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/list_lru.h>
14 #include <linux/llist.h>
15 #include <linux/radix-tree.h>
16 #include <linux/rbtree.h>
17 #include <linux/init.h>
18 #include <linux/pid.h>
19 #include <linux/bug.h>
20 #include <linux/mutex.h>
21 #include <linux/capability.h>
22 #include <linux/semaphore.h>
23 #include <linux/fiemap.h>
24 #include <linux/rculist_bl.h>
25 #include <linux/atomic.h>
26 #include <linux/shrinker.h>
27 #include <linux/migrate_mode.h>
28 #include <linux/uidgid.h>
29 #include <linux/lockdep.h>
30 #include <linux/percpu-rwsem.h>
31 #include <linux/blk_types.h>
33 #include <asm/byteorder.h>
34 #include <uapi/linux/fs.h>
36 struct export_operations
;
42 struct pipe_inode_info
;
43 struct poll_table_struct
;
45 struct vm_area_struct
;
48 struct swap_info_struct
;
50 struct workqueue_struct
;
52 extern void __init
inode_init(void);
53 extern void __init
inode_init_early(void);
54 extern void __init
files_init(unsigned long);
56 extern struct files_stat_struct files_stat
;
57 extern unsigned long get_max_files(void);
58 extern int sysctl_nr_open
;
59 extern struct inodes_stat_t inodes_stat
;
60 extern int leases_enable
, lease_break_time
;
61 extern int sysctl_protected_symlinks
;
62 extern int sysctl_protected_hardlinks
;
65 typedef int (get_block_t
)(struct inode
*inode
, sector_t iblock
,
66 struct buffer_head
*bh_result
, int create
);
67 typedef void (dio_iodone_t
)(struct kiocb
*iocb
, loff_t offset
,
68 ssize_t bytes
, void *private);
70 #define MAY_EXEC 0x00000001
71 #define MAY_WRITE 0x00000002
72 #define MAY_READ 0x00000004
73 #define MAY_APPEND 0x00000008
74 #define MAY_ACCESS 0x00000010
75 #define MAY_OPEN 0x00000020
76 #define MAY_CHDIR 0x00000040
77 /* called from RCU mode, don't block */
78 #define MAY_NOT_BLOCK 0x00000080
81 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
82 * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
85 /* file is open for reading */
86 #define FMODE_READ ((__force fmode_t)0x1)
87 /* file is open for writing */
88 #define FMODE_WRITE ((__force fmode_t)0x2)
89 /* file is seekable */
90 #define FMODE_LSEEK ((__force fmode_t)0x4)
91 /* file can be accessed using pread */
92 #define FMODE_PREAD ((__force fmode_t)0x8)
93 /* file can be accessed using pwrite */
94 #define FMODE_PWRITE ((__force fmode_t)0x10)
95 /* File is opened for execution with sys_execve / sys_uselib */
96 #define FMODE_EXEC ((__force fmode_t)0x20)
97 /* File is opened with O_NDELAY (only set for block devices) */
98 #define FMODE_NDELAY ((__force fmode_t)0x40)
99 /* File is opened with O_EXCL (only set for block devices) */
100 #define FMODE_EXCL ((__force fmode_t)0x80)
101 /* File is opened using open(.., 3, ..) and is writeable only for ioctls
102 (specialy hack for floppy.c) */
103 #define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
104 /* 32bit hashes as llseek() offset (for directories) */
105 #define FMODE_32BITHASH ((__force fmode_t)0x200)
106 /* 64bit hashes as llseek() offset (for directories) */
107 #define FMODE_64BITHASH ((__force fmode_t)0x400)
110 * Don't update ctime and mtime.
112 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
113 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
115 #define FMODE_NOCMTIME ((__force fmode_t)0x800)
117 /* Expect random access pattern */
118 #define FMODE_RANDOM ((__force fmode_t)0x1000)
120 /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
121 #define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
123 /* File is opened with O_PATH; almost nothing can be done with it */
124 #define FMODE_PATH ((__force fmode_t)0x4000)
126 /* File needs atomic accesses to f_pos */
127 #define FMODE_ATOMIC_POS ((__force fmode_t)0x8000)
129 /* File was opened by fanotify and shouldn't generate fanotify events */
130 #define FMODE_NONOTIFY ((__force fmode_t)0x1000000)
133 * Flag for rw_copy_check_uvector and compat_rw_copy_check_uvector
134 * that indicates that they should check the contents of the iovec are
135 * valid, but not check the memory that the iovec elements
138 #define CHECK_IOVEC_ONLY -1
141 * The below are the various read and write types that we support. Some of
142 * them include behavioral modifiers that send information down to the
143 * block layer and IO scheduler. Terminology:
145 * The block layer uses device plugging to defer IO a little bit, in
146 * the hope that we will see more IO very shortly. This increases
147 * coalescing of adjacent IO and thus reduces the number of IOs we
148 * have to send to the device. It also allows for better queuing,
149 * if the IO isn't mergeable. If the caller is going to be waiting
150 * for the IO, then he must ensure that the device is unplugged so
151 * that the IO is dispatched to the driver.
153 * All IO is handled async in Linux. This is fine for background
154 * writes, but for reads or writes that someone waits for completion
155 * on, we want to notify the block layer and IO scheduler so that they
156 * know about it. That allows them to make better scheduling
157 * decisions. So when the below references 'sync' and 'async', it
158 * is referencing this priority hint.
160 * With that in mind, the available types are:
162 * READ A normal read operation. Device will be plugged.
163 * READ_SYNC A synchronous read. Device is not plugged, caller can
164 * immediately wait on this read without caring about
166 * READA Used for read-ahead operations. Lower priority, and the
167 * block layer could (in theory) choose to ignore this
168 * request if it runs into resource problems.
169 * WRITE A normal async write. Device will be plugged.
170 * WRITE_SYNC Synchronous write. Identical to WRITE, but passes down
171 * the hint that someone will be waiting on this IO
172 * shortly. The write equivalent of READ_SYNC.
173 * WRITE_ODIRECT Special case write for O_DIRECT only.
174 * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
175 * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
176 * non-volatile media on completion.
177 * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
178 * by a cache flush and data is guaranteed to be on
179 * non-volatile media on completion.
182 #define RW_MASK REQ_WRITE
183 #define RWA_MASK REQ_RAHEAD
186 #define WRITE RW_MASK
187 #define READA RWA_MASK
188 #define KERNEL_READ (READ|REQ_KERNEL)
189 #define KERNEL_WRITE (WRITE|REQ_KERNEL)
191 #define READ_SYNC (READ | REQ_SYNC)
192 #define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE)
193 #define WRITE_ODIRECT (WRITE | REQ_SYNC)
194 #define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
195 #define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
196 #define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
199 * Attribute flags. These should be or-ed together to figure out what
202 #define ATTR_MODE (1 << 0)
203 #define ATTR_UID (1 << 1)
204 #define ATTR_GID (1 << 2)
205 #define ATTR_SIZE (1 << 3)
206 #define ATTR_ATIME (1 << 4)
207 #define ATTR_MTIME (1 << 5)
208 #define ATTR_CTIME (1 << 6)
209 #define ATTR_ATIME_SET (1 << 7)
210 #define ATTR_MTIME_SET (1 << 8)
211 #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
212 #define ATTR_ATTR_FLAG (1 << 10)
213 #define ATTR_KILL_SUID (1 << 11)
214 #define ATTR_KILL_SGID (1 << 12)
215 #define ATTR_FILE (1 << 13)
216 #define ATTR_KILL_PRIV (1 << 14)
217 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
218 #define ATTR_TIMES_SET (1 << 16)
221 * This is the Inode Attributes structure, used for notify_change(). It
222 * uses the above definitions as flags, to know which values have changed.
223 * Also, in this manner, a Filesystem can look at only the values it cares
224 * about. Basically, these are the attributes that the VFS layer can
225 * request to change from the FS layer.
227 * Derek Atkins <warlord@MIT.EDU> 94-10-20
230 unsigned int ia_valid
;
235 struct timespec ia_atime
;
236 struct timespec ia_mtime
;
237 struct timespec ia_ctime
;
240 * Not an attribute, but an auxiliary info for filesystems wanting to
241 * implement an ftruncate() like method. NOTE: filesystem should
242 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
244 struct file
*ia_file
;
248 * Includes for diskquotas.
250 #include <linux/quota.h>
253 * enum positive_aop_returns - aop return codes with specific semantics
255 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
256 * completed, that the page is still locked, and
257 * should be considered active. The VM uses this hint
258 * to return the page to the active list -- it won't
259 * be a candidate for writeback again in the near
260 * future. Other callers must be careful to unlock
261 * the page if they get this return. Returned by
264 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
265 * unlocked it and the page might have been truncated.
266 * The caller should back up to acquiring a new page and
267 * trying again. The aop will be taking reasonable
268 * precautions not to livelock. If the caller held a page
269 * reference, it should drop it before retrying. Returned
272 * address_space_operation functions return these large constants to indicate
273 * special semantics to the caller. These are much larger than the bytes in a
274 * page to allow for functions that return the number of bytes operated on in a
278 enum positive_aop_returns
{
279 AOP_WRITEPAGE_ACTIVATE
= 0x80000,
280 AOP_TRUNCATED_PAGE
= 0x80001,
283 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
284 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
285 #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
286 * helper code (eg buffer layer)
287 * to clear GFP_FS from alloc */
290 * oh the beauties of C type declarations.
293 struct address_space
;
294 struct writeback_control
;
297 const struct iovec
*iov
;
298 unsigned long nr_segs
;
303 size_t iov_iter_copy_from_user_atomic(struct page
*page
,
304 struct iov_iter
*i
, unsigned long offset
, size_t bytes
);
305 size_t iov_iter_copy_from_user(struct page
*page
,
306 struct iov_iter
*i
, unsigned long offset
, size_t bytes
);
307 void iov_iter_advance(struct iov_iter
*i
, size_t bytes
);
308 int iov_iter_fault_in_readable(struct iov_iter
*i
, size_t bytes
);
309 size_t iov_iter_single_seg_count(const struct iov_iter
*i
);
311 static inline void iov_iter_init(struct iov_iter
*i
,
312 const struct iovec
*iov
, unsigned long nr_segs
,
313 size_t count
, size_t written
)
316 i
->nr_segs
= nr_segs
;
318 i
->count
= count
+ written
;
320 iov_iter_advance(i
, written
);
323 static inline size_t iov_iter_count(struct iov_iter
*i
)
329 * "descriptor" for what we're up to with a read.
330 * This allows us to use the same read code yet
331 * have multiple different users of the data that
332 * we read from a file.
334 * The simplest case just copies the data to user
347 typedef int (*read_actor_t
)(read_descriptor_t
*, struct page
*,
348 unsigned long, unsigned long);
350 struct address_space_operations
{
351 int (*writepage
)(struct page
*page
, struct writeback_control
*wbc
);
352 int (*readpage
)(struct file
*, struct page
*);
354 /* Write back some dirty pages from this mapping. */
355 int (*writepages
)(struct address_space
*, struct writeback_control
*);
357 /* Set a page dirty. Return true if this dirtied it */
358 int (*set_page_dirty
)(struct page
*page
);
360 int (*readpages
)(struct file
*filp
, struct address_space
*mapping
,
361 struct list_head
*pages
, unsigned nr_pages
);
363 int (*write_begin
)(struct file
*, struct address_space
*mapping
,
364 loff_t pos
, unsigned len
, unsigned flags
,
365 struct page
**pagep
, void **fsdata
);
366 int (*write_end
)(struct file
*, struct address_space
*mapping
,
367 loff_t pos
, unsigned len
, unsigned copied
,
368 struct page
*page
, void *fsdata
);
370 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
371 sector_t (*bmap
)(struct address_space
*, sector_t
);
372 void (*invalidatepage
) (struct page
*, unsigned int, unsigned int);
373 int (*releasepage
) (struct page
*, gfp_t
);
374 void (*freepage
)(struct page
*);
375 ssize_t (*direct_IO
)(int, struct kiocb
*, const struct iovec
*iov
,
376 loff_t offset
, unsigned long nr_segs
);
377 int (*get_xip_mem
)(struct address_space
*, pgoff_t
, int,
378 void **, unsigned long *);
380 * migrate the contents of a page to the specified target. If
381 * migrate_mode is MIGRATE_ASYNC, it must not block.
