << __I_NEW)
1743 #define I_WILL_FREE (1 << 4)
1744 #define I_FREEING (1 << 5)
1745 #define I_CLEAR (1 << 6)
1747 #define I_SYNC (1 << __I_SYNC)
1748 #define I_REFERENCED (1 << 8)
1749 #define __I_DIO_WAKEUP 9
1750 #define I_DIO_WAKEUP (1 << I_DIO_WAKEUP)
1751 #define I_LINKABLE (1 << 10)
1753 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1755 extern void __mark_inode_dirty(struct inode
*, int);
1756 static inline void mark_inode_dirty(struct inode
*inode
)
1758 __mark_inode_dirty(inode
, I_DIRTY
);
1761 static inline void mark_inode_dirty_sync(struct inode
*inode
)
1763 __mark_inode_dirty(inode
, I_DIRTY_SYNC
);
1766 extern void inc_nlink(struct inode
*inode
);
1767 extern void drop_nlink(struct inode
*inode
);
1768 extern void clear_nlink(struct inode
*inode
);
1769 extern void set_nlink(struct inode
*inode
, unsigned int nlink
);
1771 static inline void inode_inc_link_count(struct inode
*inode
)
1774 mark_inode_dirty(inode
);
1777 static inline void inode_dec_link_count(struct inode
*inode
)
1780 mark_inode_dirty(inode
);
1784 * inode_inc_iversion - increments i_version
1785 * @inode: inode that need to be updated
1787 * Every time the inode is modified, the i_version field will be incremented.
1788 * The filesystem has to be mounted with i_version flag
1791 static inline void inode_inc_iversion(struct inode
*inode
)
1793 spin_lock(&inode
->i_lock
);
1795 spin_unlock(&inode
->i_lock
);
1798 enum file_time_flags
{
1805 extern void touch_atime(struct path
*);
1806 static inline void file_accessed(struct file
*file
)
1808 if (!(file
->f_flags
& O_NOATIME
))
1809 touch_atime(&file
->f_path
);
1812 int sync_inode(struct inode
*inode
, struct writeback_control
*wbc
);
1813 int sync_inode_metadata(struct inode
*inode
, int wait
);
1815 struct file_system_type
{
1818 #define FS_REQUIRES_DEV 1
1819 #define FS_BINARY_MOUNTDATA 2
1820 #define FS_HAS_SUBTYPE 4
1821 #define FS_USERNS_MOUNT 8 /* Can be mounted by userns root */
1822 #define FS_USERNS_DEV_MOUNT 16 /* A userns mount does not imply MNT_NODEV */
1823 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() during rename() internally. */
1824 struct dentry
*(*mount
) (struct file_system_type
*, int,
1825 const char *, void *);
1826 void (*kill_sb
) (struct super_block
*);
1827 struct module
*owner
;
1828 struct file_system_type
* next
;
1829 struct hlist_head fs_supers
;
1831 struct lock_class_key s_lock_key
;
1832 struct lock_class_key s_umount_key
;
1833 struct lock_class_key s_vfs_rename_key
;
1834 struct lock_class_key s_writers_key
[SB_FREEZE_LEVELS
];
1836 struct lock_class_key i_lock_key
;
1837 struct lock_class_key i_mutex_key
;
1838 struct lock_class_key i_mutex_dir_key
;
1841 #define MODULE_ALIAS_FS(NAME) MODULE_ALIAS("fs-" NAME)
1843 extern struct dentry
*mount_ns(struct file_system_type
*fs_type
, int flags
,
1844 void *data
, int (*fill_super
)(struct super_block
*, void *, int));
1845 extern struct dentry
*mount_bdev(struct file_system_type
*fs_type
,
1846 int flags
, const char *dev_name
, void *data
,
1847 int (*fill_super
)(struct super_block
*, void *, int));
1848 extern struct dentry
*mount_single(struct file_system_type
*fs_type
,
1849 int flags
, void *data
,
1850 int (*fill_super
)(struct super_block
*, void *, int));
1851 extern struct dentry
*mount_nodev(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_subtree(struct vfsmount
*mnt
, const char *path
);
1855 void generic_shutdown_super(struct super_block
*sb
);
1856 void kill_block_super(struct super_block
*sb
);
1857 void kill_anon_super(struct super_block
*sb
);
1858 void kill_litter_super(struct super_block
*sb
);
1859 void deactivate_super(struct super_block
*sb
);
1860 void deactivate_locked_super(struct super_block
*sb
);
1861 int set_anon_super(struct super_block
*s
, void *data
);
1862 int get_anon_bdev(dev_t
*);
1863 void free_anon_bdev(dev_t
);
1864 struct super_block
*sget(struct file_system_type
*type
,
1865 int (*test
)(struct super_block
*,void *),
1866 int (*set
)(struct super_block
*,void *),
1867 int flags
, void *data
);
1868 extern struct dentry
*mount_pseudo(struct file_system_type
*, char *,
1869 const struct super_operations
*ops
,
1870 const struct dentry_operations
*dops
,
1873 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1874 #define fops_get(fops) \
