5 #include <linux/linkage.h>
6 #include <linux/wait.h>
7 #include <linux/kdev_t.h>
8 #include <linux/dcache.h>
9 #include <linux/path.h>
10 #include <linux/stat.h>
11 #include <linux/cache.h>
12 #include <linux/list.h>
13 #include <linux/llist.h>
14 #include <linux/radix-tree.h>
15 #include <linux/rbtree.h>
16 #include <linux/init.h>
17 #include <linux/pid.h>
18 #include <linux/bug.h>
19 #include <linux/mutex.h>
20 #include <linux/capability.h>
21 #include <linux/semaphore.h>
22 #include <linux/fiemap.h>
23 #include <linux/rculist_bl.h>
24 #include <linux/atomic.h>
25 #include <linux/shrinker.h>
26 #include <linux/migrate_mode.h>
27 #include <linux/uidgid.h>
28 #include <linux/lockdep.h>
29 #include <linux/percpu-rwsem.h>
30 #include <linux/blk_types.h>
32 #include <asm/byteorder.h>
33 #include <uapi/linux/fs.h>
35 struct export_operations
;
41 struct pipe_inode_info
;
42 struct poll_table_struct
;
44 struct vm_area_struct
;
47 struct swap_info_struct
;
50 struct fscrypt_operations
;
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, int ret
,
71 #define MAY_EXEC 0x00000001
72 #define MAY_WRITE 0x00000002
73 #define MAY_READ 0x00000004
74 #define MAY_APPEND 0x00000008
75 #define MAY_ACCESS 0x00000010
76 #define MAY_OPEN 0x00000020
77 #define MAY_CHDIR 0x00000040
78 /* called from RCU mode, don't block */
79 #define MAY_NOT_BLOCK 0x00000080
82 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
83 * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
86 /* file is open for reading */
87 #define FMODE_READ ((__force fmode_t)0x1)
88 /* file is open for writing */
89 #define FMODE_WRITE ((__force fmode_t)0x2)
90 /* file is seekable */
91 #define FMODE_LSEEK ((__force fmode_t)0x4)
92 /* file can be accessed using pread */
93 #define FMODE_PREAD ((__force fmode_t)0x8)
94 /* file can be accessed using pwrite */
95 #define FMODE_PWRITE ((__force fmode_t)0x10)
96 /* File is opened for execution with sys_execve / sys_uselib */
97 #define FMODE_EXEC ((__force fmode_t)0x20)
98 /* File is opened with O_NDELAY (only set for block devices) */
99 #define FMODE_NDELAY ((__force fmode_t)0x40)
100 /* File is opened with O_EXCL (only set for block devices) */
101 #define FMODE_EXCL ((__force fmode_t)0x80)
102 /* File is opened using open(.., 3, ..) and is writeable only for ioctls
103 (specialy hack for floppy.c) */
104 #define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
105 /* 32bit hashes as llseek() offset (for directories) */
106 #define FMODE_32BITHASH ((__force fmode_t)0x200)
107 /* 64bit hashes as llseek() offset (for directories) */
108 #define FMODE_64BITHASH ((__force fmode_t)0x400)
111 * Don't update ctime and mtime.
113 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
114 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
116 #define FMODE_NOCMTIME ((__force fmode_t)0x800)
118 /* Expect random access pattern */
119 #define FMODE_RANDOM ((__force fmode_t)0x1000)
121 /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
122 #define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
124 /* File is opened with O_PATH; almost nothing can be done with it */
125 #define FMODE_PATH ((__force fmode_t)0x4000)
127 /* File hasn't page cache and can't be mmaped, for stackable filesystem */
128 #define FMODE_NONMAPPABLE ((__force fmode_t)0x400000)
130 /* File page don't need to be cached, for stackable filesystem's lower file */
131 #define FMODE_NONCACHEABLE ((__force fmode_t)0x800000)
133 /* File was opened by fanotify and shouldn't generate fanotify events */
134 #define FMODE_NONOTIFY ((__force fmode_t)0x1000000)
136 /* File can be read using splice */
137 #define FMODE_SPLICE_READ ((__force fmode_t)0x8000000)
138 /* File can be written using splice */
139 #define FMODE_SPLICE_WRITE ((__force fmode_t)0x10000000)
142 * Flag for rw_copy_check_uvector and compat_rw_copy_check_uvector
143 * that indicates that they should check the contents of the iovec are
144 * valid, but not check the memory that the iovec elements
147 #define CHECK_IOVEC_ONLY -1
150 * The below are the various read and write types that we support. Some of
151 * them include behavioral modifiers that send information down to the
152 * block layer and IO scheduler. Terminology:
154 * The block layer uses device plugging to defer IO a little bit, in
155 * the hope that we will see more IO very shortly. This increases
156 * coalescing of adjacent IO and thus reduces the number of IOs we
157 * have to send to the device. It also allows for better queuing,
158 * if the IO isn't mergeable. If the caller is going to be waiting
159 * for the IO, then he must ensure that the device is unplugged so
160 * that the IO is dispatched to the driver.
162 * All IO is handled async in Linux. This is fine for background
163 * writes, but for reads or writes that someone waits for completion
164 * on, we want to notify the block layer and IO scheduler so that they
165 * know about it. That allows them to make better scheduling
166 * decisions. So when the below references 'sync' and 'async', it
167 * is referencing this priority hint.
169 * With that in mind, the available types are:
171 * READ A normal read operation. Device will be plugged.
172 * READ_SYNC A synchronous read. Device is not plugged, caller can
173 * immediately wait on this read without caring about
175 * READA Used for read-ahead operations. Lower priority, and the
176 * block layer could (in theory) choose to ignore this
177 * request if it runs into resource problems.
178 * WRITE A normal async write. Device will be plugged.
179 * WRITE_SYNC Synchronous write. Identical to WRITE, but passes down
180 * the hint that someone will be waiting on this IO
181 * shortly. The write equivalent of READ_SYNC.
182 * WRITE_ODIRECT Special case write for O_DIRECT only.
183 * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
184 * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
185 * non-volatile media on completion.
186 * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
187 * by a cache flush and data is guaranteed to be on
188 * non-volatile media on completion.
191 #define RW_MASK REQ_WRITE
192 #define RWA_MASK REQ_RAHEAD
195 #define WRITE RW_MASK
196 #define READA RWA_MASK
197 #define KERNEL_READ (READ|REQ_KERNEL)
198 #define KERNEL_WRITE (WRITE|REQ_KERNEL)
200 #define READ_SYNC (READ | REQ_SYNC)
201 #define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE)
202 #define WRITE_ODIRECT (WRITE | REQ_SYNC)
203 #define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
204 #define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
205 #define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
208 * Attribute flags. These should be or-ed together to figure out what
211 #define ATTR_MODE (1 << 0)
212 #define ATTR_UID (1 << 1)
213 #define ATTR_GID (1 << 2)
214 #define ATTR_SIZE (1 << 3)
215 #define ATTR_ATIME (1 << 4)
216 #define ATTR_MTIME (1 << 5)
217 #define ATTR_CTIME (1 << 6)
218 #define ATTR_ATIME_SET (1 << 7)
219 #define ATTR_MTIME_SET (1 << 8)
220 #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
221 #define ATTR_ATTR_FLAG (1 << 10)
222 #define ATTR_KILL_SUID (1 << 11)
223 #define ATTR_KILL_SGID (1 << 12)
224 #define ATTR_FILE (1 << 13)
225 #define ATTR_KILL_PRIV (1 << 14)
226 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
227 #define ATTR_TIMES_SET (1 << 16)
230 * This is the Inode Attributes structure, used for notify_change(). It
231 * uses the above definitions as flags, to know which values have changed.
232 * Also, in this manner, a Filesystem can look at only the values it cares
233 * about. Basically, these are the attributes that the VFS layer can
234 * request to change from the FS layer.
236 * Derek Atkins <warlord@MIT.EDU> 94-10-20
239 unsigned int ia_valid
;
244 struct timespec ia_atime
;
245 struct timespec ia_mtime
;
246 struct timespec ia_ctime
;
249 * Not an attribute, but an auxiliary info for filesystems wanting to
250 * implement an ftruncate() like method. NOTE: filesystem should
251 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
253 struct file
*ia_file
;
257 * Includes for diskquotas.
259 #include <linux/quota.h>
262 * Maximum number of layers of fs stack. Needs to be limited to
263 * prevent kernel stack overflow
265 #define FILESYSTEM_MAX_STACK_DEPTH 2
268 * enum positive_aop_returns - aop return codes with specific semantics
270 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
271 * completed, that the page is still locked, and
272 * should be considered active. The VM uses this hint
273 * to return the page to the active list -- it won't
274 * be a candidate for writeback again in the near
275 * future. Other callers must be careful to unlock
276 * the page if they get this return. Returned by
279 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
280 * unlocked it and the page might have been truncated.
281 * The caller should back up to acquiring a new page and
282 * trying again. The aop will be taking reasonable
283 * precautions not to livelock. If the caller held a page
284 * reference, it should drop it before retrying. Returned
287 * address_space_operation functions return these large constants to indicate
288 * special semantics to the caller. These are much larger than the bytes in a
289 * page to allow for functions that return the number of bytes operated on in a
293 enum positive_aop_returns
{
294 AOP_WRITEPAGE_ACTIVATE
= 0x80000,
295 AOP_TRUNCATED_PAGE
= 0x80001,
298 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
299 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
300 #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
301 * helper code (eg buffer layer)
302 * to clear GFP_FS from alloc */
305 * oh the beauties of C type declarations.
308 struct address_space
;
309 struct writeback_control
;
312 const struct iovec
*iov
;
313 unsigned long nr_segs
;
318 size_t iov_iter_copy_from_user_atomic(struct page
*page
,
319 struct iov_iter
*i
, unsigned long offset
, size_t bytes
);
320 size_t iov_iter_copy_from_user(struct page
*page
,
321 struct iov_iter
*i
, unsigned long offset
, size_t bytes
);
322 void iov_iter_advance(struct iov_iter
*i
, size_t bytes
);
323 int iov_iter_fault_in_readable(struct iov_iter
*i
, size_t bytes
);
324 size_t iov_iter_single_seg_count(const struct iov_iter
*i
);
326 static inline void iov_iter_init(struct iov_iter
*i
,
327 const struct iovec
*iov
, unsigned long nr_segs
,
328 size_t count
, size_t written
)
331 i
->nr_segs
= nr_segs
;
333 i
->count
= count
+ written
;
335 iov_iter_advance(i
, written
);
338 static inline size_t iov_iter_count(struct iov_iter
*i
)
344 * "descriptor" for what we're up to with a read.
345 * This allows us to use the same read code yet
346 * have multiple different users of the data that
347 * we read from a file.
349 * The simplest case just copies the data to user
362 typedef int (*read_actor_t
)(read_descriptor_t
*, struct page
*,
363 unsigned long, unsigned long);
365 struct address_space_operations
{
366 int (*writepage
)(struct page
*page
, struct writeback_control
*wbc
);
367 int (*readpage
)(struct file
*, struct page
*);
369 /* Write back some dirty pages from this mapping. */
370 int (*writepages
)(struct address_space
*, struct writeback_control
*);
372 /* Set a page dirty. Return true if this dirtied it */
373 int (*set_page_dirty
)(struct page
*page
);
375 int (*readpages
)(struct file
*filp
, struct address_space
*mapping
,
376 struct list_head
*pages
, unsigned nr_pages
);
378 int (*write_begin
)(struct file
*, struct address_space
*mapping
,
379 loff_t pos
, unsigned len
, unsigned flags
,
380 struct page
**pagep
, void **fsdata
);
381 int (*write_end
)(struct file
*, struct address_space
*mapping
,
382 loff_t pos
, unsigned len
, unsigned copied
,
383 struct page
*page
, void *fsdata
);
385 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
386 sector_t (*bmap
)(struct address_space
*, sector_t
);
387 void (*invalidatepage
) (struct page
*, unsigned long);
388 int (*releasepage
) (struct page
*, gfp_t
);
389 void (*freepage
)(struct page
*);
390 ssize_t (*direct_IO
)(int, struct kiocb
*, const struct iovec
*iov
,
391 loff_t offset
, unsigned long nr_segs
);
392 int (*get_xip_mem
)(struct address_space
*, pgoff_t
, int,
393 void **, unsigned long *);
395 * migrate the contents of a page to the specified target. If sync
396 * is false, it must not block.