383 int (*migratepage
) (struct address_space
*,
384 struct page
*, struct page
*, enum migrate_mode
);
385 int (*launder_page
) (struct page
*);
386 int (*is_partially_uptodate
) (struct page
*, read_descriptor_t
*,
388 void (*is_dirty_writeback
) (struct page
*, bool *, bool *);
389 int (*error_remove_page
)(struct address_space
*, struct page
*);
391 /* swapfile support */
392 int (*swap_activate
)(struct swap_info_struct
*sis
, struct file
*file
,
394 void (*swap_deactivate
)(struct file
*file
);
397 extern const struct address_space_operations empty_aops
;
400 * pagecache_write_begin/pagecache_write_end must be used by general code
401 * to write into the pagecache.
403 int pagecache_write_begin(struct file
*, struct address_space
*mapping
,
404 loff_t pos
, unsigned len
, unsigned flags
,
405 struct page
**pagep
, void **fsdata
);
407 int pagecache_write_end(struct file
*, struct address_space
*mapping
,
408 loff_t pos
, unsigned len
, unsigned copied
,
409 struct page
*page
, void *fsdata
);
411 struct backing_dev_info
;
412 struct address_space
{
413 struct inode
*host
; /* owner: inode, block_device */
414 struct radix_tree_root page_tree
; /* radix tree of all pages */
415 spinlock_t tree_lock
; /* and lock protecting it */
416 unsigned int i_mmap_writable
;/* count VM_SHARED mappings */
417 struct rb_root i_mmap
; /* tree of private and shared mappings */
418 struct list_head i_mmap_nonlinear
;/*list VM_NONLINEAR mappings */
419 struct mutex i_mmap_mutex
; /* protect tree, count, list */
420 /* Protected by tree_lock together with the radix tree */
421 unsigned long nrpages
; /* number of total pages */
422 pgoff_t writeback_index
;/* writeback starts here */
423 const struct address_space_operations
*a_ops
; /* methods */
424 unsigned long flags
; /* error bits/gfp mask */
425 struct backing_dev_info
*backing_dev_info
; /* device readahead, etc */
426 spinlock_t private_lock
; /* for use by the address_space */
427 struct list_head private_list
; /* ditto */
428 void *private_data
; /* ditto */
429 } __attribute__((aligned(sizeof(long))));
431 * On most architectures that alignment is already the case; but
432 * must be enforced here for CRIS, to let the least significant bit
433 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
435 struct request_queue
;
437 struct block_device
{
438 dev_t bd_dev
; /* not a kdev_t - it's a search key */
440 struct inode
* bd_inode
; /* will die */
441 struct super_block
* bd_super
;
442 struct mutex bd_mutex
; /* open/close mutex */
443 struct list_head bd_inodes
;
447 bool bd_write_holder
;
449 struct list_head bd_holder_disks
;
451 struct block_device
* bd_contains
;
452 unsigned bd_block_size
;
453 struct hd_struct
* bd_part
;
454 /* number of times partitions within this device have been opened. */
455 unsigned bd_part_count
;
457 struct gendisk
* bd_disk
;
458 struct request_queue
* bd_queue
;
459 struct list_head bd_list
;
461 * Private data. You must have bd_claim'ed the block_device
462 * to use this. NOTE: bd_claim allows an owner to claim
463 * the same device multiple times, the owner must take special
464 * care to not mess up bd_private for that case.
466 unsigned long bd_private
;
468 /* The counter of freeze processes */
469 int bd_fsfreeze_count
;
470 /* Mutex for freeze */
471 struct mutex bd_fsfreeze_mutex
;
475 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
478 #define PAGECACHE_TAG_DIRTY 0
479 #define PAGECACHE_TAG_WRITEBACK 1
480 #define PAGECACHE_TAG_TOWRITE 2
482 int mapping_tagged(struct address_space
*mapping
, int tag
);
485 * Might pages of this file be mapped into userspace?
487 static inline int mapping_mapped(struct address_space
*mapping
)
489 return !RB_EMPTY_ROOT(&mapping
->i_mmap
) ||
490 !list_empty(&mapping
->i_mmap_nonlinear
);
494 * Might pages of this file have been modified in userspace?
495 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
496 * marks vma as VM_SHARED if it is shared, and the file was opened for
497 * writing i.e. vma may be mprotected writable even if now readonly.
499 static inline int mapping_writably_mapped(struct address_space
*mapping
)
501 return mapping
->i_mmap_writable
!= 0;
505 * Use sequence counter to get consistent i_size on 32-bit processors.
507 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
508 #include <linux/seqlock.h>
509 #define __NEED_I_SIZE_ORDERED
510 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
512 #define i_size_ordered_init(inode) do { } while (0)
516 #define ACL_NOT_CACHED ((void *)(-1))
518 #define IOP_FASTPERM 0x0001
519 #define IOP_LOOKUP 0x0002
520 #define IOP_NOFOLLOW 0x0004
523 * Keep mostly read-only and often accessed (especially for
524 * the RCU path lookup and 'stat' data) fields at the beginning
525 * of the 'struct inode'
529 unsigned short i_opflags
;
532 unsigned int i_flags
;
534 #ifdef CONFIG_FS_POSIX_ACL
535 struct posix_acl
*i_acl
;
536 struct posix_acl
*i_default_acl
;
539 const struct inode_operations
*i_op
;
540 struct super_block
*i_sb
;
541 struct address_space
*i_mapping
;
543 #ifdef CONFIG_SECURITY
547 /* Stat data, not accessed from path walking */
550 * Filesystems may only read i_nlink directly. They shall use the
551 * following functions for modification:
553 * (set|clear|inc|drop)_nlink
554 * inode_(inc|dec)_link_count
557 const unsigned int i_nlink
;
558 unsigned int __i_nlink
;
562 struct timespec i_atime
;
563 struct timespec i_mtime
;
564 struct timespec i_ctime
;
565 spinlock_t i_lock
; /* i_blocks, i_bytes, maybe i_size */
566 unsigned short i_bytes
;
567 unsigned int i_blkbits
;
570 #ifdef __NEED_I_SIZE_ORDERED
571 seqcount_t i_size_seqcount
;
575 unsigned long i_state
;
576 struct mutex i_mutex
;
578 unsigned long dirtied_when
; /* jiffies of first dirtying */
580 struct hlist_node i_hash
;
581 struct list_head i_wb_list
; /* backing dev IO list */
582 struct list_head i_lru
; /* inode LRU list */
583 struct list_head i_sb_list
;
585 struct hlist_head i_dentry
;
586 struct rcu_head i_rcu
;
590 atomic_t i_dio_count
;
591 atomic_t i_writecount
;
592 const struct file_operations
*i_fop
; /* former ->i_op->default_file_ops */
593 struct file_lock
*i_flock
;
594 struct address_space i_data
;
596 struct dquot
*i_dquot
[MAXQUOTAS
];
598 struct list_head i_devices
;
600 struct pipe_inode_info
*i_pipe
;
601 struct block_device
*i_bdev
;
607 #ifdef CONFIG_FSNOTIFY
608 __u32 i_fsnotify_mask
; /* all events this inode cares about */
609 struct hlist_head i_fsnotify_marks
;
613 atomic_t i_readcount
; /* struct files open RO */
615 void *i_private
; /* fs or device private pointer */
618 static inline int inode_unhashed(struct inode
*inode
)
620 return hlist_unhashed(&inode
->i_hash
);
624 * inode->i_mutex nesting subclasses for the lock validator:
626 * 0: the object of the current VFS operation
630 * 4: second non-directory
631 * The last is for certain operations (such as rename) which lock two
632 * non-directories at once.
634 * The locking order between these classes is
635 * parent -> child -> normal -> xattr -> second non-directory
637 enum inode_i_mutex_lock_class
646 void lock_two_nondirectories(struct inode
*, struct inode
*);
647 void unlock_two_nondirectories(struct inode
*, struct inode
*);
650 * NOTE: in a 32bit arch with a preemptable kernel and
651 * an UP compile the i_size_read/write must be atomic
652 * with respect to the local cpu (unlike with preempt disabled),
653 * but they don't need to be atomic with respect to other cpus like in
654 * true SMP (so they need either to either locally disable irq around
655 * the read or for example on x86 they can be still implemented as a
656 * cmpxchg8b without the need of the lock prefix). For SMP compiles
657 * and 64bit archs it makes no difference if preempt is enabled or not.
659 static inline loff_t
i_size_read(const struct inode
*inode
)
661 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
666 seq
= read_seqcount_begin(&inode
->i_size_seqcount
);
667 i_size
= inode
->i_size
;
668 } while (read_seqcount_retry(&inode
->i_size_seqcount
, seq
));
670 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
674 i_size
= inode
->i_size
;
678 return inode
->i_size
;
683 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
684 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
685 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
687 static inline void i_size_write(struct inode
*inode
, loff_t i_size
)
689 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
691 write_seqcount_begin(&inode
->i_size_seqcount
);
692 inode
->i_size
= i_size
;
693 write_seqcount_end(&inode
->i_size_seqcount
);
695 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
697 inode
->i_size
= i_size
;
700 inode
->i_size
= i_size
;
704 /* Helper functions so that in most cases filesystems will
705 * not need to deal directly with kuid_t and kgid_t and can
706 * instead deal with the raw numeric values that are stored
709 static inline uid_t
i_uid_read(const struct inode
*inode
)
711 return from_kuid(&init_user_ns
, inode
->i_uid
);
714 static inline gid_t
i_gid_read(const struct inode
*inode
)
716 return from_kgid(&init_user_ns
, inode
->i_gid
);
719 static inline void i_uid_write(struct inode
*inode
, uid_t uid
)
721 inode
->i_uid
= make_kuid(&init_user_ns
, uid
);
724 static inline void i_gid_write(struct inode
*inode
, gid_t gid
)
726 inode
->i_gid
= make_kgid(&init_user_ns
, gid
);
729 static inline unsigned iminor(const struct inode
*inode
)
731 return MINOR(inode
->i_rdev
);
734 static inline unsigned imajor(const struct inode
*inode
)
736 return MAJOR(inode
->i_rdev
);
739 extern struct block_device
*I_BDEV(struct inode
*inode
);
742 rwlock_t lock
; /* protects pid, uid, euid fields */
743 struct pid
*pid
; /* pid or -pgrp where SIGIO should be sent */
744 enum pid_type pid_type
; /* Kind of process group SIGIO should be sent to */
745 kuid_t uid
, euid
; /* uid/euid of process setting the owner */
746 int signum
; /* posix.1b rt signal to be delivered on IO */
750 * Track a single file's readahead state
752 struct file_ra_state
{
753 pgoff_t start
; /* where readahead started */
754 unsigned int size
; /* # of readahead pages */
755 unsigned int async_size
; /* do asynchronous readahead when
756 there are only # of pages ahead */
758 unsigned int ra_pages
; /* Maximum readahead window */
759 unsigned int mmap_miss
; /* Cache miss stat for mmap accesses */
760 loff_t prev_pos
; /* Cache last read() position */
764 * Check if @index falls in the readahead windows.
766 static inline int ra_has_index(struct file_ra_state
*ra
, pgoff_t index
)
768 return (index
>= ra
->start
&&
769 index
< ra
->start
+ ra
->size
);
774 struct llist_node fu_llist
;
775 struct rcu_head fu_rcuhead
;
778 #define f_dentry f_path.dentry
779 struct inode
*f_inode
; /* cached value */
780 const struct file_operations
*f_op
;
783 * Protects f_ep_links, f_flags.
784 * Must not be taken from IRQ context.
787 atomic_long_t f_count
;
788 unsigned int f_flags
;
790 struct mutex f_pos_lock
;
792 struct fown_struct f_owner
;
793 const struct cred
*f_cred
;
794 struct file_ra_state f_ra
;
797 #ifdef CONFIG_SECURITY
800 /* needed for tty driver, and maybe others */
804 /* Used by fs/eventpoll.c to link all the hooks to this file */
805 struct list_head f_ep_links
;
806 struct list_head f_tfile_llink
;
807 #endif /* #ifdef CONFIG_EPOLL */
808 struct address_space
*f_mapping
;
809 } __attribute__((aligned(4))); /* lest something weird decides that 2 is OK */
814 /* file identifier */
815 unsigned char f_handle
[0];
818 static inline struct file
*get_file(struct file
*f
)
820 atomic_long_inc(&f
->f_count
);
823 #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
824 #define file_count(x) atomic_long_read(&(x)->f_count)
826 #define MAX_NON_LFS ((1UL<<31) - 1)
828 /* Page cache limit. The filesystems should put that into their s_maxbytes
829 limits, otherwise bad things can happen in VM. */
830 #if BITS_PER_LONG==32
831 #define MAX_LFS_FILESIZE (((loff_t)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
832 #elif BITS_PER_LONG==64
833 #define MAX_LFS_FILESIZE ((loff_t)0x7fffffffffffffffLL)
838 #define FL_DELEG 4 /* NFSv4 delegation */
839 #define FL_ACCESS 8 /* not trying to lock, just looking */
840 #define FL_EXISTS 16 /* when unlocking, test for existence */
841 #define FL_LEASE 32 /* lease held on this file */
842 #define FL_CLOSE 64 /* unlock on close */
843 #define FL_SLEEP 128 /* A blocking lock */
844 #define FL_DOWNGRADE_PENDING 256 /* Lease is being downgraded */
845 #define FL_UNLOCK_PENDING 512 /* Lease is being broken */
848 * Special return value from posix_lock_file() and vfs_lock_file() for
849 * asynchronous locking.