1875 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1876 #define fops_put(fops) \
1877 do { if (fops) module_put((fops)->owner); } while(0)
1879 extern int register_filesystem(struct file_system_type
*);
1880 extern int unregister_filesystem(struct file_system_type
*);
1881 extern struct vfsmount
*kern_mount_data(struct file_system_type
*, void *data
);
1882 #define kern_mount(type) kern_mount_data(type, NULL)
1883 extern void kern_unmount(struct vfsmount
*mnt
);
1884 extern int may_umount_tree(struct vfsmount
*);
1885 extern int may_umount(struct vfsmount
*);
1886 extern long do_mount(const char *, const char *, const char *, unsigned long, void *);
1887 extern struct vfsmount
*collect_mounts(struct path
*);
1888 extern void drop_collected_mounts(struct vfsmount
*);
1889 extern int iterate_mounts(int (*)(struct vfsmount
*, void *), void *,
1891 extern int vfs_statfs(struct path
*, struct kstatfs
*);
1892 extern int user_statfs(const char __user
*, struct kstatfs
*);
1893 extern int fd_statfs(int, struct kstatfs
*);
1894 extern int vfs_ustat(dev_t
, struct kstatfs
*);
1895 extern int freeze_super(struct super_block
*super
);
1896 extern int thaw_super(struct super_block
*super
);
1897 extern bool our_mnt(struct vfsmount
*mnt
);
1899 extern int current_umask(void);
1902 extern struct kobject
*fs_kobj
;
1904 #define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
1906 #define FLOCK_VERIFY_READ 1
1907 #define FLOCK_VERIFY_WRITE 2
1909 #ifdef CONFIG_FILE_LOCKING
1910 extern int locks_mandatory_locked(struct inode
*);
1911 extern int locks_mandatory_area(int, struct inode
*, struct file
*, loff_t
, size_t);
1914 * Candidates for mandatory locking have the setgid bit set
1915 * but no group execute bit - an otherwise meaningless combination.
1918 static inline int __mandatory_lock(struct inode
*ino
)
1920 return (ino
->i_mode
& (S_ISGID
| S_IXGRP
)) == S_ISGID
;
1924 * ... and these candidates should be on MS_MANDLOCK mounted fs,
1925 * otherwise these will be advisory locks
1928 static inline int mandatory_lock(struct inode
*ino
)
1930 return IS_MANDLOCK(ino
) && __mandatory_lock(ino
);
1933 static inline int locks_verify_locked(struct inode
*inode
)
1935 if (mandatory_lock(inode
))
1936 return locks_mandatory_locked(inode
);
1940 static inline int locks_verify_truncate(struct inode
*inode
,
1944 if (inode
->i_flock
&& mandatory_lock(inode
))
1945 return locks_mandatory_area(
1946 FLOCK_VERIFY_WRITE
, inode
, filp
,
1947 size
< inode
->i_size
? size
: inode
->i_size
,
1948 (size
< inode
->i_size
? inode
->i_size
- size
1949 : size
- inode
->i_size
)
1954 static inline int break_lease(struct inode
*inode
, unsigned int mode
)
1957 return __break_lease(inode
, mode
);
1960 #else /* !CONFIG_FILE_LOCKING */
1961 static inline int locks_mandatory_locked(struct inode
*inode
)
1966 static inline int locks_mandatory_area(int rw
, struct inode
*inode
,
1967 struct file
*filp
, loff_t offset
,
1973 static inline int __mandatory_lock(struct inode
*inode
)
1978 static inline int mandatory_lock(struct inode
*inode
)
1983 static inline int locks_verify_locked(struct inode
*inode
)
1988 static inline int locks_verify_truncate(struct inode
*inode
, struct file
*filp
,
1994 static inline int break_lease(struct inode
*inode
, unsigned int mode
)
1999 #endif /* CONFIG_FILE_LOCKING */
2004 const char *name
; /* pointer to actual string */
2005 const __user
char *uptr
; /* original userland pointer */
2006 struct audit_names
*aname
;
2007 bool separate
; /* should "name" be freed? */
2010 extern long vfs_truncate(struct path
*, loff_t
);
2011 extern int do_truncate(struct dentry
*, loff_t start
, unsigned int time_attrs
,
2013 extern int do_fallocate(struct file
*file
, int mode
, loff_t offset
,
2015 extern long do_sys_open(int dfd
, const char __user
*filename
, int flags
,
2017 extern struct file
*file_open_name(struct filename
*, int, umode_t
);
2018 extern struct file
*filp_open(const char *, int, umode_t
);
2019 extern struct file
*file_open_root(struct dentry
*, struct vfsmount
*,
2021 extern struct file
* dentry_open(const struct path
*, int, const struct cred
*);
2022 extern int filp_close(struct file
*, fl_owner_t id
);
2024 extern struct filename
*getname(const char __user
*);
2030 extern int finish_open(struct file
*file
, struct dentry
*dentry
,
2031 int (*open
)(struct inode