398 int (*migratepage
) (struct address_space
*,
399 struct page
*, struct page
*, enum migrate_mode
);
400 int (*launder_page
) (struct page
*);
401 int (*is_partially_uptodate
) (struct page
*, read_descriptor_t
*,
403 int (*error_remove_page
)(struct address_space
*, struct page
*);
405 /* swapfile support */
406 int (*swap_activate
)(struct swap_info_struct
*sis
, struct file
*file
,
408 void (*swap_deactivate
)(struct file
*file
);
411 extern const struct address_space_operations empty_aops
;
414 * pagecache_write_begin/pagecache_write_end must be used by general code
415 * to write into the pagecache.
417 int pagecache_write_begin(struct file
*, struct address_space
*mapping
,
418 loff_t pos
, unsigned len
, unsigned flags
,
419 struct page
**pagep
, void **fsdata
);
421 int pagecache_write_end(struct file
*, struct address_space
*mapping
,
422 loff_t pos
, unsigned len
, unsigned copied
,
423 struct page
*page
, void *fsdata
);
425 struct backing_dev_info
;
426 struct address_space
{
427 struct inode
*host
; /* owner: inode, block_device */
428 struct radix_tree_root page_tree
; /* radix tree of all pages */
429 spinlock_t tree_lock
; /* and lock protecting it */
430 atomic_t i_mmap_writable
;/* count VM_SHARED mappings */
431 struct rb_root i_mmap
; /* tree of private and shared mappings */
432 struct list_head i_mmap_nonlinear
;/*list VM_NONLINEAR mappings */
433 struct mutex i_mmap_mutex
; /* protect tree, count, list */
434 /* Protected by tree_lock together with the radix tree */
435 unsigned long nrpages
; /* number of total pages */
436 unsigned long nrshadows
; /* number of shadow entries */
437 pgoff_t writeback_index
;/* writeback starts here */
438 const struct address_space_operations
*a_ops
; /* methods */
439 unsigned long flags
; /* error bits/gfp mask */
440 struct backing_dev_info
*backing_dev_info
; /* device readahead, etc */
441 spinlock_t private_lock
; /* for use by the address_space */
442 struct list_head private_list
; /* ditto */
443 void *private_data
; /* ditto */
444 #if defined(CONFIG_MMC_DW_FMP_ECRYPT_FS) || defined(CONFIG_UFS_FMP_ECRYPT_FS)
445 unsigned char *iv
; /* iv */
446 unsigned char *key
; /* key */
447 unsigned long key_length
; /* key length */
448 char *alg
; /* algorithm */
449 pgoff_t sensitive_data_index
; /* data starts here */
450 struct crypto_hash
*hash_tfm
; /* hash transform */
451 #ifdef CONFIG_CRYPTO_FIPS
452 bool cc_enable
; /* cc flag */
458 } __attribute__((aligned(sizeof(long))));
460 * On most architectures that alignment is already the case; but
461 * must be enforced here for CRIS, to let the least significant bit
462 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
464 struct request_queue
;
466 struct block_device
{
467 dev_t bd_dev
; /* not a kdev_t - it's a search key */
469 struct inode
* bd_inode
; /* will die */
470 struct super_block
* bd_super
;
471 struct mutex bd_mutex
; /* open/close mutex */
472 struct list_head bd_inodes
;
476 bool bd_write_holder
;
478 struct list_head bd_holder_disks
;
480 struct block_device
* bd_contains
;
481 unsigned bd_block_size
;
482 struct hd_struct
* bd_part
;
483 /* number of times partitions within this device have been opened. */
484 unsigned bd_part_count
;
486 struct gendisk
* bd_disk
;
487 struct request_queue
* bd_queue
;
488 struct list_head bd_list
;
490 * Private data. You must have bd_claim'ed the block_device
491 * to use this. NOTE: bd_claim allows an owner to claim
492 * the same device multiple times, the owner must take special
493 * care to not mess up bd_private for that case.
495 unsigned long bd_private
;
497 /* The counter of freeze processes */
498 int bd_fsfreeze_count
;
499 /* Mutex for freeze */
500 struct mutex bd_fsfreeze_mutex
;
504 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
507 #define PAGECACHE_TAG_DIRTY 0
508 #define PAGECACHE_TAG_WRITEBACK 1
509 #define PAGECACHE_TAG_TOWRITE 2
511 int mapping_tagged(struct address_space
*mapping
, int tag
);
514 * Might pages of this file be mapped into userspace?
516 static inline int mapping_mapped(struct address_space
*mapping
)
518 return !RB_EMPTY_ROOT(&mapping
->i_mmap
) ||
519 !list_empty(&mapping
->i_mmap_nonlinear
);
523 * Might pages of this file have been modified in userspace?
524 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
525 * marks vma as VM_SHARED if it is shared, and the file was opened for
526 * writing i.e. vma may be mprotected writable even if now readonly.
528 * If i_mmap_writable is negative, no new writable mappings are allowed. You
529 * can only deny writable mappings, if none exists right now.
531 static inline int mapping_writably_mapped(struct address_space
*mapping
)
533 return atomic_read(&mapping
->i_mmap_writable
) > 0;
536 static inline int mapping_map_writable(struct address_space
*mapping
)
538 return atomic_inc_unless_negative(&mapping
->i_mmap_writable
) ?
542 static inline void mapping_unmap_writable(struct address_space
*mapping
)
544 atomic_dec(&mapping
->i_mmap_writable
);
547 static inline int mapping_deny_writable(struct address_space
*mapping
)
549 return atomic_dec_unless_positive(&mapping
->i_mmap_writable
) ?
553 static inline void mapping_allow_writable(struct address_space
*mapping
)
555 atomic_inc(&mapping
->i_mmap_writable
);
559 * Use sequence counter to get consistent i_size on 32-bit processors.
561 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
562 #include <linux/seqlock.h>
563 #define __NEED_I_SIZE_ORDERED
564 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
566 #define i_size_ordered_init(inode) do { } while (0)
570 #define ACL_NOT_CACHED ((void *)(-1))
572 #define IOP_FASTPERM 0x0001
573 #define IOP_LOOKUP 0x0002
574 #define IOP_NOFOLLOW 0x0004
577 * Keep mostly read-only and often accessed (especially for
578 * the RCU path lookup and 'stat' data) fields at the beginning
579 * of the 'struct inode'
583 unsigned short i_opflags
;
586 unsigned int i_flags
;
588 #ifdef CONFIG_FS_POSIX_ACL
589 struct posix_acl
*i_acl
;
590 struct posix_acl
*i_default_acl
;
593 const struct inode_operations
*i_op
;
594 struct super_block
*i_sb
;
595 struct address_space
*i_mapping
;
597 #ifdef CONFIG_SECURITY
601 /* Stat data, not accessed from path walking */
604 * Filesystems may only read i_nlink directly. They shall use the
605 * following functions for modification:
607 * (set|clear|inc|drop)_nlink
608 * inode_(inc|dec)_link_count
611 const unsigned int i_nlink
;
612 unsigned int __i_nlink
;
616 struct timespec i_atime
;
617 struct timespec i_mtime
;
618 struct timespec i_ctime
;
619 spinlock_t i_lock
; /* i_blocks, i_bytes, maybe i_size */
620 unsigned short i_bytes
;
621 unsigned int i_blkbits
;
624 #ifdef __NEED_I_SIZE_ORDERED
625 seqcount_t i_size_seqcount
;
629 unsigned long i_state
;
630 struct mutex i_mutex
;
632 unsigned long dirtied_when
; /* jiffies of first dirtying */
634 struct hlist_node i_hash
;
635 struct list_head i_wb_list
; /* backing dev IO list */
636 struct list_head i_lru
; /* inode LRU list */
637 struct list_head i_sb_list
;
639 struct hlist_head i_dentry
;
640 struct rcu_head i_rcu
;
644 atomic_t i_dio_count
;
645 atomic_t i_writecount
;
646 const struct file_operations
*i_fop
; /* former ->i_op->default_file_ops */
647 struct file_lock
*i_flock
;
648 struct address_space i_data
;
650 struct dquot
*i_dquot
[MAXQUOTAS
];
652 struct list_head i_devices
;
654 struct pipe_inode_info
*i_pipe
;
655 struct block_device
*i_bdev
;
661 #ifdef CONFIG_FSNOTIFY
662 __u32 i_fsnotify_mask
; /* all events this inode cares about */
663 struct hlist_head i_fsnotify_marks
;
667 atomic_t i_readcount
; /* struct files open RO */
670 #ifdef CONFIG_FS_ENCRYPTION
671 struct fscrypt_info
*i_crypt_info
;
674 void *i_private
; /* fs or device private pointer */
677 static inline int inode_unhashed(struct inode
*inode
)
679 return hlist_unhashed(&inode
->i_hash
);
683 * inode->i_mutex nesting subclasses for the lock validator:
685 * 0: the object of the current VFS operation
690 * The locking order between these classes is
691 * parent -> child -> normal -> xattr -> quota
693 enum inode_i_mutex_lock_class
703 * NOTE: in a 32bit arch with a preemptable kernel and
704 * an UP compile the i_size_read/write must be atomic
705 * with respect to the local cpu (unlike with preempt disabled),
706 * but they don't need to be atomic with respect to other cpus like in
707 * true SMP (so they need either to either locally disable irq around
708 * the read or for example on x86 they can be still implemented as a
709 * cmpxchg8b without the need of the lock prefix). For SMP compiles
710 * and 64bit archs it makes no difference if preempt is enabled or not.
712 static inline loff_t
i_size_read(const struct inode
*inode
)
714 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
719 seq
= read_seqcount_begin(&inode
->i_size_seqcount
);
720 i_size
= inode
->i_size
;
721 } while (read_seqcount_retry(&inode
->i_size_seqcount
, seq
));
723 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
727 i_size
= inode
->i_size
;
731 return inode
->i_size
;
736 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
737 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
738 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
740 static inline void i_size_write(struct inode
*inode
, loff_t i_size
)
742 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
744 write_seqcount_begin(&inode
->i_size_seqcount
);
745 inode
->i_size
= i_size
;
746 write_seqcount_end(&inode
->i_size_seqcount
);
748 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
750 inode
->i_size
= i_size
;
753 inode
->i_size
= i_size
;
757 /* Helper functions so that in most cases filesystems will
758 * not need to deal directly with kuid_t and kgid_t and can
759 * instead deal with the raw numeric values that are stored
762 static inline uid_t
i_uid_read(const struct inode
*inode
)
764 return from_kuid(&init_user_ns
, inode
->i_uid
);
767 static inline gid_t
i_gid_read(const struct inode
*inode
)
769 return from_kgid(&init_user_ns
, inode
->i_gid
);
772 static inline void i_uid_write(struct inode
*inode
, uid_t uid
)
774 inode
->i_uid
= make_kuid(&init_user_ns
, uid
);
777 static inline void i_gid_write(struct inode
*inode
, gid_t gid
)
779 inode
->i_gid
= make_kgid(&init_user_ns
, gid
);
782 static inline unsigned iminor(const struct inode
*inode
)
784 return MINOR(inode
->i_rdev
);
787 static inline unsigned imajor(const struct inode
*inode
)
789 return MAJOR(inode
->i_rdev
);
792 extern struct block_device
*I_BDEV(struct inode
*inode
);
795 rwlock_t lock
; /* protects pid, uid, euid fields */
796 struct pid
*pid
; /* pid or -pgrp where SIGIO should be sent */
797 enum pid_type pid_type
; /* Kind of process group SIGIO should be sent to */
798 kuid_t uid
, euid
; /* uid/euid of process setting the owner */
799 int signum
; /* posix.1b rt signal to be delivered on IO */
803 * Track a single file's readahead state
805 struct file_ra_state
{
806 pgoff_t start
; /* where readahead started */
807 unsigned int size
; /* # of readahead pages */
808 unsigned int async_size
; /* do asynchronous readahead when
809 there are only # of pages ahead */
811 unsigned int ra_pages
; /* Maximum readahead window */
812 unsigned int mmap_miss
; /* Cache miss stat for mmap accesses */
813 loff_t prev_pos
; /* Cache last read() position */
817 * Check if @index falls in the readahead windows.