851 #define FILE_LOCK_DEFERRED 1
854 * The POSIX file lock owner is determined by
855 * the "struct files_struct" in the thread group
856 * (or NULL for no owner - BSD locks).
858 * Lockd stuffs a "host" pointer into this.
860 typedef struct files_struct
*fl_owner_t
;
862 struct file_lock_operations
{
863 void (*fl_copy_lock
)(struct file_lock
*, struct file_lock
*);
864 void (*fl_release_private
)(struct file_lock
*);
867 struct lock_manager_operations
{
868 int (*lm_compare_owner
)(struct file_lock
*, struct file_lock
*);
869 unsigned long (*lm_owner_key
)(struct file_lock
*);
870 void (*lm_notify
)(struct file_lock
*); /* unblock callback */
871 int (*lm_grant
)(struct file_lock
*, struct file_lock
*, int);
872 void (*lm_break
)(struct file_lock
*);
873 int (*lm_change
)(struct file_lock
**, int);
876 struct lock_manager
{
877 struct list_head list
;
881 void locks_start_grace(struct net
*, struct lock_manager
*);
882 void locks_end_grace(struct lock_manager
*);
883 int locks_in_grace(struct net
*);
885 /* that will die - we need it for nfs_lock_info */
886 #include <linux/nfs_fs_i.h>
889 * struct file_lock represents a generic "file lock". It's used to represent
890 * POSIX byte range locks, BSD (flock) locks, and leases. It's important to
891 * note that the same struct is used to represent both a request for a lock and
892 * the lock itself, but the same object is never used for both.
894 * FIXME: should we create a separate "struct lock_request" to help distinguish
897 * The i_flock list is ordered by:
899 * 1) lock type -- FL_LEASEs first, then FL_FLOCK, and finally FL_POSIX
901 * 3) lock range start
904 * Obviously, the last two criteria only matter for POSIX locks.
907 struct file_lock
*fl_next
; /* singly linked list for this inode */
908 struct hlist_node fl_link
; /* node in global lists */
909 struct list_head fl_block
; /* circular list of blocked processes */
911 unsigned int fl_flags
;
912 unsigned char fl_type
;
914 int fl_link_cpu
; /* what cpu's list is this on? */
915 struct pid
*fl_nspid
;
916 wait_queue_head_t fl_wait
;
917 struct file
*fl_file
;
921 struct fasync_struct
* fl_fasync
; /* for lease break notifications */
922 /* for lease breaks: */
923 unsigned long fl_break_time
;
924 unsigned long fl_downgrade_time
;
926 const struct file_lock_operations
*fl_ops
; /* Callbacks for filesystems */
927 const struct lock_manager_operations
*fl_lmops
; /* Callbacks for lockmanagers */
929 struct nfs_lock_info nfs_fl
;
930 struct nfs4_lock_info nfs4_fl
;
932 struct list_head link
; /* link in AFS vnode's pending_locks list */
933 int state
; /* state of grant or error if -ve */
938 /* The following constant reflects the upper bound of the file/locking space */
940 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
941 #define OFFSET_MAX INT_LIMIT(loff_t)
942 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
945 #include <linux/fcntl.h>
947 extern void send_sigio(struct fown_struct
*fown
, int fd
, int band
);
949 #ifdef CONFIG_FILE_LOCKING
950 extern int fcntl_getlk(struct file
*, struct flock __user
*);
951 extern int fcntl_setlk(unsigned int, struct file
*, unsigned int,
952 struct flock __user
*);
954 #if BITS_PER_LONG == 32
955 extern int fcntl_getlk64(struct file
*, struct flock64 __user
*);
956 extern int fcntl_setlk64(unsigned int, struct file
*, unsigned int,
957 struct flock64 __user
*);
960 extern int fcntl_setlease(unsigned int fd
, struct file
*filp
, long arg
);
961 extern int fcntl_getlease(struct file
*filp
);
964 void locks_free_lock(struct file_lock
*fl
);
965 extern void locks_init_lock(struct file_lock
*);
966 extern struct file_lock
* locks_alloc_lock(void);
967 extern void locks_copy_lock(struct file_lock
*, struct file_lock
*);
968 extern void __locks_copy_lock(struct file_lock
*, const struct file_lock
*);
969 extern void locks_remove_posix(struct file
*, fl_owner_t
);
970 extern void locks_remove_flock(struct file
*);
971 extern void locks_release_private(struct file_lock
*);
972 extern void posix_test_lock(struct file
*, struct file_lock
*);
973 extern int posix_lock_file(struct file
*, struct file_lock
*, struct file_lock
*);
974 extern int posix_lock_file_wait(struct file
*, struct file_lock
*);
975 extern int posix_unblock_lock(struct file_lock
*);
976 extern int vfs_test_lock(struct file
*, struct file_lock
*);
977 extern int vfs_lock_file(struct file
*, unsigned int, struct file_lock
*, struct file_lock
*);
978 extern int vfs_cancel_lock(struct file
*filp
, struct file_lock
*fl
);
979 extern int flock_lock_file_wait(struct file
*filp
, struct file_lock
*fl
);
980 extern int __break_lease(struct inode
*inode
, unsigned int flags
, unsigned int type
);
981 extern void lease_get_mtime(struct inode
*, struct timespec
*time
);
982 extern int generic_setlease(struct file
*, long, struct file_lock
**);
983 extern int vfs_setlease(struct file
*, long, struct file_lock
**);
984 extern int lease_modify(struct file_lock
**, int);
985 extern int lock_may_read(struct inode
*, loff_t start
, unsigned long count
);
986 extern int lock_may_write(struct inode
*, loff_t start
, unsigned long count
);
987 #else /* !CONFIG_FILE_LOCKING */
988 static inline int fcntl_getlk(struct file
*file
, struct flock __user
*user
)
993 static inline int fcntl_setlk(unsigned int fd
, struct file
*file
,
994 unsigned int cmd
, struct flock __user
*user
)
999 #if BITS_PER_LONG == 32
1000 static inline int fcntl_getlk64(struct file
*file
, struct flock64 __user
*user
)
1005 static inline int fcntl_setlk64(unsigned int fd
, struct file
*file
,
1006 unsigned int cmd
, struct flock64 __user
*user
)
1011 static inline int fcntl_setlease(unsigned int fd
, struct file
*filp
, long arg
)
1016 static inline int fcntl_getlease(struct file
*filp
)
1021 static inline void locks_init_lock(struct file_lock
*fl
)
1026 static inline void __locks_copy_lock(struct file_lock
*new, struct file_lock
*fl
)
1031 static inline void locks_copy_lock(struct file_lock
*new, struct file_lock
*fl
)
1036 static inline void locks_remove_posix(struct file
*filp
, fl_owner_t owner
)
1041 static inline void locks_remove_flock(struct file
*filp
)
1046 static inline void posix_test_lock(struct file
*filp
, struct file_lock
*fl
)
1051 static inline int posix_lock_file(struct file
*filp
, struct file_lock
*fl
,
1052 struct file_lock
*conflock
)
1057 static inline int posix_lock_file_wait(struct file
*filp
, struct file_lock
*fl
)
1062 static inline int posix_unblock_lock(struct file_lock
*waiter
)
1067 static inline int vfs_test_lock(struct file
*filp
, struct file_lock
*fl
)
1072 static inline int vfs_lock_file(struct file
*filp
, unsigned int cmd
,
1073 struct file_lock
*fl
, struct file_lock
*conf
)
1078 static inline int vfs_cancel_lock(struct file
*filp
, struct file_lock
*fl
)
1083 static inline int flock_lock_file_wait(struct file
*filp
,
1084 struct file_lock
*request
)
1089 static inline int __break_lease(struct inode
*inode
, unsigned int mode
, unsigned int type
)
1094 static inline void lease_get_mtime(struct inode
*inode
, struct timespec
*time
)
1099 static inline int generic_setlease(struct file
*filp
, long arg
,
1100 struct file_lock
**flp
)
1105 static inline int vfs_setlease(struct file
*filp
, long arg
,
1106 struct file_lock
**lease
)
1111 static inline int lease_modify(struct file_lock
**before
, int arg
)
1116 static inline int lock_may_read(struct inode
*inode
, loff_t start
,
1122 static inline int lock_may_write(struct inode
*inode
, loff_t start
,
1127 #endif /* !CONFIG_FILE_LOCKING */
1130 struct fasync_struct
{
1134 struct fasync_struct
*fa_next
; /* singly linked list */
1135 struct file
*fa_file
;
1136 struct rcu_head fa_rcu
;
1139 #define FASYNC_MAGIC 0x4601
1141 /* SMP safe fasync helpers: */
1142 extern int fasync_helper(int, struct file
*, int, struct fasync_struct
**);
1143 extern struct fasync_struct
*fasync_insert_entry(int, struct file
*, struct fasync_struct
**, struct fasync_struct
*);
1144 extern int fasync_remove_entry(struct file
*, struct fasync_struct
**);
1145 extern struct fasync_struct
*fasync_alloc(void);
1146 extern void fasync_free(struct fasync_struct
*);
1148 /* can be called from interrupts */
1149 extern void kill_fasync(struct fasync_struct
**, int, int);
1151 extern int __f_setown(struct file
*filp
, struct pid
*, enum pid_type
, int force
);
1152 extern int f_setown(struct file
*filp
, unsigned long arg
, int force
);
1153 extern void f_delown(struct file
*filp
);
1154 extern pid_t
f_getown(struct file
*filp
);
1155 extern int send_sigurg(struct fown_struct
*fown
);
1163 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1164 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1165 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1166 #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1167 #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
1169 extern struct list_head super_blocks
;
1170 extern spinlock_t sb_lock
;
1172 /* Possible states of 'frozen' field */
1174 SB_UNFROZEN
= 0, /* FS is unfrozen */
1175 SB_FREEZE_WRITE
= 1, /* Writes, dir ops, ioctls frozen */
1176 SB_FREEZE_PAGEFAULT
= 2, /* Page faults stopped as well */
1177 SB_FREEZE_FS
= 3, /* For internal FS use (e.g. to stop
1178 * internal threads if needed) */
1179 SB_FREEZE_COMPLETE
= 4, /* ->freeze_fs finished successfully */
1182 #define SB_FREEZE_LEVELS (SB_FREEZE_COMPLETE - 1)
1185 /* Counters for counting writers at each level */
1186 struct percpu_counter counter
[SB_FREEZE_LEVELS
];
1187 wait_queue_head_t wait
; /* queue for waiting for
1188 writers / faults to finish */
1189 int frozen
; /* Is sb frozen? */
1190 wait_queue_head_t wait_unfrozen
; /* queue for waiting for
1192 #ifdef CONFIG_DEBUG_LOCK_ALLOC
1193 struct lockdep_map lock_map
[SB_FREEZE_LEVELS
];
1197 struct super_block
{
1198 struct list_head s_list
; /* Keep this first */
1199 dev_t s_dev
; /* search index; _not_ kdev_t */
1200 unsigned char s_blocksize_bits
;
1201 unsigned long s_blocksize
;
1202 loff_t s_maxbytes
; /* Max file size */
1203 struct file_system_type
*s_type
;
1204 const struct super_operations
*s_op
;
1205 const struct dquot_operations
*dq_op
;
1206 const struct quotactl_ops
*s_qcop
;
1207 const struct export_operations
*s_export_op
;
1208 unsigned long s_flags
;
1209 unsigned long s_magic
;
1210 struct dentry
*s_root
;
1211 struct rw_semaphore s_umount
;
1214 #ifdef CONFIG_SECURITY
1217 const struct xattr_handler
**s_xattr
;
1219 struct list_head s_inodes
; /* all inodes */
1220 struct hlist_bl_head s_anon
; /* anonymous dentries for (nfs) exporting */
1221 struct list_head s_mounts
; /* list of mounts; _not_ for fs use */
1222 struct block_device
*s_bdev
;
1223 struct backing_dev_info
*s_bdi
;
1224 struct mtd_info
*s_mtd
;
1225 struct hlist_node s_instances
;
1226 struct quota_info s_dquot
; /* Diskquota specific options */
1228 struct sb_writers s_writers
;
1230 char s_id
[32]; /* Informational name */
1231 u8 s_uuid
[16]; /* UUID */
1233 void *s_fs_info
; /* Filesystem private info */
1234 unsigned int s_max_links
;
1237 /* Granularity of c/m/atime in ns.