*, struct file
*),
2033 extern int finish_no_open(struct file
*file
, struct dentry
*dentry
);
2037 extern int ioctl_preallocate(struct file
*filp
, void __user
*argp
);
2040 extern void __init
vfs_caches_init_early(void);
2041 extern void __init
vfs_caches_init(unsigned long);
2043 extern struct kmem_cache
*names_cachep
;
2045 extern void final_putname(struct filename
*name
);
2047 #define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL)
2048 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
2049 #ifndef CONFIG_AUDITSYSCALL
2050 #define putname(name) final_putname(name)
2052 extern void putname(struct filename
*name
);
2056 extern int register_blkdev(unsigned int, const char *);
2057 extern void unregister_blkdev(unsigned int, const char *);
2058 extern struct block_device
*bdget(dev_t
);
2059 extern struct block_device
*bdgrab(struct block_device
*bdev
);
2060 extern void bd_set_size(struct block_device
*, loff_t size
);
2061 extern void bd_forget(struct inode
*inode
);
2062 extern void bdput(struct block_device
*);
2063 extern void invalidate_bdev(struct block_device
*);
2064 extern void iterate_bdevs(void (*)(struct block_device
*, void *), void *);
2065 extern int sync_blockdev(struct block_device
*bdev
);
2066 extern void kill_bdev(struct block_device
*);
2067 extern struct super_block
*freeze_bdev(struct block_device
*);
2068 extern void emergency_thaw_all(void);
2069 extern int thaw_bdev(struct block_device
*bdev
, struct super_block
*sb
);
2070 extern int fsync_bdev(struct block_device
*);
2072 static inline void bd_forget(struct inode
*inode
) {}
2073 static inline int sync_blockdev(struct block_device
*bdev
) { return 0; }
2074 static inline void kill_bdev(struct block_device
*bdev
) {}
2075 static inline void invalidate_bdev(struct block_device
*bdev
) {}
2077 static inline struct super_block
*freeze_bdev(struct block_device
*sb
)
2082 static inline int thaw_bdev(struct block_device
*bdev
, struct super_block
*sb
)
2087 static inline void iterate_bdevs(void (*f
)(struct block_device
*, void *), void *arg
)
2091 extern int sync_filesystem(struct super_block
*);
2092 extern const struct file_operations def_blk_fops
;
2093 extern const struct file_operations def_chr_fops
;
2094 extern const struct file_operations bad_sock_fops
;
2096 extern int ioctl_by_bdev(struct block_device
*, unsigned, unsigned long);
2097 extern int blkdev_ioctl(struct block_device
*, fmode_t
, unsigned, unsigned long);
2098 extern long compat_blkdev_ioctl(struct file
*, unsigned, unsigned long);
2099 extern int blkdev_get(struct block_device
*bdev
, fmode_t mode
, void *holder
);
2100 extern struct block_device
*blkdev_get_by_path(const char *path
, fmode_t mode
,
2102 extern struct block_device
*blkdev_get_by_dev(dev_t dev
, fmode_t mode
,
2104 extern void blkdev_put(struct block_device
*bdev
, fmode_t mode
);
2106 extern int bd_link_disk_holder(struct block_device
*bdev
, struct gendisk
*disk
);
2107 extern void bd_unlink_disk_holder(struct block_device
*bdev
,
2108 struct gendisk
*disk
);
2110 static inline int bd_link_disk_holder(struct block_device
*bdev
,
2111 struct gendisk
*disk
)
2115 static inline void bd_unlink_disk_holder(struct block_device
*bdev
,
2116 struct gendisk
*disk
)
2123 #define CHRDEV_MAJOR_HASH_SIZE 255
2124 extern int alloc_chrdev_region(dev_t
*, unsigned, unsigned, const char *);
2125 extern int register_chrdev_region(dev_t
, unsigned, const char *);
2126 extern int __register_chrdev(unsigned int major
, unsigned int baseminor
,
2127 unsigned int count
, const char *name
,
2128 const struct file_operations
*fops
);
2129 extern void __unregister_chrdev(unsigned int major
, unsigned int baseminor
,
2130 unsigned int count
, const char *name
);
2131 extern void unregister_chrdev_region(dev_t
, unsigned);
2132 extern void chrdev_show(struct seq_file
*,off_t
);
2134 static inline int register_chrdev(unsigned int major
, const char *name
,
2135 const struct file_operations
*fops
)
2137 return __register_chrdev(major
, 0, 256, name
, fops
);
2140 static inline void unregister_chrdev(unsigned int major
, const char *name
)
2142 __unregister_chrdev(major
, 0, 256, name
);
2145 /* fs/block_dev.c */
2146 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
2147 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
2150 #define BLKDEV_MAJOR_HASH_SIZE 255