819 static inline int ra_has_index(struct file_ra_state
*ra
, pgoff_t index
)
821 return (index
>= ra
->start
&&
822 index
< ra
->start
+ ra
->size
);
825 #define FILE_MNT_WRITE_TAKEN 1
826 #define FILE_MNT_WRITE_RELEASED 2
830 struct llist_node fu_llist
;
831 struct rcu_head fu_rcuhead
;
834 #define f_dentry f_path.dentry
835 struct inode
*f_inode
; /* cached value */
836 const struct file_operations
*f_op
;
839 * Protects f_ep_links, f_flags, f_pos vs i_size in lseek SEEK_CUR.
840 * Must not be taken from IRQ context.
843 atomic_long_t f_count
;
844 unsigned int f_flags
;
847 struct fown_struct f_owner
;
848 const struct cred
*f_cred
;
849 struct file_ra_state f_ra
;
852 #ifdef CONFIG_SECURITY
855 /* needed for tty driver, and maybe others */
859 /* Used by fs/eventpoll.c to link all the hooks to this file */
860 struct list_head f_ep_links
;
861 struct list_head f_tfile_llink
;
862 #endif /* #ifdef CONFIG_EPOLL */
863 struct address_space
*f_mapping
;
864 #ifdef CONFIG_DEBUG_WRITECOUNT
865 unsigned long f_mnt_write_state
;
872 /* file identifier */
873 unsigned char f_handle
[0];
876 static inline struct file
*get_file(struct file
*f
)
878 atomic_long_inc(&f
->f_count
);
881 #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
882 #define file_count(x) atomic_long_read(&(x)->f_count)
884 #ifdef CONFIG_DEBUG_WRITECOUNT
885 static inline void file_take_write(struct file
*f
)
887 WARN_ON(f
->f_mnt_write_state
!= 0);
888 f
->f_mnt_write_state
= FILE_MNT_WRITE_TAKEN
;
890 static inline void file_release_write(struct file
*f
)
892 f
->f_mnt_write_state
|= FILE_MNT_WRITE_RELEASED
;
894 static inline void file_reset_write(struct file
*f
)
896 f
->f_mnt_write_state
= 0;
898 static inline void file_check_state(struct file
*f
)
901 * At this point, either both or neither of these bits
904 WARN_ON(f
->f_mnt_write_state
== FILE_MNT_WRITE_TAKEN
);
905 WARN_ON(f
->f_mnt_write_state
== FILE_MNT_WRITE_RELEASED
);
907 static inline int file_check_writeable(struct file
*f
)
909 if (f
->f_mnt_write_state
== FILE_MNT_WRITE_TAKEN
)
911 printk(KERN_WARNING
"writeable file with no "
912 "mnt_want_write()\n");
916 #else /* !CONFIG_DEBUG_WRITECOUNT */
917 static inline void file_take_write(struct file
*filp
) {}
918 static inline void file_release_write(struct file
*filp
) {}
919 static inline void file_reset_write(struct file
*filp
) {}
920 static inline void file_check_state(struct file
*filp
) {}
921 static inline int file_check_writeable(struct file
*filp
)
925 #endif /* CONFIG_DEBUG_WRITECOUNT */
927 #define MAX_NON_LFS ((1UL<<31) - 1)
929 /* Page cache limit. The filesystems should put that into their s_maxbytes
930 limits, otherwise bad things can happen in VM. */
931 #if BITS_PER_LONG==32
932 #define MAX_LFS_FILESIZE (((loff_t)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
933 #elif BITS_PER_LONG==64
934 #define MAX_LFS_FILESIZE ((loff_t)0x7fffffffffffffffLL)
939 #define FL_ACCESS 8 /* not trying to lock, just looking */
940 #define FL_EXISTS 16 /* when unlocking, test for existence */
941 #define FL_LEASE 32 /* lease held on this file */
942 #define FL_CLOSE 64 /* unlock on close */
943 #define FL_SLEEP 128 /* A blocking lock */
944 #define FL_DOWNGRADE_PENDING 256 /* Lease is being downgraded */
945 #define FL_UNLOCK_PENDING 512 /* Lease is being broken */
946 #define FL_FILE_PVT 1024 /* lock is private to the file */
949 * Special return value from posix_lock_file() and vfs_lock_file() for
950 * asynchronous locking.
952 #define FILE_LOCK_DEFERRED 1
955 * The POSIX file lock owner is determined by
956 * the "struct files_struct" in the thread group
957 * (or NULL for no owner - BSD locks).
959 * Lockd stuffs a "host" pointer into this.
961 typedef struct files_struct
*fl_owner_t
;
963 struct file_lock_operations
{
964 void (*fl_copy_lock
)(struct file_lock
*, struct file_lock
*);
965 void (*fl_release_private
)(struct file_lock
*);
968 struct lock_manager_operations
{
969 int (*lm_compare_owner
)(struct file_lock
*, struct file_lock
*);
970 void (*lm_notify
)(struct file_lock
*); /* unblock callback */
971 int (*lm_grant
)(struct file_lock
*, struct file_lock
*, int);
972 void (*lm_break
)(struct file_lock
*);
973 int (*lm_change
)(struct file_lock
**, int);
976 struct lock_manager
{
977 struct list_head list
;
981 void locks_start_grace(struct net
*, struct lock_manager
*);
982 void locks_end_grace(struct lock_manager
*);
983 int locks_in_grace(struct net
*);
985 /* that will die - we need it for nfs_lock_info */
986 #include <linux/nfs_fs_i.h>
989 struct file_lock
*fl_next
; /* singly linked list for this inode */
990 struct list_head fl_link
; /* doubly linked list of all locks */
991 struct list_head fl_block
; /* circular list of blocked processes */
993 unsigned int fl_flags
;
994 unsigned char fl_type
;
996 struct pid
*fl_nspid
;
997 wait_queue_head_t fl_wait
;
998 struct file
*fl_file
;
1002 struct fasync_struct
* fl_fasync
; /* for lease break notifications */
1003 /* for lease breaks: */
1004 unsigned long fl_break_time
;
1005 unsigned long fl_downgrade_time
;
1007 const struct file_lock_operations
*fl_ops
; /* Callbacks for filesystems */
1008 const struct lock_manager_operations
*fl_lmops
; /* Callbacks for lockmanagers */
1010 struct nfs_lock_info nfs_fl
;
1011 struct nfs4_lock_info nfs4_fl
;
1013 struct list_head link
; /* link in AFS vnode's pending_locks list */
1014 int state
; /* state of grant or error if -ve */
1019 /* The following constant reflects the upper bound of the file/locking space */
1021 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
1022 #define OFFSET_MAX INT_LIMIT(loff_t)
1023 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
1026 #include <linux/fcntl.h>
1028 extern void send_sigio(struct fown_struct
*fown
, int fd
, int band
);
1030 #ifdef CONFIG_FILE_LOCKING
1031 extern int fcntl_getlk(struct file
*, struct flock __user
*);
1032 extern int fcntl_setlk(unsigned int, struct file
*, unsigned int,
1033 struct flock __user
*);
1035 #if BITS_PER_LONG == 32
1036 extern int fcntl_getlk64(struct file
*, struct flock64 __user
*);
1037 extern int fcntl_setlk64(unsigned int, struct file
*, unsigned int,
1038 struct flock64 __user
*);
1041 extern int fcntl_setlease(unsigned int fd
, struct file
*filp
, long arg
);
1042 extern int fcntl_getlease(struct file
*filp
);
1045 void locks_free_lock(struct file_lock
*fl
);
1046 extern void locks_init_lock(struct file_lock
*);
1047 extern struct file_lock
* locks_alloc_lock(void);
1048 extern void locks_copy_lock(struct file_lock
*, struct file_lock
*);
1049 extern void __locks_copy_lock(struct file_lock
*, const struct file_lock
*);
1050 extern void locks_remove_posix(struct file
*, fl_owner_t
);
1051 extern void locks_remove_file(struct file
*);
1052 extern void locks_release_private(struct file_lock
*);
1053 extern void posix_test_lock(struct file
*, struct file_lock
*);
1054 extern int posix_lock_file(struct file
*, struct file_lock
*, struct file_lock
*);
1055 extern int posix_lock_file_wait(struct file
*, struct file_lock
*);
1056 extern int posix_unblock_lock(struct file
*, struct file_lock
*);
1057 extern int vfs_test_lock(struct file
*, struct file_lock
*);
1058 extern int vfs_lock_file(struct file
*, unsigned int, struct file_lock
*, struct file_lock
*);
1059 extern int vfs_cancel_lock(struct file
*filp
, struct file_lock
*fl
);
1060 extern int flock_lock_file_wait(struct file
*filp
, struct file_lock
*fl
);
1061 extern int __break_lease(struct inode
*inode
, unsigned int flags
);
1062 extern void lease_get_mtime(struct inode
*, struct timespec
*time
);
1063 extern int generic_setlease(struct file
*, long, struct file_lock
**);
1064 extern int vfs_setlease(struct file
*, long, struct file_lock
**);
1065 extern int lease_modify(struct file_lock
**, int);
1066 extern int lock_may_read(struct inode
*, loff_t start
, unsigned long count
);
1067 extern int lock_may_write(struct inode
*, loff_t start
, unsigned long count
);
1068 extern void locks_delete_block(struct file_lock
*waiter
);
1069 extern void lock_flocks(void);
1070 extern void unlock_flocks(void);
1071 #else /* !CONFIG_FILE_LOCKING */
1072 static inline int fcntl_getlk(struct file
*file
, struct flock __user
*user
)
1077 static inline int fcntl_setlk(unsigned int fd
, struct file
*file
,
1078 unsigned int cmd
, struct flock __user
*user
)
1083 #if BITS_PER_LONG == 32
1084 static inline int fcntl_getlk64(struct file
*file
, struct flock64 __user
*user
)
1089 static inline int fcntl_setlk64(unsigned int fd
, struct file
*file
,
1090 unsigned int cmd
, struct flock64 __user
*user
)
1095 static inline int fcntl_setlease(unsigned int fd
, struct file
*filp
, long arg
)
1100 static inline int fcntl_getlease(struct file
*filp
)
1105 static inline void locks_init_lock(struct file_lock
*fl
)
1110 static inline void __locks_copy_lock(struct file_lock
*new, struct file_lock
*fl
)
1115 static inline void locks_copy_lock(struct file_lock
*new, struct file_lock
*fl
)
1120 static inline void locks_remove_posix(struct file
*filp
, fl_owner_t owner
)
1125 static inline void locks_remove_file(struct file
*filp
)
1130 static inline void posix_test_lock(struct file
*filp
, struct file_lock
*fl
)
1135 static inline int posix_lock_file(struct file
*filp
, struct file_lock
*fl
,
1136 struct file_lock
*conflock
)
1141 static inline int posix_lock_file_wait(struct file
*filp
, struct file_lock
*fl
)
1146 static inline int posix_unblock_lock(struct file
*filp
,
1147 struct file_lock
*waiter
)
1152 static inline int vfs_test_lock(struct file
*filp
, struct file_lock
*fl
)
1157 static inline int vfs_lock_file(struct file
*filp
, unsigned int cmd
,
1158 struct file_lock
*fl
, struct file_lock
*conf
)
1163 static inline int vfs_cancel_lock(struct file
*filp
, struct file_lock
*fl
)
1168 static inline int flock_lock_file_wait(struct file
*filp
,
1169 struct file_lock
*request
)
1174 static inline int __break_lease(struct inode
*inode
, unsigned int mode
)
1179 static inline void lease_get_mtime(struct inode
*inode
, struct timespec
*time
)
1184 static inline int generic_setlease(struct file
*filp
, long arg
,
1185 struct file_lock
**flp
)
1190 static inline int vfs_setlease(struct file
*filp
, long arg
,
1191 struct file_lock
**lease
)
1196 static inline int lease_modify(struct file_lock
**before
, int arg
)
1201 static inline int lock_may_read(struct inode
*inode
, loff_t start
,
1207 static inline int lock_may_write(struct inode
*inode
, loff_t start
,
1213 static inline void locks_delete_block(struct file_lock
*waiter
)
1217 static inline void lock_flocks(void)
1221 static inline void unlock_flocks(void)
1225 #endif /* !CONFIG_FILE_LOCKING */
1228 struct fasync_struct
{
1232 struct fasync_struct
*fa_next
; /* singly linked list */
1233 struct file
*fa_file
;
1234 struct rcu_head fa_rcu
;
1237 #define FASYNC_MAGIC 0x4601
1239 /* SMP safe fasync helpers: */
1240 extern int fasync_helper(int, struct file
*, int, struct fasync_struct
**);
1241 extern struct fasync_struct