1238 Cannot be worse than a second */
1242 * The next field is for VFS *only*. No filesystems have any business
1243 * even looking at it. You had been warned.
1245 struct mutex s_vfs_rename_mutex
; /* Kludge */
1248 * Filesystem subtype. If non-empty the filesystem type field
1249 * in /proc/mounts will be "type.subtype"
1254 * Saved mount options for lazy filesystems using
1255 * generic_show_options()
1257 char __rcu
*s_options
;
1258 const struct dentry_operations
*s_d_op
; /* default d_op for dentries */
1261 * Saved pool identifier for cleancache (-1 means none)
1263 int cleancache_poolid
;
1265 struct shrinker s_shrink
; /* per-sb shrinker handle */
1267 /* Number of inodes with nlink == 0 but still referenced */
1268 atomic_long_t s_remove_count
;
1270 /* Being remounted read-only */
1271 int s_readonly_remount
;
1273 /* AIO completions deferred from interrupt context */
1274 struct workqueue_struct
*s_dio_done_wq
;
1277 * Keep the lru lists last in the structure so they always sit on their
1278 * own individual cachelines.
1280 struct list_lru s_dentry_lru ____cacheline_aligned_in_smp
;
1281 struct list_lru s_inode_lru ____cacheline_aligned_in_smp
;
1282 struct rcu_head rcu
;
1285 extern struct timespec
current_fs_time(struct super_block
*sb
);
1288 * Snapshotting support.
1291 void __sb_end_write(struct super_block
*sb
, int level
);
1292 int __sb_start_write(struct super_block
*sb
, int level
, bool wait
);
1295 * sb_end_write - drop write access to a superblock
1296 * @sb: the super we wrote to
1298 * Decrement number of writers to the filesystem. Wake up possible waiters
1299 * wanting to freeze the filesystem.
1301 static inline void sb_end_write(struct super_block
*sb
)
1303 __sb_end_write(sb
, SB_FREEZE_WRITE
);
1307 * sb_end_pagefault - drop write access to a superblock from a page fault
1308 * @sb: the super we wrote to
1310 * Decrement number of processes handling write page fault to the filesystem.
1311 * Wake up possible waiters wanting to freeze the filesystem.
1313 static inline void sb_end_pagefault(struct super_block
*sb
)
1315 __sb_end_write(sb
, SB_FREEZE_PAGEFAULT
);
1319 * sb_end_intwrite - drop write access to a superblock for internal fs purposes
1320 * @sb: the super we wrote to
1322 * Decrement fs-internal number of writers to the filesystem. Wake up possible
1323 * waiters wanting to freeze the filesystem.
1325 static inline void sb_end_intwrite(struct super_block
*sb
)
1327 __sb_end_write(sb
, SB_FREEZE_FS
);
1331 * sb_start_write - get write access to a superblock
1332 * @sb: the super we write to
1334 * When a process wants to write data or metadata to a file system (i.e. dirty
1335 * a page or an inode), it should embed the operation in a sb_start_write() -
1336 * sb_end_write() pair to get exclusion against file system freezing. This
1337 * function increments number of writers preventing freezing. If the file
1338 * system is already frozen, the function waits until the file system is
1341 * Since freeze protection behaves as a lock, users have to preserve
1342 * ordering of freeze protection and other filesystem locks. Generally,
1343 * freeze protection should be the outermost lock. In particular, we have:
1346 * -> i_mutex (write path, truncate, directory ops, ...)
1347 * -> s_umount (freeze_super, thaw_super)
1349 static inline void sb_start_write(struct super_block
*sb
)
1351 __sb_start_write(sb
, SB_FREEZE_WRITE
, true);
1354 static inline int sb_start_write_trylock(struct super_block
*sb
)
1356 return __sb_start_write(sb
, SB_FREEZE_WRITE
, false);
1360 * sb_start_pagefault - get write access to a superblock from a page fault
1361 * @sb: the super we write to
1363 * When a process starts handling write page fault, it should embed the
1364 * operation into sb_start_pagefault() - sb_end_pagefault() pair to get
1365 * exclusion against file system freezing. This is needed since the page fault
1366 * is going to dirty a page. This function increments number of running page
1367 * faults preventing freezing. If the file system is already frozen, the
1368 * function waits until the file system is thawed.
1370 * Since page fault freeze protection behaves as a lock, users have to preserve
1371 * ordering of freeze protection and other filesystem locks. It is advised to
1372 * put sb_start_pagefault() close to mmap_sem in lock ordering. Page fault
1373 * handling code implies lock dependency:
1376 * -> sb_start_pagefault
1378 static inline void sb_start_pagefault(struct super_block
*sb
)
1380 __sb_start_write(sb
, SB_FREEZE_PAGEFAULT
, true);
1384 * sb_start_intwrite - get write access to a superblock for internal fs purposes
1385 * @sb: the super we write to
1387 * This is the third level of protection against filesystem freezing. It is
1388 * free for use by a filesystem. The only requirement is that it must rank
1389 * below sb_start_pagefault.
1391 * For example filesystem can call sb_start_intwrite() when starting a
1392 * transaction which somewhat eases handling of freezing for internal sources
1393 * of filesystem changes (internal fs threads, discarding preallocation on file
1396 static inline void sb_start_intwrite(struct super_block
*sb
)
1398 __sb_start_write(sb
, SB_FREEZE_FS
, true);
1402 extern bool inode_owner_or_capable(const struct inode
*inode
);
1405 * VFS helper functions..
1407 extern int vfs_create(struct inode
*, struct dentry
*, umode_t
, bool);
1408 extern int vfs_mkdir(struct inode
*, struct dentry
*, umode_t
);
1409 extern int vfs_mknod(struct inode
*, struct dentry
*, umode_t
, dev_t
);
1410 extern int vfs_symlink(struct inode
*, struct dentry
*, const char *);
1411 extern int vfs_link(struct dentry
*, struct inode
*, struct dentry
*, struct inode
**);
1412 extern int vfs_rmdir(struct inode
*, struct dentry
*);
1413 extern int vfs_unlink(struct inode
*, struct dentry
*, struct inode
**);
1414 extern int vfs_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*, struct inode
**);
1417 * VFS dentry helper functions.
1419 extern void dentry_unhash(struct dentry
*dentry
);
1422 * VFS file helper functions.
1424 extern void inode_init_owner(struct inode
*inode
, const struct inode
*dir
,
1427 * VFS FS_IOC_FIEMAP helper definitions.
1429 struct fiemap_extent_info
{
1430 unsigned int fi_flags
; /* Flags as passed from user */
1431 unsigned int fi_extents_mapped
; /* Number of mapped extents */
1432 unsigned int fi_extents_max
; /* Size of fiemap_extent array */
1433 struct fiemap_extent __user
*fi_extents_start
; /* Start of
1434 fiemap_extent array */
1436 int fiemap_fill_next_extent(struct fiemap_extent_info
*info
, u64 logical
,
1437 u64 phys
, u64 len
, u32 flags
);
1438 int fiemap_check_flags(struct fiemap_extent_info
*fieinfo
, u32 fs_flags
);
1443 * NOTE! These match bits 12..15 of stat.st_mode
1444 * (ie "(i_mode >> 12) & 15").
1446 #define DT_UNKNOWN 0
1457 * This is the "filldir" function type, used by readdir() to let
1458 * the kernel specify what kind of dirent layout it wants to have.
1459 * This allows the kernel to read directories into kernel space or
1460 * to have different dirent layouts depending on the binary type.
1462 typedef int (*filldir_t
)(void *, const char *, int, loff_t
, u64
, unsigned);
1463 struct dir_context
{
1464 const filldir_t actor
;
1468 struct block_device_operations
;
1470 /* These macros are for out of kernel modules to test that
1471 * the kernel supports the unlocked_ioctl and compat_ioctl
1472 * fields in struct file_operations. */
1473 #define HAVE_COMPAT_IOCTL 1
1474 #define HAVE_UNLOCKED_IOCTL 1
1476 struct file_operations
{
1477 struct module
*owner
;
1478 loff_t (*llseek
) (struct file
*, loff_t
, int);
1479 ssize_t (*read
) (struct file
*, char __user
*, size_t, loff_t
*);
1480 ssize_t (*write
) (struct file
*, const char __user
*, size_t, loff_t
*);
1481 ssize_t (*aio_read
) (struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1482 ssize_t (*aio_write
) (struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1483 int (*iterate
) (struct file
*, struct dir_context
*);
1484 unsigned int (*poll
) (struct file
*, struct poll_table_struct
*);
1485 long (*unlocked_ioctl
) (struct file
*, unsigned int, unsigned long);
1486 long (*compat_ioctl
) (struct file
*, unsigned int, unsigned long);
1487 int (*mmap
) (struct file
*, struct vm_area_struct
*);
1488 int (*open
) (struct inode
*, struct file
*);
1489 int (*flush
) (struct file
*, fl_owner_t id
);
1490 int (*release
) (struct inode
*, struct file
*);
1491 int (*fsync
) (struct file
*, loff_t
, loff_t
, int datasync
);
1492 int (*aio_fsync
) (struct kiocb
*, int datasync
);
1493 int (*fasync
) (int, struct file
*, int);
1494 int (*lock
) (struct file
*, int, struct file_lock
*);
1495 ssize_t (*sendpage
) (struct file
*, struct page
*, int, size_t, loff_t
*, int);
1496 unsigned long (*get_unmapped_area
)(struct file
*, unsigned long, unsigned long, unsigned long, unsigned long);
1497 int (*check_flags
)(int);
1498 int (*flock
) (struct file
*, int, struct file_lock
*);
1499 ssize_t (*splice_write
)(struct pipe_inode_info
*, struct file
*, loff_t
*, size_t, unsigned int);
1500 ssize_t (*splice_read
)(struct file
*, loff_t
*, struct pipe_inode_info
*, size_t, unsigned int);
1501 int (*setlease
)(struct file
*, long, struct file_lock
**);
1502 long (*fallocate
)(struct file
*file
, int mode
, loff_t offset
,
1504 int (*show_fdinfo
)(struct seq_file
*m
, struct file
*f
);
1507 struct inode_operations
{
1508 struct dentry
* (*lookup
) (struct inode
*,struct dentry
*, unsigned int);
1509 void * (*follow_link
) (struct dentry
*, struct nameidata
*);
1510 int (*permission
) (struct inode
*, int);
1511 struct posix_acl
* (*get_acl
)(struct inode
*, int);
1513 int (*readlink
) (struct dentry
*, char __user
*,int);
1514 void (*put_link
) (struct dentry
*, struct nameidata
*, void *);
1516 int (*create
) (struct inode
*,struct dentry
*, umode_t
, bool);
1517 int (*link
) (struct dentry
*,struct inode
*,struct dentry
*);
1518 int (*unlink
) (struct inode
*,struct dentry
*);
1519 int (*symlink
) (struct inode
*,struct dentry
*,const char *);
1520 int (*mkdir
) (struct inode
*,struct dentry
*,umode_t
);
1521 int (*rmdir
) (struct inode
*,struct dentry
*);
1522 int (*mknod
) (struct inode
*,struct dentry
*,umode_t
,dev_t
);
1523 int (*rename
) (struct inode
*, struct dentry
*,
1524 struct inode
*, struct dentry
*);
1525 int (*setattr
) (struct dentry
*, struct iattr
*);
1526 int (*getattr
) (struct vfsmount
*mnt
, struct dentry
*, struct kstat
*);
1527 int (*setxattr
) (struct dentry
*, const char *,const void *,size_t,int);
1528 ssize_t (*getxattr
) (struct dentry
*, const char *, void *, size_t);
1529 ssize_t (*listxattr
) (struct dentry
*, char *, size_t);
1530 int (*removexattr
) (struct dentry
*, const char *);
1531 int (*fiemap
)(struct inode
*, struct fiemap_extent_info
*, u64 start
,
1533 int (*update_time
)(struct inode
*, struct timespec
*, int);
1534 int (*atomic_open
)(struct inode
*, struct dentry
*,
1535 struct file
*, unsigned open_flag
,
1536 umode_t create_mode
, int *opened
);
1537 int (*tmpfile
) (struct inode
*, struct dentry
*, umode_t
);
1538 int (*set_acl
)(struct inode
*, struct posix_acl
*, int);
1539 } ____cacheline_aligned
;
1541 ssize_t
rw_copy_check_uvector(int type
, const struct iovec __user
* uvector
,
1542 unsigned long nr_segs
, unsigned long fast_segs
,
1543 struct iovec
*fast_pointer
,
1544 struct iovec
**ret_pointer
);
1546 extern ssize_t
vfs_read(struct file
*, char __user