2151 extern const char *__bdevname(dev_t
, char *buffer
);
2152 extern const char *bdevname(struct block_device
*bdev
, char *buffer
);
2153 extern struct block_device
*lookup_bdev(const char *);
2154 extern void blkdev_show(struct seq_file
*,off_t
);
2157 #define BLKDEV_MAJOR_HASH_SIZE 0
2160 extern void init_special_inode(struct inode
*, umode_t
, dev_t
);
2162 /* Invalid inode operations -- fs/bad_inode.c */
2163 extern void make_bad_inode(struct inode
*);
2164 extern int is_bad_inode(struct inode
*);
2168 * return READ, READA, or WRITE
2170 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2173 * return data direction, READ or WRITE
2175 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
2177 extern void check_disk_size_change(struct gendisk
*disk
,
2178 struct block_device
*bdev
);
2179 extern int revalidate_disk(struct gendisk
*);
2180 extern int check_disk_change(struct block_device
*);
2181 extern int __invalidate_device(struct block_device
*, bool);
2182 extern int invalidate_partition(struct gendisk
*, int);
2184 unsigned long invalidate_mapping_pages(struct address_space
*mapping
,
2185 pgoff_t start
, pgoff_t end
);
2187 static inline void invalidate_remote_inode(struct inode
*inode
)
2189 if (S_ISREG(inode
->i_mode
) || S_ISDIR(inode
->i_mode
) ||
2190 S_ISLNK(inode
->i_mode
))
2191 invalidate_mapping_pages(inode
->i_mapping
, 0, -1);
2193 extern int invalidate_inode_pages2(struct address_space
*mapping
);
2194 extern int invalidate_inode_pages2_range(struct address_space
*mapping
,
2195 pgoff_t start
, pgoff_t end
);
2196 extern int write_inode_now(struct inode
*, int);
2197 extern int filemap_fdatawrite(struct address_space
*);
2198 extern int filemap_flush(struct address_space
*);
2199 extern int filemap_fdatawait(struct address_space
*);
2200 extern int filemap_fdatawait_range(struct address_space
*, loff_t lstart
,
2202 extern int filemap_write_and_wait(struct address_space
*mapping
);
2203 extern int filemap_write_and_wait_range(struct address_space
*mapping
,
2204 loff_t lstart
, loff_t lend
);
2205 extern int __filemap_fdatawrite_range(struct address_space
*mapping
,
2206 loff_t start
, loff_t end
, int sync_mode
);
2207 extern int filemap_fdatawrite_range(struct address_space
*mapping
,
2208 loff_t start
, loff_t end
);
2210 extern int vfs_fsync_range(struct file
*file
, loff_t start
, loff_t end
,
2212 extern int vfs_fsync(struct file
*file
, int datasync
);
2213 extern int generic_write_sync(struct file
*file
, loff_t pos
, loff_t count
);
2214 extern void emergency_sync(void);
2215 extern void emergency_remount(void);
2217 extern sector_t
bmap(struct inode
*, sector_t
);
2219 extern int notify_change(struct dentry
*, struct iattr
*);
2220 extern int inode_permission(struct inode
*, int);
2221 extern int generic_permission(struct inode
*, int);
2223 static inline bool execute_ok(struct inode
*inode
)
2225 return (inode
->i_mode
& S_IXUGO
) || S_ISDIR(inode
->i_mode
);
2228 static inline struct inode
*file_inode(struct file
*f
)
2233 static inline void file_start_write(struct file
*file
)
2235 if (!S_ISREG(file_inode(file
)->i_mode
))
2237 __sb_start_write(file_inode(file
)->i_sb
, SB_FREEZE_WRITE
, true);
2240 static inline bool file_start_write_trylock(struct file
*file
)
2242 if (!S_ISREG(file_inode(file
)->i_mode
))
2244 return __sb_start_write(file_inode(file
)->i_sb
, SB_FREEZE_WRITE
, false);
2247 static inline void file_end_write(struct file
*file
)
2249 if (!S_ISREG(file_inode(file
)->i_mode
))
2251 __sb_end_write(file_inode(file
)->i_sb
, SB_FREEZE_WRITE
);
2255 * get_write_access() gets write permission for a file.
2256 * put_write_access() releases this write permission.
2257 * This is used for regular files.
2258 * We cannot support write (and maybe mmap read-write shared) accesses and
2259 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2260 * can have the following values:
2261 * 0: no writers, no VM_DENYWRITE mappings
2262 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2263 * > 0: (i_writecount) users are writing to the file.
2265 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2266 * except for the cases where we don't hold i_writecount yet. Then we need to
2267 * use {get,deny}_write_access() - these functions check the sign and refuse
2268 * to do the change if sign is wrong.