*fasync_insert_entry(int, struct file
*, struct fasync_struct
**, struct fasync_struct
*);
1242 extern int fasync_remove_entry(struct file
*, struct fasync_struct
**);
1243 extern struct fasync_struct
*fasync_alloc(void);
1244 extern void fasync_free(struct fasync_struct
*);
1246 /* can be called from interrupts */
1247 extern void kill_fasync(struct fasync_struct
**, int, int);
1249 extern int __f_setown(struct file
*filp
, struct pid
*, enum pid_type
, int force
);
1250 extern int f_setown(struct file
*filp
, unsigned long arg
, int force
);
1251 extern void f_delown(struct file
*filp
);
1252 extern pid_t
f_getown(struct file
*filp
);
1253 extern int send_sigurg(struct fown_struct
*fown
);
1261 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1262 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1263 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1264 #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1265 #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
1267 extern struct list_head super_blocks
;
1268 extern spinlock_t sb_lock
;
1270 /* Possible states of 'frozen' field */
1272 SB_UNFROZEN
= 0, /* FS is unfrozen */
1273 SB_FREEZE_WRITE
= 1, /* Writes, dir ops, ioctls frozen */
1274 SB_FREEZE_PAGEFAULT
= 2, /* Page faults stopped as well */
1275 SB_FREEZE_FS
= 3, /* For internal FS use (e.g. to stop
1276 * internal threads if needed) */
1277 SB_FREEZE_COMPLETE
= 4, /* ->freeze_fs finished successfully */
1280 #define SB_FREEZE_LEVELS (SB_FREEZE_COMPLETE - 1)
1283 /* Counters for counting writers at each level */
1284 struct percpu_counter counter
[SB_FREEZE_LEVELS
];
1285 wait_queue_head_t wait
; /* queue for waiting for
1286 writers / faults to finish */
1287 int frozen
; /* Is sb frozen? */
1288 wait_queue_head_t wait_unfrozen
; /* queue for waiting for
1290 #ifdef CONFIG_DEBUG_LOCK_ALLOC
1291 struct lockdep_map lock_map
[SB_FREEZE_LEVELS
];
1295 struct super_block
{
1296 struct list_head s_list
; /* Keep this first */
1297 dev_t s_dev
; /* search index; _not_ kdev_t */
1298 unsigned char s_blocksize_bits
;
1299 unsigned long s_blocksize
;
1300 loff_t s_maxbytes
; /* Max file size */
1301 struct file_system_type
*s_type
;
1302 const struct super_operations
*s_op
;
1303 const struct dquot_operations
*dq_op
;
1304 const struct quotactl_ops
*s_qcop
;
1305 const struct export_operations
*s_export_op
;
1306 unsigned long s_flags
;
1307 unsigned long s_magic
;
1308 struct dentry
*s_root
;
1309 struct rw_semaphore s_umount
;
1312 #ifdef CONFIG_SECURITY
1315 const struct xattr_handler
**s_xattr
;
1317 const struct fscrypt_operations
*s_cop
;
1319 struct list_head s_inodes
; /* all inodes */
1320 struct hlist_bl_head s_anon
; /* anonymous dentries for (nfs) exporting */
1321 struct list_head s_mounts
; /* list of mounts; _not_ for fs use */
1322 /* s_dentry_lru, s_nr_dentry_unused protected by dcache.c lru locks */
1323 struct list_head s_dentry_lru
; /* unused dentry lru */
1324 int s_nr_dentry_unused
; /* # of dentry on lru */
1326 /* s_inode_lru_lock protects s_inode_lru and s_nr_inodes_unused */
1327 spinlock_t s_inode_lru_lock ____cacheline_aligned_in_smp
;
1328 struct list_head s_inode_lru
; /* unused inode lru */
1329 int s_nr_inodes_unused
; /* # of inodes on lru */
1331 struct block_device
*s_bdev
;
1332 struct backing_dev_info
*s_bdi
;
1333 struct mtd_info
*s_mtd
;
1334 struct hlist_node s_instances
;
1335 struct quota_info s_dquot
; /* Diskquota specific options */
1337 struct sb_writers s_writers
;
1339 char s_id
[32]; /* Informational name */
1340 u8 s_uuid
[16]; /* UUID */
1342 void *s_fs_info
; /* Filesystem private info */
1343 unsigned int s_max_links
;
1346 /* Granularity of c/m/atime in ns.
1347 Cannot be worse than a second */
1351 * The next field is for VFS *only*. No filesystems have any business
1352 * even looking at it. You had been warned.
1354 struct mutex s_vfs_rename_mutex
; /* Kludge */
1357 * Filesystem subtype. If non-empty the filesystem type field
1358 * in /proc/mounts will be "type.subtype"
1363 * Saved mount options for lazy filesystems using
1364 * generic_show_options()
1366 char __rcu
*s_options
;
1367 const struct dentry_operations
*s_d_op
; /* default d_op for dentries */
1370 * Saved pool identifier for cleancache (-1 means none)
1372 int cleancache_poolid
;
1374 struct shrinker s_shrink
; /* per-sb shrinker handle */
1376 /* Number of inodes with nlink == 0 but still referenced */
1377 atomic_long_t s_remove_count
;
1379 /* Being remounted read-only */
1380 int s_readonly_remount
;
1383 * Indicates how deep in a filesystem stack this SB is
1388 /* superblock cache pruning functions */
1389 extern void prune_icache_sb(struct super_block
*sb
, int nr_to_scan
);
1390 extern void prune_dcache_sb(struct super_block
*sb
, int nr_to_scan
);
1392 extern struct timespec
current_fs_time(struct super_block
*sb
);
1395 * Snapshotting support.
1398 void __sb_end_write(struct super_block
*sb
, int level
);
1399 int __sb_start_write(struct super_block
*sb
, int level
, bool wait
);
1402 * sb_end_write - drop write access to a superblock
1403 * @sb: the super we wrote to
1405 * Decrement number of writers to the filesystem. Wake up possible waiters
1406 * wanting to freeze the filesystem.
1408 static inline void sb_end_write(struct super_block
*sb
)
1410 __sb_end_write(sb
, SB_FREEZE_WRITE
);
1414 * sb_end_pagefault - drop write access to a superblock from a page fault
1415 * @sb: the super we wrote to
1417 * Decrement number of processes handling write page fault to the filesystem.
1418 * Wake up possible waiters wanting to freeze the filesystem.
1420 static inline void sb_end_pagefault(struct super_block
*sb
)
1422 __sb_end_write(sb
, SB_FREEZE_PAGEFAULT
);
1426 * sb_end_intwrite - drop write access to a superblock for internal fs purposes
1427 * @sb: the super we wrote to
1429 * Decrement fs-internal number of writers to the filesystem. Wake up possible
1430 * waiters wanting to freeze the filesystem.
1432 static inline void sb_end_intwrite(struct super_block
*sb
)
1434 __sb_end_write(sb
, SB_FREEZE_FS
);
1438 * sb_start_write - get write access to a superblock
1439 * @sb: the super we write to
1441 * When a process wants to write data or metadata to a file system (i.e. dirty
1442 * a page or an inode), it should embed the operation in a sb_start_write() -
1443 * sb_end_write() pair to get exclusion against file system freezing. This
1444 * function increments number of writers preventing freezing. If the file
1445 * system is already frozen, the function waits until the file system is
1448 * Since freeze protection behaves as a lock, users have to preserve
1449 * ordering of freeze protection and other filesystem locks. Generally,
1450 * freeze protection should be the outermost lock. In particular, we have:
1453 * -> i_mutex (write path, truncate, directory ops, ...)
1454 * -> s_umount (freeze_super, thaw_super)
1456 static inline void sb_start_write(struct super_block
*sb
)
1458 __sb_start_write(sb
, SB_FREEZE_WRITE
, true);
1461 static inline int sb_start_write_trylock(struct super_block
*sb
)
1463 return __sb_start_write(sb
, SB_FREEZE_WRITE
, false);
1467 * sb_start_pagefault - get write access to a superblock from a page fault
1468 * @sb: the super we write to
1470 * When a process starts handling write page fault, it should embed the
1471 * operation into sb_start_pagefault() - sb_end_pagefault() pair to get
1472 * exclusion against file system freezing. This is needed since the page fault
1473 * is going to dirty a page. This function increments number of running page
1474 * faults preventing freezing. If the file system is already frozen, the
1475 * function waits until the file system is thawed.
1477 * Since page fault freeze protection behaves as a lock, users have to preserve
1478 * ordering of freeze protection and other filesystem locks. It is advised to
1479 * put sb_start_pagefault() close to mmap_sem in lock ordering. Page fault
1480 * handling code implies lock dependency:
1483 * -> sb_start_pagefault
1485 static inline void sb_start_pagefault(struct super_block
*sb
)
1487 __sb_start_write(sb
, SB_FREEZE_PAGEFAULT
, true);
1491 * sb_start_intwrite - get write access to a superblock for internal fs purposes
1492 * @sb: the super we write to
1494 * This is the third level of protection against filesystem freezing. It is
1495 * free for use by a filesystem. The only requirement is that it must rank
1496 * below sb_start_pagefault.
1498 * For example filesystem can call sb_start_intwrite() when starting a
1499 * transaction which somewhat eases handling of freezing for internal sources
1500 * of filesystem changes (internal fs threads, discarding preallocation on file
1503 static inline void sb_start_intwrite(struct super_block
*sb
)
1505 __sb_start_write(sb
, SB_FREEZE_FS
, true);
1509 extern bool inode_owner_or_capable(const struct inode
*inode
);
1512 * VFS helper functions..
1514 extern int vfs_create(struct inode
*, struct dentry
*, umode_t
, bool);
1515 extern int vfs_create2(struct vfsmount
*, struct inode
*, struct dentry
*, umode_t
, bool);
1516 extern int vfs_mkdir(struct inode
*, struct dentry
*, umode_t
);
1517 extern int vfs_mkdir2(struct vfsmount
*, struct inode
*, struct dentry
*, umode_t
);
1518 extern int vfs_mknod(struct inode
*, struct dentry
*, umode_t
, dev_t
);
1519 extern int vfs_mknod2(struct vfsmount
*, struct inode
*, struct dentry
*, umode_t
, dev_t
);
1520 extern int vfs_symlink(struct inode
*, struct dentry
*, const char *);
1521 extern int vfs_link(struct dentry
*, struct inode
*, struct dentry
*);
1522 extern int vfs_link2(struct vfsmount
*, struct dentry
*, struct inode
*, struct dentry
*);
1523 extern int vfs_rmdir(struct inode
*, struct dentry
*);
1524 extern int vfs_rmdir2(struct vfsmount
*, struct inode
*, struct dentry
*);
1525 extern int vfs_unlink(struct inode
*, struct dentry
*);
1526 extern int vfs_unlink2(struct vfsmount
*, struct inode
*, struct dentry
*);
1527 extern int vfs_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
1528 extern int vfs_rename2(struct vfsmount
*, struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
1531 * VFS dentry helper functions.
1533 extern void dentry_unhash(struct dentry
*dentry
);
1536 * VFS file helper functions.
1538 extern void inode_init_owner(struct inode
*inode
, const struct inode
*dir
,
1541 * VFS FS_IOC_FIEMAP helper definitions.
1543 struct fiemap_extent_info
{
1544 unsigned int fi_flags
; /* Flags as passed from user */
1545 unsigned int fi_extents_mapped
; /* Number of mapped extents */
1546 unsigned int fi_extents_max
; /* Size of fiemap_extent array */
1547 struct fiemap_extent __user
*fi_extents_start
; /* Start of
1548 fiemap_extent array */
1550 int fiemap_fill_next_extent(struct fiemap_extent_info
*info
, u64 logical
,
1551 u64 phys
, u64 len
, u32 flags
);
1552 int fiemap_check_flags(struct fiemap_extent_info
*fieinfo
, u32 fs_flags
);
1557 * NOTE! These match bits 12..15 of stat.st_mode
1558 * (ie "(i_mode >> 12) & 15").