*, size_t, loff_t
*);
1547 extern ssize_t
vfs_write(struct file
*, const char __user
*, size_t, loff_t
*);
1548 extern ssize_t
vfs_readv(struct file
*, const struct iovec __user
*,
1549 unsigned long, loff_t
*);
1550 extern ssize_t
vfs_writev(struct file
*, const struct iovec __user
*,
1551 unsigned long, loff_t
*);
1553 struct super_operations
{
1554 struct inode
*(*alloc_inode
)(struct super_block
*sb
);
1555 void (*destroy_inode
)(struct inode
*);
1557 void (*dirty_inode
) (struct inode
*, int flags
);
1558 int (*write_inode
) (struct inode
*, struct writeback_control
*wbc
);
1559 int (*drop_inode
) (struct inode
*);
1560 void (*evict_inode
) (struct inode
*);
1561 void (*put_super
) (struct super_block
*);
1562 int (*sync_fs
)(struct super_block
*sb
, int wait
);
1563 int (*freeze_fs
) (struct super_block
*);
1564 int (*unfreeze_fs
) (struct super_block
*);
1565 int (*statfs
) (struct dentry
*, struct kstatfs
*);
1566 int (*remount_fs
) (struct super_block
*, int *, char *);
1567 void (*umount_begin
) (struct super_block
*);
1569 int (*show_options
)(struct seq_file
*, struct dentry
*);
1570 int (*show_devname
)(struct seq_file
*, struct dentry
*);
1571 int (*show_path
)(struct seq_file
*, struct dentry
*);
1572 int (*show_stats
)(struct seq_file
*, struct dentry
*);
1574 ssize_t (*quota_read
)(struct super_block
*, int, char *, size_t, loff_t
);
1575 ssize_t (*quota_write
)(struct super_block
*, int, const char *, size_t, loff_t
);
1577 int (*bdev_try_to_free_page
)(struct super_block
*, struct page
*, gfp_t
);
1578 long (*nr_cached_objects
)(struct super_block
*, int);
1579 long (*free_cached_objects
)(struct super_block
*, long, int);
1583 * Inode flags - they have no relation to superblock flags now
1585 #define S_SYNC 1 /* Writes are synced at once */
1586 #define S_NOATIME 2 /* Do not update access times */
1587 #define S_APPEND 4 /* Append-only file */
1588 #define S_IMMUTABLE 8 /* Immutable file */
1589 #define S_DEAD 16 /* removed, but still open directory */
1590 #define S_NOQUOTA 32 /* Inode is not counted to quota */
1591 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
1592 #define S_NOCMTIME 128 /* Do not update file c/mtime */
1593 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
1594 #define S_PRIVATE 512 /* Inode is fs-internal */
1595 #define S_IMA 1024 /* Inode has an associated IMA struct */
1596 #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
1597 #define S_NOSEC 4096 /* no suid or xattr security attributes */
1600 * Note that nosuid etc flags are inode-specific: setting some file-system
1601 * flags just means all the inodes inherit those flags by default. It might be
1602 * possible to override it selectively if you really wanted to with some
1603 * ioctl() that is not currently implemented.
1605 * Exception: MS_RDONLY is always applied to the entire file system.
1607 * Unfortunately, it is possible to change a filesystems flags with it mounted
1608 * with files in use. This means that all of the inodes will not have their
1609 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
1610 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
1612 #define __IS_FLG(inode, flg) ((inode)->i_sb->s_flags & (flg))
1614 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
1615 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
1616 ((inode)->i_flags & S_SYNC))
1617 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
1618 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
1619 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
1620 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
1621 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
1623 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
1624 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
1625 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
1626 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
1628 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
1629 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
1630 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
1631 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
1632 #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
1633 #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
1634 #define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
1637 * Inode state bits. Protected by inode->i_lock
1639 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1640 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1642 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1643 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1644 * various stages of removing an inode.
1646 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1648 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1649 * fdatasync(). i_atime is the usual cause.
1650 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1651 * these changes separately from I_DIRTY_SYNC so that we
1652 * don't have to write inode on fdatasync() when only
1653 * mtime has changed in it.
1654 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1655 * I_NEW Serves as both a mutex and completion notification.
1656 * New inodes set I_NEW. If two processes both create
1657 * the same inode, one of them will release its inode and
1658 * wait for I_NEW to be released before returning.
1659 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1660 * also cause waiting on I_NEW, without I_NEW actually
1661 * being set. find_inode() uses this to prevent returning
1662 * nearly-dead inodes.
1663 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1664 * is zero. I_FREEING must be set when I_WILL_FREE is
1666 * I_FREEING Set when inode is about to be freed but still has dirty
1667 * pages or buffers attached or the inode itself is still
1669 * I_CLEAR Added by clear_inode(). In this state the inode is
1670 * clean and can be destroyed. Inode keeps I_FREEING.
1672 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1673 * prohibited for many purposes. iget() must wait for
1674 * the inode to be completely released, then create it
1675 * anew. Other functions will just ignore such inodes,
1676 * if appropriate. I_NEW is used for waiting.
1678 * I_SYNC Writeback of inode is running. The bit is set during
1679 * data writeback, and cleared with a wakeup on the bit
1680 * address once it is done. The bit is also used to pin
1681 * the inode in memory for flusher thread.
1683 * I_REFERENCED Marks the inode as recently references on the LRU list.
1685 * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit().
1687 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1689 #define I_DIRTY_SYNC (1 << 0)
1690 #define I_DIRTY_DATASYNC (1 << 1)
1691 #define I_DIRTY_PAGES (1 << 2)
1693 #define I_NEW (1 << __I_NEW)
1694 #define I_WILL_FREE (1 << 4)
1695 #define I_FREEING (1 << 5)
1696 #define I_CLEAR (1 << 6)
1698 #define I_SYNC (1 << __I_SYNC)
1699 #define I_REFERENCED (1 << 8)
1700 #define __I_DIO_WAKEUP 9
1701 #define I_DIO_WAKEUP (1 << I_DIO_WAKEUP)
1702 #define I_LINKABLE (1 << 10)
1704 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1706 extern void __mark_inode_dirty(struct inode
*, int);
1707 static inline void mark_inode_dirty(struct inode
*inode
)
1709 __mark_inode_dirty(inode
, I_DIRTY
);
1712 static inline void mark_inode_dirty_sync(struct inode
*inode
)
1714 __mark_inode_dirty(inode
, I_DIRTY_SYNC
);
1717 extern void inc_nlink(struct inode
*inode
);
1718 extern void drop_nlink(struct inode
*inode
);
1719 extern void clear_nlink(struct inode
*inode
);
1720 extern void set_nlink(struct inode
*inode
, unsigned int nlink
);
1722 static inline void inode_inc_link_count(struct inode
*inode
)
1725 mark_inode_dirty(inode
);
1728 static inline void inode_dec_link_count(struct inode
*inode
)
1731 mark_inode_dirty(inode
);
1735 * inode_inc_iversion - increments i_version
1736 * @inode: inode that need to be updated
1738 * Every time the inode is modified, the i_version field will be incremented.
1739 * The filesystem has to be mounted with i_version flag
1742 static inline void inode_inc_iversion(struct inode
*inode
)
1744 spin_lock(&inode
->i_lock
);
1746 spin_unlock(&inode
->i_lock
);
1749 enum file_time_flags
{
1756 extern void touch_atime(const struct path
*);
1757 static inline void file_accessed(struct file
*file
)
1759 if (!(file
->f_flags
& O_NOATIME
))
1760 touch_atime(&file
->f_path
);
1763 int sync_inode(struct inode
*inode
, struct writeback_control
*wbc
);
1764 int sync_inode_metadata(struct inode
*inode
, int wait
);
1766 struct file_system_type
{
1769 #define FS_REQUIRES_DEV 1
1770 #define FS_BINARY_MOUNTDATA 2
1771 #define FS_HAS_SUBTYPE 4
1772 #define FS_USERNS_MOUNT 8 /* Can be mounted by userns root */
1773 #define FS_USERNS_DEV_MOUNT 16 /* A userns mount does not imply MNT_NODEV */
1774 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() during rename() internally. */
1775 struct dentry
*(*mount
) (struct file_system_type
*, int,
1776 const char *, void *);
1777 void (*kill_sb
) (struct super_block
*);
1778 struct module
*owner
;
1779 struct file_system_type
* next
;
1780 struct hlist_head fs_supers
;
1782 struct lock_class_key s_lock_key
;
1783 struct lock_class_key s_umount_key
;
1784 struct lock_class_key s_vfs_rename_key
;
1785 struct lock_class_key s_writers_key
[SB_FREEZE_LEVELS
];
1787 struct lock_class_key i_lock_key
;
1788 struct lock_class_key i_mutex_key
;
1789 struct lock_class_key i_mutex_dir_key
;
1792 #define MODULE_ALIAS_FS(NAME) MODULE_ALIAS("fs-" NAME)
1794 extern struct dentry
*mount_ns(struct file_system_type
*fs_type
, int flags
,
1795 void *data
, int (*fill_super
)(struct super_block
*, void *, int));
1796 extern struct dentry
*mount_bdev(struct file_system_type
*fs_type
,
1797 int flags
, const char *dev_name
, void *data
,
1798 int (*fill_super
)(struct super_block
*, void *, int));
1799 extern struct dentry
*mount_single(struct file_system_type
*fs_type
,
1800 int flags
, void *data
,
1801 int (*fill_super
)(struct super_block
*, void *, int));
1802 extern struct dentry
*mount_nodev(struct file_system_type
*fs_type
,
1803 int flags
, void *data
,
1804 int (*fill_super
)(struct super_block
*, void *, int));
1805 extern struct dentry
*mount_subtree(struct vfsmount
*mnt
, const char *path
);
1806 void generic_shutdown_super(struct super_block
*sb
);
1807 void kill_block_super(struct super_block
*sb
);
1808 void kill_anon_super(struct super_block
*sb
);
1809 void kill_litter_super(struct super_block
*sb
);
1810 void deactivate_super(struct super_block
*sb
);
1811 void deactivate_locked_super(struct super_block
*sb
);
1812 int set_anon_super(struct super_block
*s
, void *data
);
1813 int get_anon_bdev(dev_t
*);
1814 void free_anon_bdev(dev_t
);
1815 struct super_block
*sget(struct file_system_type
*type
,
1816 int (*test
)(struct super_block
*,void *),
1817 int (*set
)(struct super_block
*,void *),
1818 int flags
, void *data
);
1819 extern struct dentry
*mount_pseudo(struct file_system_type
*, char *,
1820 const struct super_operations
*ops
,
1821 const struct dentry_operations
*dops
,
1824 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1825 #define fops_get(fops) \
1826 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1827 #define fops_put(fops) \
1828 do { if (fops) module_put((fops)->owner); } while(0)
1830 * This one is to be used *ONLY* from ->open() instances.
1831 * fops must be non-NULL, pinned down *and* module dependencies
1832 * should be sufficient to pin the caller down as well.