2270 static inline int get_write_access(struct inode
*inode
)
2272 return atomic_inc_unless_negative(&inode
->i_writecount
) ? 0 : -ETXTBSY
;
2274 static inline int deny_write_access(struct file
*file
)
2276 struct inode
*inode
= file_inode(file
);
2277 return atomic_dec_unless_positive(&inode
->i_writecount
) ? 0 : -ETXTBSY
;
2279 static inline void put_write_access(struct inode
* inode
)
2281 atomic_dec(&inode
->i_writecount
);
2283 static inline void allow_write_access(struct file
*file
)
2286 atomic_inc(&file_inode(file
)->i_writecount
);
2289 static inline void i_readcount_dec(struct inode
*inode
)
2291 BUG_ON(!atomic_read(&inode
->i_readcount
));
2292 atomic_dec(&inode
->i_readcount
);
2294 static inline void i_readcount_inc(struct inode
*inode
)
2296 atomic_inc(&inode
->i_readcount
);
2299 static inline void i_readcount_dec(struct inode
*inode
)
2303 static inline void i_readcount_inc(struct inode
*inode
)
2308 extern int do_pipe_flags(int *, int);
2310 extern int kernel_read(struct file
*, loff_t
, char *, unsigned long);
2311 extern ssize_t
kernel_write(struct file
*, const char *, size_t, loff_t
);
2312 extern struct file
* open_exec(const char *);
2314 /* fs/dcache.c -- generic fs support functions */
2315 extern int is_subdir(struct dentry
*, struct dentry
*);
2316 extern int path_is_under(struct path
*, struct path
*);
2318 #include <linux/err.h>
2320 /* needed for stackable file system support */
2321 extern loff_t
default_llseek(struct file
*file
, loff_t offset
, int whence
);
2323 extern loff_t
vfs_llseek(struct file
*file
, loff_t offset
, int whence
);
2325 extern int inode_init_always(struct super_block
*, struct inode
*);
2326 extern void inode_init_once(struct inode
*);
2327 extern void address_space_init_once(struct address_space
*mapping
);
2328 extern void ihold(struct inode
* inode
);
2329 extern void iput(struct inode
*);
2330 extern struct inode
* igrab(struct inode
*);
2331 extern ino_t
iunique(struct super_block
*, ino_t
);
2332 extern int inode_needs_sync(struct inode
*inode
);
2333 extern int generic_delete_inode(struct inode
*inode
);
2334 static inline int generic_drop_inode(struct inode
*inode
)
2336 return !inode
->i_nlink
|| inode_unhashed(inode
);
2339 extern struct inode
*ilookup5_nowait(struct super_block
*sb
,
2340 unsigned long hashval
, int (*test
)(struct inode
*, void *),
2342 extern struct inode
*ilookup5(struct super_block
*sb
, unsigned long hashval
,
2343 int (*test
)(struct inode
*, void *), void *data
);
2344 extern struct inode
*ilookup(struct super_block
*sb
, unsigned long ino
);
2346 extern struct inode
* iget5_locked(struct super_block
*, unsigned long, int (*test
)(struct inode
*, void *), int (*set
)(struct inode
*, void *), void *);
2347 extern struct inode
* iget_locked(struct super_block
*, unsigned long);
2348 extern int insert_inode_locked4(struct inode
*, unsigned long, int (*test
)(struct inode
*, void *), void *);
2349 extern int insert_inode_locked(struct inode
*);
2350 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2351 extern void lockdep_annotate_inode_mutex_key(struct inode
*inode
);
2353 static inline void lockdep_annotate_inode_mutex_key(struct inode
*inode
) { };
2355 extern void unlock_new_inode(struct inode
*);
2356 extern unsigned int get_next_ino(void);
2358 extern void __iget(struct inode
* inode
);
2359 extern void iget_failed(struct inode
*);
2360 extern void clear_inode(struct inode
*);
2361 extern void __destroy_inode(struct inode
*);
2362 extern struct inode
*new_inode_pseudo(struct super_block
*sb
);
2363 extern struct inode
*new_inode(struct super_block
*sb
);
2364 extern void free_inode_nonrcu(struct inode
*inode
);
2365 extern int should_remove_suid(struct dentry
*);
2366 extern int file_remove_suid(struct file
*);
2368 extern void __insert_inode_hash(struct inode
*, unsigned long hashval
);
2369 static inline void insert_inode_hash(struct inode
*inode
)
2371 __insert_inode_hash(inode
, inode
->i_ino
);
2374 extern void __remove_inode_hash(struct inode
*);
2375 static inline void remove_inode_hash(struct inode
*inode
)
2377 if (!inode_unhashed(inode
))
2378 __remove_inode_hash(inode
);
2381 extern void inode_sb_list_add(struct inode
*inode
);
2384 extern void submit_bio(int, struct bio
*);
2385 extern int bdev_read_only(struct block_device
*);
2387 extern int set_blocksize(struct block_device
*, int);
2388 extern int sb_set_blocksize(struct super_block
*, int);
2389 extern int sb_min_blocksize(struct super_block
*, int);
2391 extern int generic_file_mmap(struct file
*, struct vm_area_struct
*);
2392 extern int generic_file_readonly_mmap(struct file
*, struct vm_area_struct
*);
2393 extern int generic_file_remap_pages(struct vm_area_struct
*, unsigned long addr
,
2394 unsigned long size
, pgoff_t pgoff
);
2395 extern int file_read_actor(read_descriptor_t
* desc
, struct page
*page
, unsigned long offset
, unsigned long size
);
2396 int generic_write_checks(struct file
*file
, loff_t
*pos
, size_t *count
, int isblk
);
2397 extern ssize_t