1560 #define DT_UNKNOWN 0
1571 * This is the "filldir" function type, used by readdir() to let
1572 * the kernel specify what kind of dirent layout it wants to have.
1573 * This allows the kernel to read directories into kernel space or
1574 * to have different dirent layouts depending on the binary type.
1576 typedef int (*filldir_t
)(void *, const char *, int, loff_t
, u64
, unsigned);
1577 struct dir_context
{
1578 const filldir_t actor
;
1583 static inline bool dir_emit(struct dir_context
*ctx
,
1584 const char *name
, int namelen
,
1585 u64 ino
, unsigned type
)
1587 return ctx
->actor(ctx
, name
, namelen
, ctx
->pos
, ino
, type
) == 0;
1589 struct block_device_operations
;
1591 /* These macros are for out of kernel modules to test that
1592 * the kernel supports the unlocked_ioctl and compat_ioctl
1593 * fields in struct file_operations. */
1594 #define HAVE_COMPAT_IOCTL 1
1595 #define HAVE_UNLOCKED_IOCTL 1
1597 struct file_operations
{
1598 struct module
*owner
;
1599 loff_t (*llseek
) (struct file
*, loff_t
, int);
1600 ssize_t (*read
) (struct file
*, char __user
*, size_t, loff_t
*);
1601 ssize_t (*write
) (struct file
*, const char __user
*, size_t, loff_t
*);
1602 ssize_t (*aio_read
) (struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1603 ssize_t (*aio_write
) (struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1604 int (*readdir
) (struct file
*, void *, filldir_t
);
1605 int (*iterate
) (struct file
*, struct dir_context
*);
1606 unsigned int (*poll
) (struct file
*, struct poll_table_struct
*);
1607 long (*unlocked_ioctl
) (struct file
*, unsigned int, unsigned long);
1608 long (*compat_ioctl
) (struct file
*, unsigned int, unsigned long);
1609 int (*mmap
) (struct file
*, struct vm_area_struct
*);
1610 int (*open
) (struct inode
*, struct file
*);
1611 int (*flush
) (struct file
*, fl_owner_t id
);
1612 int (*release
) (struct inode
*, struct file
*);
1613 int (*fsync
) (struct file
*, loff_t
, loff_t
, int datasync
);
1614 int (*aio_fsync
) (struct kiocb
*, int datasync
);
1615 int (*fasync
) (int, struct file
*, int);
1616 int (*lock
) (struct file
*, int, struct file_lock
*);
1617 ssize_t (*sendpage
) (struct file
*, struct page
*, int, size_t, loff_t
*, int);
1618 unsigned long (*get_unmapped_area
)(struct file
*, unsigned long, unsigned long, unsigned long, unsigned long);
1619 int (*check_flags
)(int);
1620 int (*flock
) (struct file
*, int, struct file_lock
*);
1621 ssize_t (*splice_write
)(struct pipe_inode_info
*, struct file
*, loff_t
*, size_t, unsigned int);
1622 ssize_t (*splice_read
)(struct file
*, loff_t
*, struct pipe_inode_info
*, size_t, unsigned int);
1623 int (*setlease
)(struct file
*, long, struct file_lock
**);
1624 long (*fallocate
)(struct file
*file
, int mode
, loff_t offset
,
1626 int (*show_fdinfo
)(struct seq_file
*m
, struct file
*f
);
1627 /* get_lower_file is for stackable file system */
1628 struct file
* (*get_lower_file
)(struct file
*f
);
1631 struct inode_operations
{
1632 struct dentry
* (*lookup
) (struct inode
*,struct dentry
*, unsigned int);
1633 void * (*follow_link
) (struct dentry
*, struct nameidata
*);
1634 int (*permission
) (struct inode
*, int);
1635 int (*permission2
) (struct vfsmount
*, struct inode
*, int);
1636 struct posix_acl
* (*get_acl
)(struct inode
*, int);
1638 int (*readlink
) (struct dentry
*, char __user
*,int);
1639 void (*put_link
) (struct dentry
*, struct nameidata
*, void *);
1641 int (*create
) (struct inode
*,struct dentry
*, umode_t
, bool);
1642 int (*link
) (struct dentry
*,struct inode
*,struct dentry
*);
1643 int (*unlink
) (struct inode
*,struct dentry
*);
1644 int (*symlink
) (struct inode
*,struct dentry
*,const char *);
1645 int (*mkdir
) (struct inode
*,struct dentry
*,umode_t
);
1646 int (*rmdir
) (struct inode
*,struct dentry
*);
1647 int (*mknod
) (struct inode
*,struct dentry
*,umode_t
,dev_t
);
1648 int (*rename
) (struct inode
*, struct dentry
*,
1649 struct inode
*, struct dentry
*);
1650 int (*setattr
) (struct dentry
*, struct iattr
*);
1651 int (*setattr2
) (struct vfsmount
*, struct dentry
*, struct iattr
*);
1652 int (*getattr
) (struct vfsmount
*mnt
, struct dentry
*, struct kstat
*);
1653 int (*setxattr
) (struct dentry
*, const char *,const void *,size_t,int);
1654 ssize_t (*getxattr
) (struct dentry
*, const char *, void *, size_t);
1655 ssize_t (*listxattr
) (struct dentry
*, char *, size_t);
1656 int (*removexattr
) (struct dentry
*, const char *);
1657 int (*fiemap
)(struct inode
*, struct fiemap_extent_info
*, u64 start
,
1659 int (*update_time
)(struct inode
*, struct timespec
*, int);
1660 int (*atomic_open
)(struct inode
*, struct dentry
*,
1661 struct file
*, unsigned open_flag
,
1662 umode_t create_mode
, int *opened
);
1663 int (*tmpfile
) (struct inode
*, struct dentry
*, umode_t
);
1664 } ____cacheline_aligned
;
1666 ssize_t
rw_copy_check_uvector(int type
, const struct iovec __user
* uvector
,
1667 unsigned long nr_segs
, unsigned long fast_segs
,
1668 struct iovec
*fast_pointer
,
1669 struct iovec
**ret_pointer
);
1671 extern ssize_t
vfs_read(struct file
*, char __user
*, size_t, loff_t
*);
1672 extern ssize_t
vfs_write(struct file
*, const char __user
*, size_t, loff_t
*);
1673 extern ssize_t
vfs_readv(struct file
*, const struct iovec __user
*,
1674 unsigned long, loff_t
*);
1675 extern ssize_t
vfs_writev(struct file
*, const struct iovec __user
*,
1676 unsigned long, loff_t
*);
1678 struct super_operations
{
1679 struct inode
*(*alloc_inode
)(struct super_block
*sb
);
1680 void (*destroy_inode
)(struct inode
*);
1682 void (*dirty_inode
) (struct inode
*, int flags
);
1683 int (*write_inode
) (struct inode
*, struct writeback_control
*wbc
);
1684 int (*drop_inode
) (struct inode
*);
1685 void (*evict_inode
) (struct inode
*);
1686 void (*put_super
) (struct super_block
*);
1687 int (*sync_fs
)(struct super_block
*sb
, int wait
);
1688 int (*freeze_fs
) (struct super_block
*);
1689 int (*unfreeze_fs
) (struct super_block
*);
1690 int (*statfs
) (struct dentry
*, struct kstatfs
*);
1691 int (*remount_fs
) (struct super_block
*, int *, char *);
1692 int (*remount_fs2
) (struct vfsmount
*, struct super_block
*, int *, char *);
1693 void *(*clone_mnt_data
) (void *);
1694 void (*copy_mnt_data
) (void *, void *);
1695 void (*umount_begin
) (struct super_block
*);
1697 int (*show_options
)(struct seq_file
*, struct dentry
*);
1698 int (*show_options2
)(struct vfsmount
*,struct seq_file
*, struct dentry
*);
1699 int (*show_devname
)(struct seq_file
*, struct dentry
*);
1700 int (*show_path
)(struct seq_file
*, struct dentry
*);
1701 int (*show_stats
)(struct seq_file
*, struct dentry
*);
1703 ssize_t (*quota_read
)(struct super_block
*, int, char *, size_t, loff_t
);
1704 ssize_t (*quota_write
)(struct super_block
*, int, const char *, size_t, loff_t
);
1706 int (*bdev_try_to_free_page
)(struct super_block
*, struct page
*, gfp_t
);
1707 int (*nr_cached_objects
)(struct super_block
*);
1708 void (*free_cached_objects
)(struct super_block
*, int);
1709 long (*unlink_callback
)(struct super_block
*, char *);
1713 * Inode flags - they have no relation to superblock flags now
1715 #define S_SYNC 1 /* Writes are synced at once */
1716 #define S_NOATIME 2 /* Do not update access times */
1717 #define S_APPEND 4 /* Append-only file */
1718 #define S_IMMUTABLE 8 /* Immutable file */
1719 #define S_DEAD 16 /* removed, but still open directory */
1720 #define S_NOQUOTA 32 /* Inode is not counted to quota */
1721 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
1722 #define S_NOCMTIME 128 /* Do not update file c/mtime */
1723 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
1724 #define S_PRIVATE 512 /* Inode is fs-internal */
1725 #define S_IMA 1024 /* Inode has an associated IMA struct */
1726 #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
1727 #define S_NOSEC 4096 /* no suid or xattr security attributes */
1730 * Note that nosuid etc flags are inode-specific: setting some file-system
1731 * flags just means all the inodes inherit those flags by default. It might be
1732 * possible to override it selectively if you really wanted to with some
1733 * ioctl() that is not currently implemented.
1735 * Exception: MS_RDONLY is always applied to the entire file system.
1737 * Unfortunately, it is possible to change a filesystems flags with it mounted
1738 * with files in use. This means that all of the inodes will not have their
1739 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
1740 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
1742 #define __IS_FLG(inode, flg) ((inode)->i_sb->s_flags & (flg))
1744 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
1745 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
1746 ((inode)->i_flags & S_SYNC))
1747 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
1748 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
1749 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
1750 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
1751 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
1753 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
1754 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
1755 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
1756 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
1758 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
1759 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
1760 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
1761 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
1762 #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
1763 #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
1764 #define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
1767 * Inode state bits. Protected by inode->i_lock
1769 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1770 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1772 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1773 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1774 * various stages of removing an inode.
1776 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1778 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1779 * fdatasync(). i_atime is the usual cause.
1780 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1781 * these changes separately from I_DIRTY_SYNC so that we
1782 * don't have to write inode on fdatasync() when only
1783 * mtime has changed in it.
1784 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1785 * I_NEW Serves as both a mutex and completion notification.
1786 * New inodes set I_NEW. If two processes both create
1787 * the same inode, one of them will release its inode and
1788 * wait for I_NEW to be released before returning.
1789 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1790 * also cause waiting on I_NEW, without I_NEW actually
1791 * being set. find_inode() uses this to prevent returning
1792 * nearly-dead inodes.
1793 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1794 * is zero. I_FREEING must be set when I_WILL_FREE is
1796 * I_FREEING Set when inode is about to be freed but still has dirty
1797 * pages or buffers attached or the inode itself is still
1799 * I_CLEAR Added by clear_inode(). In this state the inode is
1800 * clean and can be destroyed. Inode keeps I_FREEING.
1802 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1803 * prohibited for many purposes. iget() must wait for
1804 * the inode to be completely released, then create it
1805 * anew. Other functions will just ignore such inodes,
1806 * if appropriate. I_NEW is used for waiting.
1808 * I_SYNC Writeback of inode is running. The bit is set during
1809 * data writeback, and cleared with a wakeup on the bit
1810 * address once it is done. The bit is also used to pin
1811 * the inode in memory for flusher thread.
1813 * I_REFERENCED Marks the inode as recently references on the LRU list.
1815 * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit().