1834 #define replace_fops(f, fops) \
1836 struct file *__file = (f); \
1837 fops_put(__file->f_op); \
1838 BUG_ON(!(__file->f_op = (fops))); \
1841 extern int register_filesystem(struct file_system_type
*);
1842 extern int unregister_filesystem(struct file_system_type
*);
1843 extern struct vfsmount
*kern_mount_data(struct file_system_type
*, void *data
);
1844 #define kern_mount(type) kern_mount_data(type, NULL)
1845 extern void kern_unmount(struct vfsmount
*mnt
);
1846 extern int may_umount_tree(struct vfsmount
*);
1847 extern int may_umount(struct vfsmount
*);
1848 extern long do_mount(const char *, const char *, const char *, unsigned long, void *);
1849 extern struct vfsmount
*collect_mounts(struct path
*);
1850 extern void drop_collected_mounts(struct vfsmount
*);
1851 extern int iterate_mounts(int (*)(struct vfsmount
*, void *), void *,
1853 extern int vfs_statfs(struct path
*, struct kstatfs
*);
1854 extern int user_statfs(const char __user
*, struct kstatfs
*);
1855 extern int fd_statfs(int, struct kstatfs
*);
1856 extern int vfs_ustat(dev_t
, struct kstatfs
*);
1857 extern int freeze_super(struct super_block
*super
);
1858 extern int thaw_super(struct super_block
*super
);
1859 extern bool our_mnt(struct vfsmount
*mnt
);
1860 extern bool fs_fully_visible(struct file_system_type
*);
1862 extern int current_umask(void);
1864 extern void ihold(struct inode
* inode
);
1865 extern void iput(struct inode
*);
1868 extern struct kobject
*fs_kobj
;
1870 #define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
1872 #define FLOCK_VERIFY_READ 1
1873 #define FLOCK_VERIFY_WRITE 2
1875 #ifdef CONFIG_FILE_LOCKING
1876 extern int locks_mandatory_locked(struct inode
*);
1877 extern int locks_mandatory_area(int, struct inode
*, struct file
*, loff_t
, size_t);
1880 * Candidates for mandatory locking have the setgid bit set
1881 * but no group execute bit - an otherwise meaningless combination.
1884 static inline int __mandatory_lock(struct inode
*ino
)
1886 return (ino
->i_mode
& (S_ISGID
| S_IXGRP
)) == S_ISGID
;
1890 * ... and these candidates should be on MS_MANDLOCK mounted fs,
1891 * otherwise these will be advisory locks
1894 static inline int mandatory_lock(struct inode
*ino
)
1896 return IS_MANDLOCK(ino
) && __mandatory_lock(ino
);
1899 static inline int locks_verify_locked(struct inode
*inode
)
1901 if (mandatory_lock(inode
))
1902 return locks_mandatory_locked(inode
);
1906 static inline int locks_verify_truncate(struct inode
*inode
,
1910 if (inode
->i_flock
&& mandatory_lock(inode
))
1911 return locks_mandatory_area(
1912 FLOCK_VERIFY_WRITE
, inode
, filp
,
1913 size
< inode
->i_size
? size
: inode
->i_size
,
1914 (size
< inode
->i_size
? inode
->i_size
- size
1915 : size
- inode
->i_size
)
1920 static inline int break_lease(struct inode
*inode
, unsigned int mode
)
1923 return __break_lease(inode
, mode
, FL_LEASE
);
1927 static inline int break_deleg(struct inode
*inode
, unsigned int mode
)
1930 return __break_lease(inode
, mode
, FL_DELEG
);
1934 static inline int try_break_deleg(struct inode
*inode
, struct inode
**delegated_inode
)
1938 ret
= break_deleg(inode
, O_WRONLY
|O_NONBLOCK
);
1939 if (ret
== -EWOULDBLOCK
&& delegated_inode
) {
1940 *delegated_inode
= inode
;
1946 static inline int break_deleg_wait(struct inode
**delegated_inode
)
1950 ret
= break_deleg(*delegated_inode
, O_WRONLY
);
1951 iput(*delegated_inode
);
1952 *delegated_inode
= NULL
;
1956 #else /* !CONFIG_FILE_LOCKING */
1957 static inline int locks_mandatory_locked(struct inode
*inode
)
1962 static inline int locks_mandatory_area(int rw
, struct inode
*inode
,
1963 struct file
*filp
, loff_t offset
,
1969 static inline int __mandatory_lock(struct inode
*inode
)
1974 static inline int mandatory_lock(struct inode
*inode
)
1979 static inline int locks_verify_locked(struct inode
*inode
)
1984 static inline int locks_verify_truncate(struct inode
*inode
, struct file
*filp
,
1990 static inline int break_lease(struct inode
*inode
, unsigned int mode
)
1995 static inline int break_deleg(struct inode
*inode
, unsigned int mode
)
2000 static inline int try_break_deleg(struct inode
*inode
, struct inode
**delegated_inode
)
2005 static inline int break_deleg_wait(struct inode
**delegated_inode
)
2011 #endif /* CONFIG_FILE_LOCKING */
2016 const char *name
; /* pointer to actual string */
2017 const __user
char *uptr
; /* original userland pointer */
2018 struct audit_names
*aname
;
2019 bool separate
; /* should "name" be freed? */
2022 extern long vfs_truncate(struct path
*, loff_t
);
2023 extern int do_truncate(struct dentry
*, loff_t start
, unsigned int time_attrs
,
2025 extern int do_fallocate(struct file
*file
, int mode
, loff_t offset
,
2027 extern long do_sys_open(int dfd
, const char __user
*filename
, int flags
,
2029 extern struct file
*file_open_name(struct filename
*, int, umode_t
);
2030 extern struct file
*filp_open(const char *, int, umode_t
);
2031 extern struct file
*file_open_root(struct dentry
*, struct vfsmount
*,
2033 extern struct file
* dentry_open(const struct path
*, int, const struct cred
*);
2034 extern int filp_close(struct file
*, fl_owner_t id
);
2036 extern struct filename
*getname(const char __user
*);
2037 extern struct filename
*getname_kernel(const char *);
2043 extern int finish_open(struct file
*file
, struct dentry
*dentry
,
2044 int (*open
)(struct inode
*, struct file
*),
2046 extern int finish_no_open(struct file
*file
, struct dentry
*dentry
);
2050 extern int ioctl_preallocate(struct file
*filp
, void __user
*argp
);
2053 extern void __init
vfs_caches_init_early(void);
2054 extern void __init
vfs_caches_init(unsigned long);
2056 extern struct kmem_cache
*names_cachep
;
2058 extern void final_putname(struct filename
*name
);
2060 #define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL)
2061 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
2062 #ifndef CONFIG_AUDITSYSCALL
2063 #define putname(name) final_putname(name)
2065 extern void putname(struct filename
*name
);
2069 extern int register_blkdev(unsigned int, const char *);
2070 extern void unregister_blkdev(unsigned int, const char *);
2071 extern struct block_device
*bdget(dev_t
);
2072 extern struct block_device
*bdgrab(struct block_device
*bdev
);
2073 extern void bd_set_size(struct block_device
*, loff_t size
);
2074 extern void bd_forget(struct inode
*inode
);
2075 extern void bdput(struct block_device
*);
2076 extern void invalidate_bdev(struct block_device
*);
2077 extern void iterate_bdevs(void (*)(struct block_device
*, void *), void *);
2078 extern int sync_blockdev(struct block_device
*bdev
);
2079 extern void kill_bdev(struct block_device
*);
2080 extern struct super_block
*freeze_bdev(struct block_device
*);
2081 extern void emergency_thaw_all(void);
2082 extern int thaw_bdev(struct block_device
*bdev
, struct super_block
*sb
);
2083 extern int fsync_bdev(struct block_device
*);
2084 extern int sb_is_blkdev_sb(struct super_block
*sb
);
2086 static inline void bd_forget(struct inode
*inode
) {}
2087 static inline int sync_blockdev(struct block_device
*bdev
) { return 0; }
2088 static inline void kill_bdev(struct block_device
*bdev
) {}
2089 static inline void invalidate_bdev(struct block_device
*bdev
) {}
2091 static inline struct super_block
*freeze_bdev(struct block_device
*sb
)
2096 static inline int thaw_bdev(struct block_device
*bdev
, struct super_block
*sb
)
2101 static inline void iterate_bdevs(void (*f
)(struct block_device
*, void *), void *arg
)
2105 static inline int sb_is_blkdev_sb(struct super_block
*sb
)
2110 extern int sync_filesystem(struct super_block
*);
2111 extern const struct file_operations def_blk_fops
;
2112 extern const struct file_operations def_chr_fops
;
2113 extern const struct file_operations bad_sock_fops
;
2115 extern int ioctl_by_bdev(struct block_device
*, unsigned, unsigned long);
2116 extern int blkdev_ioctl(struct block_device
*, fmode_t
, unsigned, unsigned long);
2117 extern long compat_blkdev_ioctl(struct file
*, unsigned, unsigned long);
2118 extern int blkdev_get(struct block_device
*bdev
, fmode_t mode
, void *holder
);
2119 extern struct block_device
*blkdev_get_by_path(const char *path
, fmode_t mode
,
2121 extern struct block_device
*blkdev_get_by_dev(dev_t dev
, fmode_t mode
,
2123 extern void blkdev_put(struct block_device
*bdev
, fmode_t mode
);
2125 extern int bd_link_disk_holder(struct block_device
*bdev
, struct gendisk
*disk
);
2126 extern void bd_unlink_disk_holder(struct block_device
*bdev
,
2127 struct gendisk
*disk
);
2129 static inline int bd_link_disk_holder(struct block_device
*bdev
,
2130 struct gendisk
*disk
)
2134 static inline void bd_unlink_disk_holder(struct block_device
*bdev
,
2135 struct gendisk
*disk
)
2142 #define CHRDEV_MAJOR_HASH_SIZE 255
2143 extern int alloc_chrdev_region(dev_t
*, unsigned, unsigned, const char *);
2144 extern int register_chrdev_region(dev_t
, unsigned, const char *);
2145 extern int __register_chrdev(unsigned int major
, unsigned int baseminor
,
2146 unsigned int count
, const char *name
,
2147 const struct file_operations
*fops
);
2148 extern void __unregister_chrdev(unsigned int major
, unsigned int baseminor
,
2149 unsigned int count
, const char *name
);
2150 extern void unregister_chrdev_region(dev_t
, unsigned);
2151 extern void chrdev_show(struct seq_file
*,off_t
);
2153 static inline int register_chrdev(unsigned int major
, const char *name
,
2154 const struct file_operations
*fops
)
2156 return __register_chrdev(major
, 0, 256, name
, fops
);
2159 static inline void unregister_chrdev(unsigned int major
, const char *name
)
2161 __unregister_chrdev(major
, 0, 256, name
);
2164 /* fs/block_dev.c */
2165 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
2166 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
2169 #define BLKDEV_MAJOR_HASH_SIZE 255
2170 extern const char *__bdevname(dev_t
, char *buffer
);
2171 extern const char *bdevname(struct block_device
*bdev
, char *buffer
);
2172 extern struct block_device
*lookup_bdev(const char *);
2173 extern void blkdev_show(struct seq_file
*,off_t
);
2176 #define BLKDEV_MAJOR_HASH_SIZE 0
2179 extern void init_special_inode(struct inode
*, umode_t
, dev_t
);
2181 /* Invalid inode operations -- fs/bad_inode.c */
2182 extern void make_bad_inode(struct inode
*);
2183 extern int is_bad_inode(struct inode
*);
2187 * return READ, READA, or WRITE
2189 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2192 * return data direction, READ or WRITE
2194 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
2196 extern void check_disk_size_change(struct gendisk
*disk
,
2197 struct block_device
*bdev
);
2198 extern int revalidate_disk(struct gendisk
*);
2199 extern int check_disk_change(struct block_device
*);
2200 extern int __invalidate_device(struct block_device
*, bool);
2201 extern int invalidate_partition(struct gendisk
*, int);
2203 unsigned long invalidate_mapping_pages(struct address_space
*mapping
,
2204 pgoff_t start
, pgoff_t end
);
2206 static inline void invalidate_remote_inode(struct inode
*inode
)
2208 if (S_ISREG(inode
->i_mode
) || S_ISDIR(inode
->i_mode
) ||
2209 S_ISLNK(inode
->i_mode
))
2210 invalidate_mapping_pages(inode
->i_mapping
, 0, -1);
2212 extern int invalidate_inode_pages2(struct address_space
*mapping
);
2213 extern int invalidate_inode_pages2_range(struct address_space
*mapping
,
2214 pgoff_t start
, pgoff_t end
);
2215 extern int write_inode_now(struct inode
*, int);
2216 extern int filemap_fdatawrite(struct address_space
*);
2217 extern int filemap_flush(struct address_space
*);
2218 extern int filemap_fdatawait(struct address_space
*);
2219 extern int filemap_fdatawait_range(struct address_space
*, loff_t lstart
,
2221 extern int filemap_write_and_wait(struct address_space
*mapping
);
2222 extern int filemap_write_and_wait_range(struct address_space
*mapping
,
2223 loff_t lstart
, loff_t lend
);
2224 extern int __filemap_fdatawrite_range(struct address_space
*mapping
,
2225 loff_t start
, loff_t end
, int sync_mode
);
2226 extern int filemap_fdatawrite_range(struct address_space
*mapping
,
2227 loff_t start
, loff_t end
);
2229 extern int vfs_fsync_range(struct file
*file
, loff_t start
, loff_t end
,
2231 extern int vfs_fsync(struct file
*file
, int datasync
);
2232 static inline int generic_write_sync(struct file
*file
, loff_t pos
, loff_t count
)
2234 if (!(file
->f_flags
& O_DSYNC
) && !IS_SYNC(file
->f_mapping
->host
))
2236 return vfs_fsync_range(file
, pos
, pos
+ count
- 1,
2237 (file
->f_flags
& __O_SYNC
) ? 0 : 1);
2239 extern void emergency_sync(void);
2240 extern void emergency_remount(void);
2242 extern sector_t
bmap(struct inode
*, sector_t
);
2244 extern int notify_change(struct dentry
*, struct iattr
*, struct inode
**);
2245 extern int inode_permission(struct inode
*, int);
2246 extern int generic_permission(struct inode
*, int);
2248 static inline bool execute_ok(struct inode
*inode
)
2250 return (inode
->i_mode
& S_IXUGO
) || S_ISDIR(inode
->i_mode
);
2253 static inline struct inode
*file_inode(struct file
*f
)
2258 static inline void file_start_write(struct file
*file
)
2260 if (!S_ISREG(file_inode(file
)->i_mode
))
2262 __sb_start_write(file_inode(file
)->i_sb
, SB_FREEZE_WRITE
, true);
2265 static inline bool file_start_write_trylock(struct file
*file
)
2267 if (!S_ISREG(file_inode(file
)->i_mode
))
2269 return __sb_start_write(file_inode(file
)->i_sb
, SB_FREEZE_WRITE
, false);
2272 static inline void file_end_write(struct file
*file
)
2274 if (!S_ISREG(file_inode(file
)->i_mode
))
2276 __sb_end_write(file_inode(file
)->i_sb
, SB_FREEZE_WRITE
);
2280 * get_write_access() gets write permission for a file.