generic_file_aio_read(struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
2398 extern ssize_t
__generic_file_aio_write(struct kiocb
*, const struct iovec
*, unsigned long,
2400 extern ssize_t
generic_file_aio_write(struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
2401 extern ssize_t
generic_file_direct_write(struct kiocb
*, const struct iovec
*,
2402 unsigned long *, loff_t
, loff_t
*, size_t, size_t);
2403 extern ssize_t
generic_file_buffered_write(struct kiocb
*, const struct iovec
*,
2404 unsigned long, loff_t
, loff_t
*, size_t, ssize_t
);
2405 extern ssize_t
do_sync_read(struct file
*filp
, char __user
*buf
, size_t len
, loff_t
*ppos
);
2406 extern ssize_t
do_sync_write(struct file
*filp
, const char __user
*buf
, size_t len
, loff_t
*ppos
);
2407 extern int generic_segment_checks(const struct iovec
*iov
,
2408 unsigned long *nr_segs
, size_t *count
, int access_flags
);
2410 /* fs/block_dev.c */
2411 extern ssize_t
blkdev_aio_write(struct kiocb
*iocb
, const struct iovec
*iov
,
2412 unsigned long nr_segs
, loff_t pos
);
2413 extern int blkdev_fsync(struct file
*filp
, loff_t start
, loff_t end
,
2415 extern void block_sync_page(struct page
*page
);
2418 extern ssize_t
generic_file_splice_read(struct file
*, loff_t
*,
2419 struct pipe_inode_info
*, size_t, unsigned int);
2420 extern ssize_t
default_file_splice_read(struct file
*, loff_t
*,
2421 struct pipe_inode_info
*, size_t, unsigned int);
2422 extern ssize_t
generic_file_splice_write(struct pipe_inode_info
*,
2423 struct file
*, loff_t
*, size_t, unsigned int);
2424 extern ssize_t
generic_splice_sendpage(struct pipe_inode_info
*pipe
,
2425 struct file
*out
, loff_t
*, size_t len
, unsigned int flags
);
2428 file_ra_state_init(struct file_ra_state
*ra
, struct address_space
*mapping
);
2429 extern loff_t
noop_llseek(struct file
*file
, loff_t offset
, int whence
);
2430 extern loff_t
no_llseek(struct file
*file
, loff_t offset
, int whence
);
2431 extern loff_t
vfs_setpos(struct file
*file
, loff_t offset
, loff_t maxsize
);
2432 extern loff_t
generic_file_llseek(struct file
*file
, loff_t offset
, int whence
);
2433 extern loff_t
generic_file_llseek_size(struct file
*file
, loff_t offset
,
2434 int whence
, loff_t maxsize
, loff_t eof
);
2435 extern loff_t
fixed_size_llseek(struct file
*file
, loff_t offset
,
2436 int whence
, loff_t size
);
2437 extern int generic_file_open(struct inode
* inode
, struct file
* filp
);
2438 extern int nonseekable_open(struct inode
* inode
, struct file
* filp
);
2440 #ifdef CONFIG_FS_XIP
2441 extern ssize_t
xip_file_read(struct file
*filp
, char __user
*buf
, size_t len
,
2443 extern int xip_file_mmap(struct file
* file
, struct vm_area_struct
* vma
);
2444 extern ssize_t
xip_file_write(struct file
*filp
, const char __user
*buf
,
2445 size_t len
, loff_t
*ppos
);
2446 extern int xip_truncate_page(struct address_space
*mapping
, loff_t from
);
2448 static inline int xip_truncate_page(struct address_space
*mapping
, loff_t from
)
2455 typedef void (dio_submit_t
)(int rw
, struct bio
*bio
, struct inode
*inode
,
2456 loff_t file_offset
);
2459 /* need locking between buffered and direct access */
2462 /* filesystem does not support filling holes */
2463 DIO_SKIP_HOLES
= 0x02,
2466 void dio_end_io(struct bio
*bio
, int error
);
2468 ssize_t
__blockdev_direct_IO(int rw
, struct kiocb
*iocb
, struct inode
*inode
,
2469 struct block_device
*bdev
, const struct iovec
*iov
, loff_t offset
,
2470 unsigned long nr_segs
, get_block_t get_block
, dio_iodone_t end_io
,
2471 dio_submit_t submit_io
, int flags
);
2473 static inline ssize_t
blockdev_direct_IO(int rw
, struct kiocb
*iocb
,
2474 struct inode
*inode
, const struct iovec
*iov
, loff_t offset
,
2475 unsigned long nr_segs
, get_block_t get_block
)
2477 return __blockdev_direct_IO(rw
, iocb
, inode
, inode
->i_sb
->s_bdev
, iov
,
2478 offset
, nr_segs
, get_block
, NULL
, NULL
,
2479 DIO_LOCKING
| DIO_SKIP_HOLES
);
2483 void inode_dio_wait(struct inode
*inode
);
2484 void inode_dio_done(struct inode
*inode
);
2486 extern const struct file_operations generic_ro_fops
;
2488 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2490 extern int vfs_readlink(struct dentry
*, char __user
*, int, const char *);
2491 extern int vfs_follow_link(struct nameidata
*, const char *);
2492 extern int page_readlink(struct dentry
*, char __user
*, int);
2493 extern void *page_follow_link_light(struct dentry
*, struct nameidata
*);
2494 extern void page_put_link(struct dentry
*, struct nameidata
*, void *);
2495 extern int __page_symlink(struct inode
*inode
, const char *symname
, int len
,
2497 extern int page_symlink(struct inode
*inode
, const char *symname
, int len
);
2498 extern const struct inode_operations page_symlink_inode_operations
;
2499 extern int generic_readlink(struct dentry
*, char __user
*, int);
2500 extern void generic_fillattr(struct inode
*, struct kstat
*);
2501 extern int vfs_getattr(struct path
*, struct kstat
*);
2502 void __inode_add_bytes(struct inode
*inode
, loff_t bytes
);
2503 void inode_add_bytes(struct inode
*inode