1817 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1819 #define I_DIRTY_SYNC (1 << 0)
1820 #define I_DIRTY_DATASYNC (1 << 1)
1821 #define I_DIRTY_PAGES (1 << 2)
1823 #define I_NEW (1 << __I_NEW)
1824 #define I_WILL_FREE (1 << 4)
1825 #define I_FREEING (1 << 5)
1826 #define I_CLEAR (1 << 6)
1828 #define I_SYNC (1 << __I_SYNC)
1829 #define I_REFERENCED (1 << 8)
1830 #define __I_DIO_WAKEUP 9
1831 #define I_DIO_WAKEUP (1 << I_DIO_WAKEUP)
1833 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1835 extern void __mark_inode_dirty(struct inode
*, int);
1836 static inline void mark_inode_dirty(struct inode
*inode
)
1838 __mark_inode_dirty(inode
, I_DIRTY
);
1841 static inline void mark_inode_dirty_sync(struct inode
*inode
)
1843 __mark_inode_dirty(inode
, I_DIRTY_SYNC
);
1846 extern void inc_nlink(struct inode
*inode
);
1847 extern void drop_nlink(struct inode
*inode
);
1848 extern void clear_nlink(struct inode
*inode
);
1849 extern void set_nlink(struct inode
*inode
, unsigned int nlink
);
1851 static inline void inode_inc_link_count(struct inode
*inode
)
1854 mark_inode_dirty(inode
);
1857 static inline void inode_dec_link_count(struct inode
*inode
)
1860 mark_inode_dirty(inode
);
1864 * inode_inc_iversion - increments i_version
1865 * @inode: inode that need to be updated
1867 * Every time the inode is modified, the i_version field will be incremented.
1868 * The filesystem has to be mounted with i_version flag
1871 static inline void inode_inc_iversion(struct inode
*inode
)
1873 spin_lock(&inode
->i_lock
);
1875 spin_unlock(&inode
->i_lock
);
1878 enum file_time_flags
{
1885 extern void touch_atime(struct path
*);
1886 static inline void file_accessed(struct file
*file
)
1888 if (!(file
->f_flags
& O_NOATIME
))
1889 touch_atime(&file
->f_path
);
1892 int sync_inode(struct inode
*inode
, struct writeback_control
*wbc
);
1893 int sync_inode_metadata(struct inode
*inode
, int wait
);
1895 struct file_system_type
{
1898 #define FS_REQUIRES_DEV 1
1899 #define FS_BINARY_MOUNTDATA 2
1900 #define FS_HAS_SUBTYPE 4
1901 #define FS_USERNS_MOUNT 8 /* Can be mounted by userns root */
1902 #define FS_USERNS_DEV_MOUNT 16 /* A userns mount does not imply MNT_NODEV */
1903 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() during rename() internally. */
1904 struct dentry
*(*mount
) (struct file_system_type
*, int,
1905 const char *, void *);
1906 struct dentry
*(*mount2
) (struct vfsmount
*, struct file_system_type
*, int,
1907 const char *, void *);
1908 void *(*alloc_mnt_data
) (void);
1909 void (*kill_sb
) (struct super_block
*);
1910 struct module
*owner
;
1911 struct file_system_type
* next
;
1912 struct hlist_head fs_supers
;
1914 struct lock_class_key s_lock_key
;
1915 struct lock_class_key s_umount_key
;
1916 struct lock_class_key s_vfs_rename_key
;
1917 struct lock_class_key s_writers_key
[SB_FREEZE_LEVELS
];
1919 struct lock_class_key i_lock_key
;
1920 struct lock_class_key i_mutex_key
;
1921 struct lock_class_key i_mutex_dir_key
;
1924 #define MODULE_ALIAS_FS(NAME) MODULE_ALIAS("fs-" NAME)
1926 extern struct dentry
*mount_ns(struct file_system_type
*fs_type
, int flags
,
1927 void *data
, int (*fill_super
)(struct super_block
*, void *, int));
1928 extern struct dentry
*mount_bdev(struct file_system_type
*fs_type
,
1929 int flags
, const char *dev_name
, void *data
,
1930 int (*fill_super
)(struct super_block
*, void *, int));
1931 extern struct dentry
*mount_single(struct file_system_type
*fs_type
,
1932 int flags
, void *data
,
1933 int (*fill_super
)(struct super_block
*, void *, int));
1934 extern struct dentry
*mount_nodev(struct file_system_type
*fs_type
,
1935 int flags
, void *data
,
1936 int (*fill_super
)(struct super_block
*, void *, int));
1937 extern struct dentry
*mount_subtree(struct vfsmount
*mnt
, const char *path
);
1938 void generic_shutdown_super(struct super_block
*sb
);
1939 void kill_block_super(struct super_block
*sb
);
1940 void kill_anon_super(struct super_block
*sb
);
1941 void kill_litter_super(struct super_block
*sb
);
1942 void deactivate_super(struct super_block
*sb
);
1943 void deactivate_locked_super(struct super_block
*sb
);
1944 int set_anon_super(struct super_block
*s
, void *data
);
1945 int get_anon_bdev(dev_t
*);
1946 void free_anon_bdev(dev_t
);
1947 struct super_block
*sget(struct file_system_type
*type
,
1948 int (*test
)(struct super_block
*,void *),
1949 int (*set
)(struct super_block
*,void *),
1950 int flags
, void *data
);
1951 extern struct dentry
*mount_pseudo(struct file_system_type
*, char *,
1952 const struct super_operations
*ops
,
1953 const struct dentry_operations
*dops
,
1956 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1957 #define fops_get(fops) \
1958 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1959 #define fops_put(fops) \
1960 do { if (fops) module_put((fops)->owner); } while(0)
1962 extern int register_filesystem(struct file_system_type
*);
1963 extern int unregister_filesystem(struct file_system_type
*);
1964 extern struct vfsmount
*kern_mount_data(struct file_system_type
*, void *data
);
1965 #define kern_mount(type) kern_mount_data(type, NULL)
1966 extern void kern_unmount(struct vfsmount
*mnt
);
1967 extern int may_umount_tree(struct vfsmount
*);
1968 extern int may_umount(struct vfsmount
*);
1969 extern long do_mount(const char *, const char *, const char *, unsigned long, void *);
1970 extern struct vfsmount
*collect_mounts(struct path
*);
1971 extern void drop_collected_mounts(struct vfsmount
*);
1972 extern int iterate_mounts(int (*)(struct vfsmount
*, void *), void *,
1974 extern int vfs_statfs(struct path
*, struct kstatfs
*);
1975 extern int user_statfs(const char __user
*, struct kstatfs
*);
1976 extern int fd_statfs(int, struct kstatfs
*);
1977 extern int vfs_ustat(dev_t
, struct kstatfs
*);
1978 extern int freeze_super(struct super_block
*super
);
1979 extern int thaw_super(struct super_block
*super
);
1980 extern bool our_mnt(struct vfsmount
*mnt
);
1982 extern int current_umask(void);
1985 extern struct kobject
*fs_kobj
;
1987 #define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
1988 extern int rw_verify_area(int, struct file
*, loff_t
*, size_t);
1990 #define FLOCK_VERIFY_READ 1
1991 #define FLOCK_VERIFY_WRITE 2
1993 #ifdef CONFIG_FILE_LOCKING
1994 extern int locks_mandatory_locked(struct inode
*);
1995 extern int locks_mandatory_area(int, struct inode
*, struct file
*, loff_t
, size_t);
1998 * Candidates for mandatory locking have the setgid bit set
1999 * but no group execute bit - an otherwise meaningless combination.
2002 static inline int __mandatory_lock(struct inode
*ino
)
2004 return (ino
->i_mode
& (S_ISGID
| S_IXGRP
)) == S_ISGID
;
2008 * ... and these candidates should be on MS_MANDLOCK mounted fs,
2009 * otherwise these will be advisory locks
2012 static inline int mandatory_lock(struct inode
*ino
)
2014 return IS_MANDLOCK(ino
) && __mandatory_lock(ino
);
2017 static inline int locks_verify_locked(struct inode
*inode
)
2019 if (mandatory_lock(inode
))
2020 return locks_mandatory_locked(inode
);
2024 static inline int locks_verify_truncate(struct inode
*inode
,
2028 if (inode
->i_flock
&& mandatory_lock(inode
))
2029 return locks_mandatory_area(
2030 FLOCK_VERIFY_WRITE
, inode
, filp
,
2031 size
< inode
->i_size
? size
: inode
->i_size
,
2032 (size
< inode
->i_size
? inode
->i_size
- size
2033 : size
- inode
->i_size
)
2038 static inline int break_lease(struct inode
*inode
, unsigned int mode
)
2041 return __break_lease(inode
, mode
);
2044 #else /* !CONFIG_FILE_LOCKING */
2045 static inline int locks_mandatory_locked(struct inode
*inode
)
2050 static inline int locks_mandatory_area(int rw
, struct inode
*inode
,
2051 struct file
*filp
, loff_t offset
,
2057 static inline int __mandatory_lock(struct inode
*inode
)
2062 static inline int mandatory_lock(struct inode
*inode
)
2067 static inline int locks_verify_locked(struct inode
*inode
)
2072 static inline int locks_verify_truncate(struct inode
*inode
, struct file
*filp
,
2078 static inline int break_lease(struct inode
*inode
, unsigned int mode
)
2083 #endif /* CONFIG_FILE_LOCKING */
2088 const char *name
; /* pointer to actual string */
2089 const __user
char *uptr
; /* original userland pointer */
2090 struct audit_names
*aname
;
2091 bool separate
; /* should "name" be freed? */
2094 extern long vfs_truncate(struct path
*, loff_t
);
2095 extern int do_truncate(struct dentry
*, loff_t start
, unsigned int time_attrs
,
2097 extern int do_truncate2(struct vfsmount
*, struct dentry
*, loff_t start
,
2098 unsigned int time_attrs
, struct file
*filp
);
2099 extern int do_fallocate(struct file
*file
, int mode
, loff_t offset
,
2101 extern long do_sys_open(int dfd
, const char __user
*filename
, int flags
,
2103 extern struct file
*file_open_name(struct filename
*, int, umode_t
);
2104 extern struct file
*filp_open(const char *, int, umode_t
);
2105 extern struct file
*file_open_root(struct dentry
*, struct vfsmount
*,
2107 extern struct file
* dentry_open(const struct path
*, int, const struct cred
*);
2108 extern int filp_close(struct file
*, fl_owner_t id
);
2110 extern struct filename
*getname(const char __user
*);
2116 extern int finish_open(struct file
*file
, struct dentry
*dentry
,
2117 int (*open
)(struct inode
*, struct file
*),
2119 extern int finish_no_open(struct file
*file
, struct dentry
*dentry
);
2123 extern int ioctl_preallocate(struct file
*filp
, void __user
*argp
);
2126 extern void __init
vfs_caches_init_early(void);
2127 extern void __init
vfs_caches_init(unsigned long);
2129 extern struct kmem_cache
*names_cachep
;
2131 extern void final_putname(struct filename
*name
);
2133 #define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL)
2134 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
2135 #ifndef CONFIG_AUDITSYSCALL
2136 #define putname(name) final_putname(name)
2138 extern void putname(struct filename
*name
);
2142 extern int register_blkdev(unsigned int, const char *);
2143 extern void unregister_blkdev(unsigned int, const char *);
2144 extern struct block_device
*bdget(dev_t
);
2145 extern struct block_device
*bdgrab(struct block_device
*bdev
);
2146 extern void bd_set_size(struct block_device
*, loff_t size
);
2147 extern void bd_forget(struct inode
*inode
);
2148 extern void bdput(struct block_device
*);
2149 extern void invalidate_bdev(struct block_device
*);
2150 extern void iterate_bdevs(void (*)(struct block_device
*, void *), void *);
2151 extern int sync_blockdev(struct block_device
*bdev
);
2152 extern void kill_bdev(struct block_device
*);
2153 extern struct super_block
*freeze_bdev(struct block_device
*);
2154 extern void emergency_thaw_all(void);
2155 extern int thaw_bdev(struct block_device
*bdev
, struct super_block
*sb
);
2156 extern int fsync_bdev(struct block_device
*);
2158 static inline void bd_forget(struct inode
*inode
) {}
2159 static inline int sync_blockdev(struct block_device
*bdev
) { return 0; }
2160 static inline void kill_bdev(struct block_device
*bdev
) {}
2161 static inline void invalidate_bdev(struct block_device