2281 * put_write_access() releases this write permission.
2282 * This is used for regular files.
2283 * We cannot support write (and maybe mmap read-write shared) accesses and
2284 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2285 * can have the following values:
2286 * 0: no writers, no VM_DENYWRITE mappings
2287 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2288 * > 0: (i_writecount) users are writing to the file.
2290 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2291 * except for the cases where we don't hold i_writecount yet. Then we need to
2292 * use {get,deny}_write_access() - these functions check the sign and refuse
2293 * to do the change if sign is wrong.
2295 static inline int get_write_access(struct inode
*inode
)
2297 return atomic_inc_unless_negative(&inode
->i_writecount
) ? 0 : -ETXTBSY
;
2299 static inline int deny_write_access(struct file
*file
)
2301 struct inode
*inode
= file_inode(file
);
2302 return atomic_dec_unless_positive(&inode
->i_writecount
) ? 0 : -ETXTBSY
;
2304 static inline void put_write_access(struct inode
* inode
)
2306 atomic_dec(&inode
->i_writecount
);
2308 static inline void allow_write_access(struct file
*file
)
2311 atomic_inc(&file_inode(file
)->i_writecount
);
2313 static inline bool inode_is_open_for_write(const struct inode
*inode
)
2315 return atomic_read(&inode
->i_writecount
) > 0;
2319 static inline void i_readcount_dec(struct inode
*inode
)
2321 BUG_ON(!atomic_read(&inode
->i_readcount
));
2322 atomic_dec(&inode
->i_readcount
);
2324 static inline void i_readcount_inc(struct inode
*inode
)
2326 atomic_inc(&inode
->i_readcount
);
2329 static inline void i_readcount_dec(struct inode
*inode
)
2333 static inline void i_readcount_inc(struct inode
*inode
)
2338 extern int do_pipe_flags(int *, int);
2340 extern int kernel_read(struct file
*, loff_t
, char *, unsigned long);
2341 extern ssize_t
kernel_write(struct file
*, const char *, size_t, loff_t
);
2342 extern struct file
* open_exec(const char *);
2344 /* fs/dcache.c -- generic fs support functions */
2345 extern int is_subdir(struct dentry
*, struct dentry
*);
2346 extern int path_is_under(struct path
*, struct path
*);
2348 #include <linux/err.h>
2350 /* needed for stackable file system support */
2351 extern loff_t
default_llseek(struct file
*file
, loff_t offset
, int whence
);
2353 extern loff_t
vfs_llseek(struct file
*file
, loff_t offset
, int whence
);
2355 extern int inode_init_always(struct super_block
*, struct inode
*);
2356 extern void inode_init_once(struct inode
*);
2357 extern void address_space_init_once(struct address_space
*mapping
);
2358 extern struct inode
* igrab(struct inode
*);
2359 extern ino_t
iunique(struct super_block
*, ino_t
);
2360 extern int inode_needs_sync(struct inode
*inode
);
2361 extern int generic_delete_inode(struct inode
*inode
);
2362 static inline int generic_drop_inode(struct inode
*inode
)
2364 return !inode
->i_nlink
|| inode_unhashed(inode
);
2367 extern struct inode
*ilookup5_nowait(struct super_block
*sb
,
2368 unsigned long hashval
, int (*test
)(struct inode
*, void *),
2370 extern struct inode
*ilookup5(struct super_block
*sb
, unsigned long hashval
,
2371 int (*test
)(struct inode
*, void *), void *data
);
2372 extern struct inode
*ilookup(struct super_block
*sb
, unsigned long ino
);
2374 extern struct inode
* iget5_locked(struct super_block
*, unsigned long, int (*test
)(struct inode
*, void *), int (*set
)(struct inode
*, void *), void *);
2375 extern struct inode
* iget_locked(struct super_block
*, unsigned long);
2376 extern int insert_inode_locked4(struct inode
*, unsigned long, int (*test
)(struct inode
*, void *), void *);
2377 extern int insert_inode_locked(struct inode
*);
2378 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2379 extern void lockdep_annotate_inode_mutex_key(struct inode
*inode
);
2381 static inline void lockdep_annotate_inode_mutex_key(struct inode
*inode
) { };
2383 extern void unlock_new_inode(struct inode
*);
2384 extern unsigned int get_next_ino(void);
2386 extern void __iget(struct inode
* inode
);
2387 extern void iget_failed(struct inode
*);
2388 extern void clear_inode(struct inode
*);
2389 extern void __destroy_inode(struct inode
*);
2390 extern struct inode
*new_inode_pseudo(struct super_block
*sb
);
2391 extern struct inode
*new_inode(struct super_block
*sb
);
2392 extern void free_inode_nonrcu(struct inode
*inode
);
2393 extern int should_remove_suid(struct dentry
*);
2394 extern int file_remove_suid(struct file
*);
2396 extern void __insert_inode_hash(struct inode
*, unsigned long hashval
);
2397 static inline void insert_inode_hash(struct inode
*inode
)
2399 __insert_inode_hash(inode
, inode
->i_ino
);
2402 extern void __remove_inode_hash(struct inode
*);
2403 static inline void remove_inode_hash(struct inode
*inode
)
2405 if (!inode_unhashed(inode
))
2406 __remove_inode_hash(inode
);
2409 extern void inode_sb_list_add(struct inode
*inode
);
2412 extern void submit_bio(int, struct bio
*);
2413 extern int bdev_read_only(struct block_device
*);
2415 extern int set_blocksize(struct block_device
*, int);
2416 extern int sb_set_blocksize(struct super_block
*, int);
2417 extern int sb_min_blocksize(struct super_block
*, int);
2419 extern int generic_file_mmap(struct file
*, struct vm_area_struct
*);
2420 extern int generic_file_readonly_mmap(struct file
*, struct vm_area_struct
*);
2421 extern int generic_file_remap_pages(struct vm_area_struct
*, unsigned long addr
,
2422 unsigned long size
, pgoff_t pgoff
);
2423 extern int file_read_actor(read_descriptor_t
* desc
, struct page
*page
, unsigned long offset
, unsigned long size
);
2424 int generic_write_checks(struct file
*file
, loff_t
*pos
, size_t *count
, int isblk
);
2425 extern ssize_t
generic_file_aio_read(struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
2426 extern ssize_t
__generic_file_aio_write(struct kiocb
*, const struct iovec
*, unsigned long,
2428 extern ssize_t
generic_file_aio_write(struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
2429 extern ssize_t
generic_file_direct_write(struct kiocb
*, const struct iovec
*,
2430 unsigned long *, loff_t
, loff_t
*, size_t, size_t);
2431 extern ssize_t
generic_file_buffered_write(struct kiocb
*, const struct iovec
*,
2432 unsigned long, loff_t
, loff_t
*, size_t, ssize_t
);
2433 extern ssize_t
do_sync_read(struct file
*filp
, char __user
*buf
, size_t len
, loff_t
*ppos
);
2434 extern ssize_t
do_sync_write(struct file
*filp
, const char __user
*buf
, size_t len
, loff_t
*ppos
);
2435 extern int generic_segment_checks(const struct iovec
*iov
,
2436 unsigned long *nr_segs
, size_t *count
, int access_flags
);
2438 /* fs/block_dev.c */
2439 extern ssize_t
blkdev_aio_write(struct kiocb
*iocb
, const struct iovec
*iov
,
2440 unsigned long nr_segs
, loff_t pos
);
2441 extern int blkdev_fsync(struct file
*filp
, loff_t start
, loff_t end
,
2443 extern void block_sync_page(struct page
*page
);
2446 extern ssize_t
generic_file_splice_read(struct file
*, loff_t
*,
2447 struct pipe_inode_info
*, size_t, unsigned int);
2448 extern ssize_t
default_file_splice_read(struct file
*, loff_t
*,
2449 struct pipe_inode_info
*, size_t, unsigned int);
2450 extern ssize_t
generic_file_splice_write(struct pipe_inode_info
*,
2451 struct file
*, loff_t
*, size_t, unsigned int);
2452 extern ssize_t
generic_splice_sendpage(struct pipe_inode_info
*pipe
,
2453 struct file
*out
, loff_t
*, size_t len
, unsigned int flags
);
2456 file_ra_state_init(struct file_ra_state
*ra
, struct address_space
*mapping
);
2457 extern loff_t
noop_llseek(struct file
*file
, loff_t offset
, int whence
);
2458 extern loff_t
no_llseek(struct file
*file
, loff_t offset
, int whence
);
2459 extern loff_t
vfs_setpos(struct file
*file
, loff_t offset
, loff_t maxsize
);
2460 extern loff_t
generic_file_llseek(struct file
*file
, loff_t offset
, int whence
);
2461 extern loff_t
generic_file_llseek_size(struct file
*file
, loff_t offset
,
2462 int whence
, loff_t maxsize
, loff_t eof
);
2463 extern loff_t
fixed_size_llseek(struct file
*file
, loff_t offset
,
2464 int whence
, loff_t size
);
2465 extern int generic_file_open(struct inode
* inode
, struct file
* filp
);
2466 extern int nonseekable_open(struct inode
* inode
, struct file
* filp
);
2468 #ifdef CONFIG_FS_XIP
2469 extern ssize_t
xip_file_read(struct file
*filp
, char __user
*buf
, size_t len
,
2471 extern int xip_file_mmap(struct file
* file
, struct vm_area_struct
* vma
);
2472 extern ssize_t
xip_file_write(struct file
*filp
, const char __user
*buf
,
2473 size_t len
, loff_t
*ppos
);
2474 extern int xip_truncate_page(struct address_space
*mapping
, loff_t from
);
2476 static inline int xip_truncate_page(struct address_space
*mapping
, loff_t from
)
2483 typedef void (dio_submit_t
)(int rw
, struct bio
*bio
, struct inode
*inode
,
2484 loff_t file_offset
);
2487 /* need locking between buffered and direct access */
2490 /* filesystem does not support filling holes */
2491 DIO_SKIP_HOLES
= 0x02,
2494 void dio_end_io(struct bio
*bio
, int error
);
2496 ssize_t
__blockdev_direct_IO(int rw
, struct kiocb
*iocb
, struct inode
*inode
,
2497 struct block_device
*bdev
, const struct iovec
*iov
, loff_t offset
,
2498 unsigned long nr_segs
, get_block_t get_block
, dio_iodone_t end_io
,
2499 dio_submit_t submit_io
, int flags
);
2501 static inline ssize_t
blockdev_direct_IO(int rw
, struct kiocb
*iocb
,
2502 struct inode
*inode
, const struct iovec
*iov
, loff_t offset
,
2503 unsigned long nr_segs
, get_block_t get_block
)
2505 return __blockdev_direct_IO(rw
, iocb
, inode
, inode
->i_sb
->s_bdev
, iov
,
2506 offset
, nr_segs
, get_block
, NULL
, NULL
,
2507 DIO_LOCKING
| DIO_SKIP_HOLES
);
2511 void inode_dio_wait(struct inode
*inode
);
2512 void inode_dio_done(struct inode
*inode
);
2514 extern const struct file_operations generic_ro_fops
;
2516 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2518 extern int vfs_readlink(struct dentry
*, char __user
*, int, const char *);
2519 extern int page_readlink(struct dentry
*, char __user
*, int);
2520 extern void *page_follow_link_light(struct dentry
*, struct nameidata
*);
2521 extern void page_put_link(struct dentry
*, struct nameidata
*, void *);
2522 extern int __page_symlink(struct inode
*inode
, const char *symname