, loff_t bytes
);
2504 void inode_sub_bytes(struct inode
*inode
, loff_t bytes
);
2505 loff_t
inode_get_bytes(struct inode
*inode
);
2506 void inode_set_bytes(struct inode
*inode
, loff_t bytes
);
2508 extern int vfs_readdir(struct file
*, filldir_t
, void *);
2509 extern int iterate_dir(struct file
*, struct dir_context
*);
2511 extern int vfs_stat(const char __user
*, struct kstat
*);
2512 extern int vfs_lstat(const char __user
*, struct kstat
*);
2513 extern int vfs_fstat(unsigned int, struct kstat
*);
2514 extern int vfs_fstatat(int , const char __user
*, struct kstat
*, int);
2516 extern int do_vfs_ioctl(struct file
*filp
, unsigned int fd
, unsigned int cmd
,
2518 extern int __generic_block_fiemap(struct inode
*inode
,
2519 struct fiemap_extent_info
*fieinfo
,
2520 loff_t start
, loff_t len
,
2521 get_block_t
*get_block
);
2522 extern int generic_block_fiemap(struct inode
*inode
,
2523 struct fiemap_extent_info
*fieinfo
, u64 start
,
2524 u64 len
, get_block_t
*get_block
);
2526 extern void get_filesystem(struct file_system_type
*fs
);
2527 extern void put_filesystem(struct file_system_type
*fs
);
2528 extern struct file_system_type
*get_fs_type(const char *name
);
2529 extern struct super_block
*get_super(struct block_device
*);
2530 extern struct super_block
*get_super_thawed(struct block_device
*);
2531 extern struct super_block
*get_active_super(struct block_device
*bdev
);
2532 extern void drop_super(struct super_block
*sb
);
2533 extern void iterate_supers(void (*)(struct super_block
*, void *), void *);
2534 extern void iterate_supers_type(struct file_system_type
*,
2535 void (*)(struct super_block
*, void *), void *);
2537 extern int dcache_dir_open(struct inode
*, struct file
*);
2538 extern int dcache_dir_close(struct inode
*, struct file
*);
2539 extern loff_t
dcache_dir_lseek(struct file
*, loff_t
, int);
2540 extern int dcache_readdir(struct file
*, struct dir_context
*);
2541 extern int simple_setattr(struct dentry
*, struct iattr
*);
2542 extern int simple_getattr(struct vfsmount
*, struct dentry
*, struct kstat
*);
2543 extern int simple_statfs(struct dentry
*, struct kstatfs
*);
2544 extern int simple_open(struct inode
*inode
, struct file
*file
);
2545 extern int simple_link(struct dentry
*, struct inode
*, struct dentry
*);
2546 extern int simple_unlink(struct inode
*, struct dentry
*);
2547 extern int simple_rmdir(struct inode
*, struct dentry
*);
2548 extern int simple_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
2549 extern int noop_fsync(struct file
*, loff_t
, loff_t
, int);
2550 extern int simple_empty(struct dentry
*);
2551 extern int simple_readpage(struct file
*file
, struct page
*page
);
2552 extern int simple_write_begin(struct file
*file
, struct address_space
*mapping
,
2553 loff_t pos
, unsigned len
, unsigned flags
,
2554 struct page
**pagep
, void **fsdata
);
2555 extern int simple_write_end(struct file
*file
, struct address_space
*mapping
,
2556 loff_t pos
, unsigned len
, unsigned copied
,
2557 struct page
*page
, void *fsdata
);
2559 extern struct dentry
*simple_lookup(struct inode
*, struct dentry
*, unsigned int flags
);
2560 extern ssize_t
generic_read_dir(struct file
*, char __user
*, size_t, loff_t
*);
2561 extern const struct file_operations simple_dir_operations
;
2562 extern const struct inode_operations simple_dir_inode_operations
;
2563 struct tree_descr
{ char *name
; const struct file_operations
*ops
; int mode
; };
2564 struct dentry
*d_alloc_name(struct dentry
*, const char *);
2565 extern int simple_fill_super(struct super_block
*, unsigned long, struct tree_descr
*);
2566 extern int simple_pin_fs(struct file_system_type
*, struct vfsmount
**mount
, int *count
);
2567 extern void simple_release_fs(struct vfsmount
**mount
, int *count
);
2569 extern ssize_t
simple_read_from_buffer(void __user
*to
, size_t count
,
2570 loff_t
*ppos
, const void *from
, size_t available
);
2571 extern ssize_t
simple_write_to_buffer(void *to
, size_t available
, loff_t
*ppos
,
2572 const void __user
*from
, size_t count
);
2574 extern int generic_file_fsync(struct file
*, loff_t
, loff_t
, int);
2576 extern int generic_check_addressable(unsigned, u64
);
2578 #ifdef CONFIG_MIGRATION
2579 extern int buffer_migrate_page(struct address_space
*,
2580 struct page
*, struct page
*,
2583 #define buffer_migrate_page NULL
2586 extern int inode_change_ok(const struct inode
*, struct iattr
*);
2587 extern int inode_newsize_ok(const struct inode
*, loff_t offset
);
2588 extern void setattr_copy(struct inode
*inode
, const struct iattr
*attr
);
2590 extern int file_update_time(struct file
*file
);
2592 extern int generic_show_options(struct seq_file
*m
, struct dentry
*root
);
2593 extern void save_mount_options(struct super_block
*sb
, char *options
);
2594 extern void replace_mount_options(struct super_block
*sb
, char *options
);
2596 static inline ino_t
parent_ino(struct dentry
*dentry
)
2601 * Don't strictly need d_lock here? If the parent ino could change
2602 * then surely we'd have a deeper race in the caller?