*bdev
) {}
2163 static inline struct super_block
*freeze_bdev(struct block_device
*sb
)
2168 static inline int thaw_bdev(struct block_device
*bdev
, struct super_block
*sb
)
2173 static inline void iterate_bdevs(void (*f
)(struct block_device
*, void *), void *arg
)
2177 extern int sync_filesystem(struct super_block
*);
2178 extern const struct file_operations def_blk_fops
;
2179 extern const struct file_operations def_chr_fops
;
2180 extern const struct file_operations bad_sock_fops
;
2182 extern int ioctl_by_bdev(struct block_device
*, unsigned, unsigned long);
2183 extern int blkdev_ioctl(struct block_device
*, fmode_t
, unsigned, unsigned long);
2184 extern long compat_blkdev_ioctl(struct file
*, unsigned, unsigned long);
2185 extern int blkdev_get(struct block_device
*bdev
, fmode_t mode
, void *holder
);
2186 extern struct block_device
*blkdev_get_by_path(const char *path
, fmode_t mode
,
2188 extern struct block_device
*blkdev_get_by_dev(dev_t dev
, fmode_t mode
,
2190 extern void blkdev_put(struct block_device
*bdev
, fmode_t mode
);
2192 extern int bd_link_disk_holder(struct block_device
*bdev
, struct gendisk
*disk
);
2193 extern void bd_unlink_disk_holder(struct block_device
*bdev
,
2194 struct gendisk
*disk
);
2196 static inline int bd_link_disk_holder(struct block_device
*bdev
,
2197 struct gendisk
*disk
)
2201 static inline void bd_unlink_disk_holder(struct block_device
*bdev
,
2202 struct gendisk
*disk
)
2209 #define CHRDEV_MAJOR_HASH_SIZE 255
2210 extern int alloc_chrdev_region(dev_t
*, unsigned, unsigned, const char *);
2211 extern int register_chrdev_region(dev_t
, unsigned, const char *);
2212 extern int __register_chrdev(unsigned int major
, unsigned int baseminor
,
2213 unsigned int count
, const char *name
,
2214 const struct file_operations
*fops
);
2215 extern void __unregister_chrdev(unsigned int major
, unsigned int baseminor
,
2216 unsigned int count
, const char *name
);
2217 extern void unregister_chrdev_region(dev_t
, unsigned);
2218 extern void chrdev_show(struct seq_file
*,off_t
);
2220 static inline int register_chrdev(unsigned int major
, const char *name
,
2221 const struct file_operations
*fops
)
2223 return __register_chrdev(major
, 0, 256, name
, fops
);
2226 static inline void unregister_chrdev(unsigned int major
, const char *name
)
2228 __unregister_chrdev(major
, 0, 256, name
);
2231 /* fs/block_dev.c */
2232 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
2233 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
2236 #define BLKDEV_MAJOR_HASH_SIZE 255
2237 extern const char *__bdevname(dev_t
, char *buffer
);
2238 extern const char *bdevname(struct block_device
*bdev
, char *buffer
);
2239 extern struct block_device
*lookup_bdev(const char *);
2240 extern void blkdev_show(struct seq_file
*,off_t
);
2243 #define BLKDEV_MAJOR_HASH_SIZE 0
2246 extern void init_special_inode(struct inode
*, umode_t
, dev_t
);
2248 /* Invalid inode operations -- fs/bad_inode.c */
2249 extern void make_bad_inode(struct inode
*);
2250 extern int is_bad_inode(struct inode
*);
2254 * return READ, READA, or WRITE
2256 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2259 * return data direction, READ or WRITE
2261 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
2263 extern void check_disk_size_change(struct gendisk
*disk
,
2264 struct block_device
*bdev
);
2265 extern int revalidate_disk(struct gendisk
*);
2266 extern int check_disk_change(struct block_device
*);
2267 extern int __invalidate_device(struct block_device
*, bool);
2268 extern int invalidate_partition(struct gendisk
*, int);
2270 unsigned long invalidate_mapping_pages(struct address_space
*mapping
,
2271 pgoff_t start
, pgoff_t end
);
2273 static inline void invalidate_remote_inode(struct inode
*inode
)
2275 if (S_ISREG(inode
->i_mode
) || S_ISDIR(inode
->i_mode
) ||
2276 S_ISLNK(inode
->i_mode
))
2277 invalidate_mapping_pages(inode
->i_mapping
, 0, -1);
2279 extern int invalidate_inode_pages2(struct address_space
*mapping
);
2280 extern int invalidate_inode_pages2_range(struct address_space
*mapping
,
2281 pgoff_t start
, pgoff_t end
);
2282 extern int write_inode_now(struct inode
*, int);
2283 extern int filemap_fdatawrite(struct address_space
*);
2284 extern int filemap_flush(struct address_space
*);
2285 extern int filemap_fdatawait(struct address_space
*);
2286 extern int filemap_fdatawait_range(struct address_space
*, loff_t lstart
,
2288 extern int filemap_write_and_wait(struct address_space
*mapping
);
2289 extern int filemap_write_and_wait_range(struct address_space
*mapping
,
2290 loff_t lstart
, loff_t lend
);
2291 extern int __filemap_fdatawrite_range(struct address_space
*mapping
,
2292 loff_t start
, loff_t end
, int sync_mode
);
2293 extern int filemap_fdatawrite_range(struct address_space
*mapping
,
2294 loff_t start
, loff_t end
);
2296 extern int vfs_fsync_range(struct file
*file
, loff_t start
, loff_t end
,
2298 extern int vfs_fsync(struct file
*file
, int datasync
);
2299 extern int generic_write_sync(struct file
*file
, loff_t pos
, loff_t count
);
2300 extern void emergency_sync(void);
2301 extern void emergency_remount(void);
2302 extern int intr_sync(int *);
2305 extern sector_t
bmap(struct inode
*, sector_t
);
2307 extern int notify_change(struct dentry
*, struct iattr
*);
2308 extern int notify_change2(struct vfsmount
*, struct dentry
*, struct iattr
*);
2309 extern int inode_permission(struct inode
*, int);
2310 extern int inode_permission2(struct vfsmount
*, struct inode
*, int);
2311 extern int generic_permission(struct inode
*, int);
2313 static inline bool execute_ok(struct inode
*inode
)
2315 return (inode
->i_mode
& S_IXUGO
) || S_ISDIR(inode
->i_mode
);
2318 static inline struct inode
*file_inode(struct file
*f
)
2323 static inline void file_start_write(struct file
*file
)
2325 if (!S_ISREG(file_inode(file
)->i_mode
))
2327 __sb_start_write(file_inode(file
)->i_sb
, SB_FREEZE_WRITE
, true);
2330 static inline bool file_start_write_trylock(struct file
*file
)
2332 if (!S_ISREG(file_inode(file
)->i_mode
))
2334 return __sb_start_write(file_inode(file
)->i_sb
, SB_FREEZE_WRITE
, false);
2337 static inline void file_end_write(struct file
*file
)
2339 if (!S_ISREG(file_inode(file
)->i_mode
))
2341 __sb_end_write(file_inode(file
)->i_sb
, SB_FREEZE_WRITE
);
2345 * get_write_access() gets write permission for a file.
2346 * put_write_access() releases this write permission.
2347 * This is used for regular files.
2348 * We cannot support write (and maybe mmap read-write shared) accesses and
2349 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2350 * can have the following values:
2351 * 0: no writers, no VM_DENYWRITE mappings
2352 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2353 * > 0: (i_writecount) users are writing to the file.
2355 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2356 * except for the cases where we don't hold i_writecount yet. Then we need to
2357 * use {get,deny}_write_access() - these functions check the sign and refuse
2358 * to do the change if sign is wrong.
2360 static inline int get_write_access(struct inode
*inode
)
2362 return atomic_inc_unless_negative(&inode
->i_writecount
) ? 0 : -ETXTBSY
;
2364 static inline int deny_write_access(struct file
*file
)
2366 struct inode
*inode
= file_inode(file
);
2367 return atomic_dec_unless_positive(&inode
->i_writecount
) ? 0 : -ETXTBSY
;
2369 static inline void put_write_access(struct inode
* inode
)
2371 atomic_dec(&inode
->i_writecount
);
2373 static inline void allow_write_access(struct file
*file
)
2376 atomic_inc(&file_inode(file
)->i_writecount
);
2379 static inline void i_readcount_dec(struct inode
*inode
)
2381 BUG_ON(!atomic_read(&inode
->i_readcount
));
2382 atomic_dec(&inode
->i_readcount
);
2384 static inline void i_readcount_inc(struct inode
*inode
)
2386 atomic_inc(&inode
->i_readcount
);
2389 static inline void i_readcount_dec(struct inode
*inode
)
2393 static inline void i_readcount_inc(struct inode
*inode
)
2398 extern int do_pipe_flags(int *, int);
2400 extern int kernel_read(struct file
*, loff_t
, char *, unsigned long);
2401 extern ssize_t
kernel_write(struct file
*, const char *, size_t, loff_t
);
2402 extern struct file
* open_exec(const char *);
2404 /* fs/dcache.c -- generic fs support functions */
2405 extern int is_subdir(struct dentry
*, struct dentry
*);
2406 extern int path_is_under(struct path
*, struct path
*);
2407 extern ino_t
find_inode_number(struct dentry
*, struct qstr
*);
2409 #include <linux/err.h>
2411 /* needed for stackable file system support */
2412 extern loff_t
default_llseek(struct file
*file
, loff_t offset
, int whence
);
2414 extern loff_t
vfs_llseek(struct file
*file
, loff_t offset
, int whence
);
2416 extern int inode_init_always(struct super_block
*, struct inode
*);
2417 extern void inode_init_once(struct inode
*);
2418 extern void address_space_init_once(struct address_space
*mapping
);
2419 extern void ihold(struct inode
* inode
);
2420 extern void iput(struct inode
*);
2421 extern struct inode
* igrab(struct inode
*);
2422 extern ino_t
iunique(struct super_block
*, ino_t
);
2423 extern int inode_needs_sync(struct inode
*inode
);
2424 extern int generic_delete_inode(struct inode
*inode
);
2425 static inline int generic_drop_inode(struct inode
*inode
)
2427 return !inode
->i_nlink
|| inode_unhashed(inode
);
2430 extern struct inode
*ilookup5_nowait(struct super_block
*sb
,
2431 unsigned long hashval
, int (*test
)(struct inode
*, void *),
2433 extern struct inode
*ilookup5(struct super_block
*sb
, unsigned long hashval
,
2434 int (*test
)(struct inode
*, void *), void *data
);
2435 extern struct inode
*ilookup(struct super_block
*sb
, unsigned long ino
);
2437 extern struct inode
* iget5_locked(struct super_block
*, unsigned long, int (*test
)(struct inode
*, void *), int (*set
)(struct inode
*, void *), void *);
2438 extern struct inode
* iget_locked(struct super_block
*, unsigned long);
2439 extern int insert_inode_locked4(struct inode
*, unsigned long, int (*test
)(struct inode
*, void *), void *);
2440 extern int insert_inode_locked(struct inode
*);
2441 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2442 extern void lockdep_annotate_inode_mutex_key(struct inode
*inode
);
2444 static inline void lockdep_annotate_inode_mutex_key(struct inode
*inode
) { };
2446 extern void unlock_new_inode(struct inode
*);
2447 extern unsigned int get_next_ino(void);
2449 extern void __iget(struct inode
* inode
);
2450 extern void iget_failed(struct inode
*);
2451 extern void clear_inode(struct inode
*);
2452 extern void __destroy_inode(struct inode
*);
2453 extern struct inode
*new_inode_pseudo(struct super_block
*sb
);
2454 extern struct inode
*new_inode(struct super_block
*sb
);
2455 extern void free_inode_nonrcu(struct inode
*inode
);
2456 extern int should_remove_suid(struct dentry
*);
2457 extern int file_remove_suid(struct file
*);
2459 extern void __insert_inode_hash(struct inode
*, unsigned long hashval
);
2460 static inline void insert_inode_hash(struct inode
*inode