, int len
,
2524 extern int page_symlink(struct inode
*inode
, const char *symname
, int len
);
2525 extern const struct inode_operations page_symlink_inode_operations
;
2526 extern void kfree_put_link(struct dentry
*, struct nameidata
*, void *);
2527 extern int generic_readlink(struct dentry
*, char __user
*, int);
2528 extern void generic_fillattr(struct inode
*, struct kstat
*);
2529 int vfs_getattr_nosec(struct path
*path
, struct kstat
*stat
);
2530 extern int vfs_getattr(struct path
*, struct kstat
*);
2531 void __inode_add_bytes(struct inode
*inode
, loff_t bytes
);
2532 void inode_add_bytes(struct inode
*inode
, loff_t bytes
);
2533 void __inode_sub_bytes(struct inode
*inode
, loff_t bytes
);
2534 void inode_sub_bytes(struct inode
*inode
, loff_t bytes
);
2535 loff_t
inode_get_bytes(struct inode
*inode
);
2536 void inode_set_bytes(struct inode
*inode
, loff_t bytes
);
2538 extern int vfs_readdir(struct file
*, filldir_t
, void *);
2539 extern int iterate_dir(struct file
*, struct dir_context
*);
2541 extern int vfs_stat(const char __user
*, struct kstat
*);
2542 extern int vfs_lstat(const char __user
*, struct kstat
*);
2543 extern int vfs_fstat(unsigned int, struct kstat
*);
2544 extern int vfs_fstatat(int , const char __user
*, struct kstat
*, int);
2546 extern int do_vfs_ioctl(struct file
*filp
, unsigned int fd
, unsigned int cmd
,
2548 extern int __generic_block_fiemap(struct inode
*inode
,
2549 struct fiemap_extent_info
*fieinfo
,
2550 loff_t start
, loff_t len
,
2551 get_block_t
*get_block
);
2552 extern int generic_block_fiemap(struct inode
*inode
,
2553 struct fiemap_extent_info
*fieinfo
, u64 start
,
2554 u64 len
, get_block_t
*get_block
);
2556 extern void get_filesystem(struct file_system_type
*fs
);
2557 extern void put_filesystem(struct file_system_type
*fs
);
2558 extern struct file_system_type
*get_fs_type(const char *name
);
2559 extern struct super_block
*get_super(struct block_device
*);
2560 extern struct super_block
*get_super_thawed(struct block_device
*);
2561 extern struct super_block
*get_active_super(struct block_device
*bdev
);
2562 extern void drop_super(struct super_block
*sb
);
2563 extern void iterate_supers(void (*)(struct super_block
*, void *), void *);
2564 extern void iterate_supers_type(struct file_system_type
*,
2565 void (*)(struct super_block
*, void *), void *);
2567 extern int dcache_dir_open(struct inode
*, struct file
*);
2568 extern int dcache_dir_close(struct inode
*, struct file
*);
2569 extern loff_t
dcache_dir_lseek(struct file
*, loff_t
, int);
2570 extern int dcache_readdir(struct file
*, struct dir_context
*);
2571 extern int simple_setattr(struct dentry
*, struct iattr
*);
2572 extern int simple_getattr(struct vfsmount
*, struct dentry
*, struct kstat
*);
2573 extern int simple_statfs(struct dentry
*, struct kstatfs
*);
2574 extern int simple_open(struct inode
*inode
, struct file
*file
);
2575 extern int simple_link(struct dentry
*, struct inode
*, struct dentry
*);
2576 extern int simple_unlink(struct inode
*, struct dentry
*);
2577 extern int simple_rmdir(struct inode
*, struct dentry
*);
2578 extern int simple_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
2579 extern int noop_fsync(struct file
*, loff_t
, loff_t
, int);
2580 extern int simple_empty(struct dentry
*);
2581 extern int simple_readpage(struct file
*file
, struct page
*page
);
2582 extern int simple_write_begin(struct file
*file
, struct address_space
*mapping
,
2583 loff_t pos
, unsigned len
, unsigned flags
,
2584 struct page
**pagep
, void **fsdata
);
2585 extern int simple_write_end(struct file
*file
, struct address_space
*mapping
,
2586 loff_t pos
, unsigned len
, unsigned copied
,
2587 struct page
*page
, void *fsdata
);
2588 extern int always_delete_dentry(const struct dentry
*);
2589 extern struct inode
*alloc_anon_inode(struct super_block
*);
2590 extern const struct dentry_operations simple_dentry_operations
;
2592 extern struct dentry
*simple_lookup(struct inode
*, struct dentry
*, unsigned int flags
);
2593 extern ssize_t
generic_read_dir(struct file
*, char __user
*, size_t, loff_t
*);
2594 extern const struct file_operations simple_dir_operations
;
2595 extern const struct inode_operations simple_dir_inode_operations
;
2596 struct tree_descr
{ char *name
; const struct file_operations
*ops
; int mode
; };
2597 struct dentry
*d_alloc_name(struct dentry
*, const char *);
2598 extern int simple_fill_super(struct super_block
*, unsigned long, struct tree_descr
*);
2599 extern int simple_pin_fs(struct file_system_type
*, struct vfsmount
**mount
, int *count
);
2600 extern void simple_release_fs(struct vfsmount
**mount
, int *count
);
2602 extern ssize_t
simple_read_from_buffer(void __user
*to
, size_t count
,
2603 loff_t
*ppos
, const void *from
, size_t available
);
2604 extern ssize_t
simple_write_to_buffer(void *to
, size_t available
, loff_t
*ppos
,
2605 const void __user
*from
, size_t count
);
2607 extern int generic_file_fsync(struct file
*, loff_t
, loff_t
, int);
2609 extern int generic_check_addressable(unsigned, u64
);
2611 #ifdef CONFIG_MIGRATION
2612 extern int buffer_migrate_page(struct address_space
*,
2613 struct page
*, struct page
*,
2616 #define buffer_migrate_page NULL
2619 extern int inode_change_ok(const struct inode
*, struct iattr
*);
2620 extern int inode_newsize_ok(const struct inode
*, loff_t offset
);
2621 extern void setattr_copy(struct inode
*inode
, const struct iattr
*attr
);
2623 extern int file_update_time(struct file
*file
);
2625 extern int generic_show_options(struct seq_file
*m
, struct dentry
*root
);
2626 extern void save_mount_options(struct super_block
*sb
, char *options
);
2627 extern void replace_mount_options(struct super_block
*sb
, char *options
);
2629 static inline ino_t
parent_ino(struct dentry
*dentry
)
2634 * Don't strictly need d_lock here? If the parent ino could change
2635 * then surely we'd have a deeper race in the caller?
2637 spin_lock(&dentry
->d_lock
);
2638 res
= dentry
->d_parent
->d_inode
->i_ino
;
2639 spin_unlock(&dentry
->d_lock
);
2643 /* Transaction based IO helpers */
2646 * An argresp is stored in an allocated page and holds the
2647 * size of the argument or response, along with its content
2649 struct simple_transaction_argresp
{
2654 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2656 char *simple_transaction_get(struct file
*file
, const char __user
*buf
,
2658 ssize_t
simple_transaction_read(struct file
*file
, char __user
*buf
,
2659 size_t size
, loff_t
*pos
);
2660 int simple_transaction_release(struct inode
*inode
, struct file
*file
);
2662 void simple_transaction_set(struct file
*file
, size_t n
);
2665 * simple attribute files
2667 * These attributes behave similar to those in sysfs:
2669 * Writing to an attribute immediately sets a value, an open file can be
2670 * written to multiple times.
2672 * Reading from an attribute creates a buffer from the value that might get
2673 * read with multiple read calls. When the attribute has been read
2674 * completely, no further read calls are possible until the file is opened
2677 * All attributes contain a text representation of a numeric value
2678 * that are accessed with the get() and set() functions.
2680 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2681 static int __fops ## _open(struct inode *inode, struct file *file) \
2683 __simple_attr_check_format(__fmt, 0ull); \
2684 return simple_attr_open(inode, file, __get, __set, __fmt); \
2686 static const struct file_operations __fops = { \
2687 .owner = THIS_MODULE, \
2688 .open = __fops ## _open, \
2689 .release = simple_attr_release, \
2690 .read = simple_attr_read, \
2691 .write = simple_attr_write, \
2692 .llseek = generic_file_llseek, \
2695 static inline __printf(1, 2)
2696 void __simple_attr_check_format(const char *fmt
, ...)
2698 /* don't do anything, just let the compiler check the arguments; */
2701 int simple_attr_open(struct inode
*inode
, struct file
*file
,
2702 int (*get
)(void *, u64
*), int (*set
)(void *, u64
),
2704 int simple_attr_release(struct inode
*inode
, struct file
*file
);
2705 ssize_t
simple_attr_read(struct file
*file
, char __user
*buf
,
2706 size_t len
, loff_t
*ppos
);
2707 ssize_t
simple_attr_write(struct file
*file
, const char __user
*buf
,
2708 size_t len
, loff_t
*ppos
);
2711 int proc_nr_files(struct ctl_table
*table
, int write
,
2712 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2713 int proc_nr_dentry(struct ctl_table
*table
, int write
,
2714 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2715 int proc_nr_inodes(struct ctl_table
*table
, int write
,
2716 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2717 int __init
get_filesystem_list(char *buf
);
2719 #define __FMODE_EXEC ((__force int) FMODE_EXEC)
2720 #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
2722 #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
2723 #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
2724 (flag & __FMODE_NONOTIFY)))
2726 static inline int is_sxid(umode_t mode
)
2728 return (mode
& S_ISUID
) || ((mode
& S_ISGID
) && (mode
& S_IXGRP
));
2731 static inline void inode_has_no_xattr(struct inode
*inode
)
2733 if (!is_sxid(inode
->i_mode
) && (inode
->i_sb
->s_flags
& MS_NOSEC
))
2734 inode
->i_flags
|= S_NOSEC
;
2737 static inline bool dir_emit(struct dir_context
*ctx
,
2738 const char *name
, int namelen
,
2739 u64 ino
, unsigned type
)
2741 return ctx
->actor(ctx
, name
, namelen
, ctx
->pos
, ino
, type
) == 0;
2743 static inline bool dir_emit_dot(struct file
*file
, struct dir_context
*ctx
)
2745 return ctx
->actor(ctx
, ".", 1, ctx
->pos
,
2746 file
->f_path
.dentry
->d_inode
->i_ino
, DT_DIR
) == 0;
2748 static inline bool dir_emit_dotdot(struct file
*file
, struct dir_context
*ctx
)
2750 return ctx
->actor(ctx
, "..", 2, ctx
->pos
,
2751 parent_ino(file
->f_path
.dentry
), DT_DIR
) == 0;
2753 static inline bool dir_emit_dots(struct file
*file
, struct dir_context
*ctx
)
2755 if (ctx
->pos
== 0) {
2756 if (!dir_emit_dot(file
, ctx
))
2760 if (ctx
->pos
== 1) {
2761 if (!dir_emit_dotdot(file
, ctx
))
2767 static inline bool dir_relax(struct inode
*inode
)
2769 mutex_unlock(&inode
->i_mutex
);
2770 mutex_lock(&inode
->i_mutex
);
2771 return !IS_DEADDIR(inode
);
2774 #endif /* _LINUX_FS_H */