2604 spin_lock(&dentry
->d_lock
);
2605 res
= dentry
->d_parent
->d_inode
->i_ino
;
2606 spin_unlock(&dentry
->d_lock
);
2610 /* Transaction based IO helpers */
2613 * An argresp is stored in an allocated page and holds the
2614 * size of the argument or response, along with its content
2616 struct simple_transaction_argresp
{
2621 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2623 char *simple_transaction_get(struct file
*file
, const char __user
*buf
,
2625 ssize_t
simple_transaction_read(struct file
*file
, char __user
*buf
,
2626 size_t size
, loff_t
*pos
);
2627 int simple_transaction_release(struct inode
*inode
, struct file
*file
);
2629 void simple_transaction_set(struct file
*file
, size_t n
);
2632 * simple attribute files
2634 * These attributes behave similar to those in sysfs:
2636 * Writing to an attribute immediately sets a value, an open file can be
2637 * written to multiple times.
2639 * Reading from an attribute creates a buffer from the value that might get
2640 * read with multiple read calls. When the attribute has been read
2641 * completely, no further read calls are possible until the file is opened
2644 * All attributes contain a text representation of a numeric value
2645 * that are accessed with the get() and set() functions.
2647 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2648 static int __fops ## _open(struct inode *inode, struct file *file) \
2650 __simple_attr_check_format(__fmt, 0ull); \
2651 return simple_attr_open(inode, file, __get, __set, __fmt); \
2653 static const struct file_operations __fops = { \
2654 .owner = THIS_MODULE, \
2655 .open = __fops ## _open, \
2656 .release = simple_attr_release, \
2657 .read = simple_attr_read, \
2658 .write = simple_attr_write, \
2659 .llseek = generic_file_llseek, \
2662 static inline __printf(1, 2)
2663 void __simple_attr_check_format(const char *fmt
, ...)
2665 /* don't do anything, just let the compiler check the arguments; */
2668 int simple_attr_open(struct inode
*inode
, struct file
*file
,
2669 int (*get
)(void *, u64
*), int (*set
)(void *, u64
),
2671 int simple_attr_release(struct inode
*inode
, struct file
*file
);
2672 ssize_t
simple_attr_read(struct file
*file
, char __user
*buf
,
2673 size_t len
, loff_t
*ppos
);
2674 ssize_t
simple_attr_write(struct file
*file
, const char __user
*buf
,
2675 size_t len
, loff_t
*ppos
);
2678 int proc_nr_files(struct ctl_table
*table
, int write
,
2679 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2680 int proc_nr_dentry(struct ctl_table
*table
, int write
,
2681 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2682 int proc_nr_inodes(struct ctl_table
*table
, int write
,
2683 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2684 int __init
get_filesystem_list(char *buf
);
2686 #define __FMODE_EXEC ((__force int) FMODE_EXEC)
2687 #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
2689 #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
2690 #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
2691 (flag & __FMODE_NONOTIFY)))
2693 static inline int is_sxid(umode_t mode
)
2695 return (mode
& S_ISUID
) || ((mode
& S_ISGID
) && (mode
& S_IXGRP
));
2698 static inline void inode_has_no_xattr(struct inode
*inode
)
2700 if (!is_sxid(inode
->i_mode
) && (inode
->i_sb
->s_flags
& MS_NOSEC
))
2701 inode
->i_flags
|= S_NOSEC
;
2704 static inline bool dir_emit(struct dir_context
*ctx
,
2705 const char *name
, int namelen
,
2706 u64 ino
, unsigned type
)
2708 return ctx
->actor(ctx
, name
, namelen
, ctx
->pos
, ino
, type
) == 0;
2710 static inline bool dir_emit_dot(struct file
*file
, struct dir_context
*ctx
)
2712 return ctx
->actor(ctx
, ".", 1, ctx
->pos
,
2713 file
->f_path
.dentry
->d_inode
->i_ino
, DT_DIR
) == 0;
2715 static inline bool dir_emit_dotdot(struct file
*file
, struct dir_context
*ctx
)
2717 return ctx
->actor(ctx
, "..", 2, ctx
->pos
,
2718 parent_ino(file
->f_path
.dentry
), DT_DIR
) == 0;
2720 static inline bool dir_emit_dots(struct file
*file
, struct dir_context
*ctx
)
2722 if (ctx
->pos
== 0) {
2723 if (!dir_emit_dot(file
, ctx
))
2727 if (ctx
->pos
== 1) {
2728 if (!dir_emit_dotdot(file
, ctx
))
2734 static inline bool dir_relax(struct inode
*inode
)
2736 mutex_unlock(&inode
->i_mutex
);
2737 mutex_lock(&inode
->i_mutex
);
2738 return !IS_DEADDIR(inode
);
2741 #endif /* _LINUX_FS_H */