)
2462 __insert_inode_hash(inode
, inode
->i_ino
);
2465 extern void __remove_inode_hash(struct inode
*);
2466 static inline void remove_inode_hash(struct inode
*inode
)
2468 if (!inode_unhashed(inode
))
2469 __remove_inode_hash(inode
);
2472 extern void inode_sb_list_add(struct inode
*inode
);
2475 extern void submit_bio(int, struct bio
*);
2476 extern int bdev_read_only(struct block_device
*);
2478 extern int set_blocksize(struct block_device
*, int);
2479 extern int sb_set_blocksize(struct super_block
*, int);
2480 extern int sb_min_blocksize(struct super_block
*, int);
2482 extern int generic_file_mmap(struct file
*, struct vm_area_struct
*);
2483 extern int generic_file_readonly_mmap(struct file
*, struct vm_area_struct
*);
2484 extern int generic_file_remap_pages(struct vm_area_struct
*, unsigned long addr
,
2485 unsigned long size
, pgoff_t pgoff
);
2486 extern int file_read_actor(read_descriptor_t
* desc
, struct page
*page
, unsigned long offset
, unsigned long size
);
2487 int generic_write_checks(struct file
*file
, loff_t
*pos
, size_t *count
, int isblk
);
2488 extern ssize_t
generic_file_aio_read(struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
2489 extern ssize_t
__generic_file_aio_write(struct kiocb
*, const struct iovec
*, unsigned long,
2491 extern ssize_t
generic_file_aio_write(struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
2492 extern ssize_t
generic_file_direct_write(struct kiocb
*, const struct iovec
*,
2493 unsigned long *, loff_t
, loff_t
*, size_t, size_t);
2494 extern ssize_t
generic_file_buffered_write(struct kiocb
*, const struct iovec
*,
2495 unsigned long, loff_t
, loff_t
*, size_t, ssize_t
);
2496 extern ssize_t
do_sync_read(struct file
*filp
, char __user
*buf
, size_t len
, loff_t
*ppos
);
2497 extern ssize_t
do_sync_write(struct file
*filp
, const char __user
*buf
, size_t len
, loff_t
*ppos
);
2498 extern int generic_segment_checks(const struct iovec
*iov
,
2499 unsigned long *nr_segs
, size_t *count
, int access_flags
);
2501 /* fs/block_dev.c */
2502 extern ssize_t
blkdev_aio_write(struct kiocb
*iocb
, const struct iovec
*iov
,
2503 unsigned long nr_segs
, loff_t pos
);
2504 extern int blkdev_fsync(struct file
*filp
, loff_t start
, loff_t end
,
2506 extern void block_sync_page(struct page
*page
);
2509 extern ssize_t
generic_file_splice_read(struct file
*, loff_t
*,
2510 struct pipe_inode_info
*, size_t, unsigned int);
2511 extern ssize_t
default_file_splice_read(struct file
*, loff_t
*,
2512 struct pipe_inode_info
*, size_t, unsigned int);
2513 extern ssize_t
generic_file_splice_write(struct pipe_inode_info
*,
2514 struct file
*, loff_t
*, size_t, unsigned int);
2515 extern ssize_t
generic_splice_sendpage(struct pipe_inode_info
*pipe
,
2516 struct file
*out
, loff_t
*, size_t len
, unsigned int flags
);
2519 file_ra_state_init(struct file_ra_state
*ra
, struct address_space
*mapping
);
2520 extern loff_t
noop_llseek(struct file
*file
, loff_t offset
, int whence
);
2521 extern loff_t
no_llseek(struct file
*file
, loff_t offset
, int whence
);
2522 extern loff_t
generic_file_llseek(struct file
*file
, loff_t offset
, int whence
);
2523 extern loff_t
generic_file_llseek_size(struct file
*file
, loff_t offset
,
2524 int whence
, loff_t maxsize
, loff_t eof
);
2525 extern int generic_file_open(struct inode
* inode
, struct file
* filp
);
2526 extern int nonseekable_open(struct inode
* inode
, struct file
* filp
);
2528 #ifdef CONFIG_FS_XIP
2529 extern ssize_t
xip_file_read(struct file
*filp
, char __user
*buf
, size_t len
,
2531 extern int xip_file_mmap(struct file
* file
, struct vm_area_struct
* vma
);
2532 extern ssize_t
xip_file_write(struct file
*filp
, const char __user
*buf
,
2533 size_t len
, loff_t
*ppos
);
2534 extern int xip_truncate_page(struct address_space
*mapping
, loff_t from
);
2536 static inline int xip_truncate_page(struct address_space
*mapping
, loff_t from
)
2543 typedef void (dio_submit_t
)(int rw
, struct bio
*bio
, struct inode
*inode
,
2544 loff_t file_offset
);
2547 /* need locking between buffered and direct access */
2550 /* filesystem does not support filling holes */
2551 DIO_SKIP_HOLES
= 0x02,
2554 void dio_end_io(struct bio
*bio
, int error
);
2556 ssize_t
__blockdev_direct_IO(int rw
, struct kiocb
*iocb
, struct inode
*inode
,
2557 struct block_device
*bdev
, const struct iovec
*iov
, loff_t offset
,
2558 unsigned long nr_segs
, get_block_t get_block
, dio_iodone_t end_io
,
2559 dio_submit_t submit_io
, int flags
);
2561 static inline ssize_t
blockdev_direct_IO(int rw
, struct kiocb
*iocb
,
2562 struct inode
*inode
, const struct iovec
*iov
, loff_t offset
,
2563 unsigned long nr_segs
, get_block_t get_block
)
2565 return __blockdev_direct_IO(rw
, iocb
, inode
, inode
->i_sb
->s_bdev
, iov
,
2566 offset
, nr_segs
, get_block
, NULL
, NULL
,
2567 DIO_LOCKING
| DIO_SKIP_HOLES
);
2571 void inode_dio_wait(struct inode
*inode
);
2572 void inode_dio_done(struct inode
*inode
);
2574 extern const struct file_operations generic_ro_fops
;
2576 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2578 extern int vfs_readlink(struct dentry
*, char __user
*, int, const char *);
2579 extern int vfs_follow_link(struct nameidata
*, const char *);
2580 extern int page_readlink(struct dentry
*, char __user
*, int);
2581 extern void *page_follow_link_light(struct dentry
*, struct nameidata
*);
2582 extern void page_put_link(struct dentry
*, struct nameidata
*, void *);
2583 extern int __page_symlink(struct inode
*inode
, const char *symname
, int len
,
2585 extern int page_symlink(struct inode
*inode
, const char *symname
, int len
);
2586 extern const struct inode_operations page_symlink_inode_operations
;
2587 extern int generic_readlink(struct dentry
*, char __user
*, int);
2588 extern void generic_fillattr(struct inode
*, struct kstat
*);
2589 extern int vfs_getattr(struct path
*, struct kstat
*);
2590 void __inode_add_bytes(struct inode
*inode
, loff_t bytes
);
2591 void inode_add_bytes(struct inode
*inode
, loff_t bytes
);
2592 void __inode_sub_bytes(struct inode
*inode
, loff_t bytes
);
2593 void inode_sub_bytes(struct inode
*inode
, loff_t bytes
);
2594 loff_t
inode_get_bytes(struct inode
*inode
);
2595 void inode_set_bytes(struct inode
*inode
, loff_t bytes
);
2597 extern int vfs_readdir(struct file
*, filldir_t
, void *);
2598 extern int iterate_dir(struct file
*, struct dir_context
*);
2600 extern int vfs_stat(const char __user
*, struct kstat
*);
2601 extern int vfs_lstat(const char __user
*, struct kstat
*);
2602 extern int vfs_fstat(unsigned int, struct kstat
*);
2603 extern int vfs_fstatat(int , const char __user
*, struct kstat
*, int);
2605 extern int do_vfs_ioctl(struct file
*filp
, unsigned int fd
, unsigned int cmd
,
2607 extern int __generic_block_fiemap(struct inode
*inode
,
2608 struct fiemap_extent_info
*fieinfo
,
2609 loff_t start
, loff_t len
,
2610 get_block_t
*get_block
);
2611 extern int generic_block_fiemap(struct inode
*inode
,
2612 struct fiemap_extent_info
*fieinfo
, u64 start
,
2613 u64 len
, get_block_t
*get_block
);
2615 extern void get_filesystem(struct file_system_type
*fs
);
2616 extern void put_filesystem(struct file_system_type
*fs
);
2617 extern struct file_system_type
*get_fs_type(const char *name
);
2618 extern struct super_block
*get_super(struct block_device
*);
2619 extern struct super_block
*get_super_thawed(struct block_device
*);
2620 extern struct super_block
*get_active_super(struct block_device
*bdev
);
2621 extern void drop_super(struct super_block
*sb
);
2622 extern void iterate_supers(void (*)(struct super_block
*, void *), void *);
2623 extern void iterate_supers_type(struct file_system_type
*,
2624 void (*)(struct super_block
*, void *), void *);
2626 extern int dcache_dir_open(struct inode
*, struct file
*);
2627 extern int dcache_dir_close(struct inode
*, struct file
*);
2628 extern loff_t
dcache_dir_lseek(struct file
*, loff_t
, int);
2629 extern int dcache_readdir(struct file
*, void *, filldir_t
);
2630 extern int simple_setattr(struct dentry
*, struct iattr
*);
2631 extern int simple_getattr(struct vfsmount
*, struct dentry
*, struct kstat
*);
2632 extern int simple_statfs(struct dentry
*, struct kstatfs
*);
2633 extern int simple_open(struct inode
*inode
, struct file
*file
);
2634 extern int simple_link(struct dentry
*, struct inode
*, struct dentry
*);
2635 extern int simple_unlink(struct inode
*, struct dentry
*);
2636 extern int simple_rmdir(struct inode
*, struct dentry
*);
2637 extern int simple_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
2638 extern int noop_fsync(struct file
*, loff_t
, loff_t
, int);
2639 extern int simple_empty(struct dentry
*);
2640 extern int simple_readpage(struct file
*file
, struct page
*page
);
2641 extern int simple_write_begin(struct file
*file
, struct address_space
*mapping
,
2642 loff_t pos
, unsigned len
, unsigned flags
,
2643 struct page
**pagep
, void **fsdata
);
2644 extern int simple_write_end(struct file
*file
, struct address_space
*mapping
,
2645 loff_t pos
, unsigned len
, unsigned copied
,
2646 struct page
*page
, void *fsdata
);
2648 extern struct dentry
*simple_lookup(struct inode
*, struct dentry
*, unsigned int flags
);
2649 extern ssize_t
generic_read_dir(struct file
*, char __user
*, size_t, loff_t
*);
2650 extern const struct file_operations simple_dir_operations
;
2651 extern const struct inode_operations simple_dir_inode_operations
;
2652 struct tree_descr
{ char *name
; const struct file_operations
*ops
; int mode
; };
2653 struct dentry
*d_alloc_name(struct dentry
*, const char *);
2654 extern int simple_fill_super(struct super_block
*, unsigned long, struct tree_descr
*);
2655 extern int simple_pin_fs(struct file_system_type
*, struct vfsmount
**mount
, int *count
);
2656 extern void simple_release_fs(struct vfsmount
**mount
, int *count
);
2658 extern ssize_t
simple_read_from_buffer(void __user
*to
, size_t count
,
2659 loff_t
*ppos
, const void *from
, size_t available
);
2660 extern ssize_t
simple_write_to_buffer(void *to
, size_t available
, loff_t
*ppos
,
2661 const void __user
*from
, size_t count
);
2663 extern int generic_file_fsync(struct file
*, loff_t
, loff_t
, int);
2665 extern int generic_check_addressable(unsigned, u64
);
2667 #ifdef CONFIG_MIGRATION
2668 extern int buffer_migrate_page(struct address_space
*,
2669 struct page
*, struct page
*,
2672 #define buffer_migrate_page NULL
2675 extern int inode_change_ok(const struct inode
*, struct iattr
*);
2676 extern int inode_newsize_ok(const struct inode
*, loff_t offset
);
2677 extern void setattr_copy(struct inode
*inode
, const struct iattr
*attr
);
2679 extern int file_update_time(struct file
*file
);
2681 extern int generic_show_options(struct seq_file
*m
, struct dentry
*root
);
2682 extern void save_mount_options(struct super_block
*sb
, char *options
);
2683 extern void replace_mount_options(struct super_block
*sb
, char *options
);
2685 static inline ino_t
parent_ino(struct dentry
*dentry
)
2690 * Don't strictly need d_lock here? If the parent ino could change
2691 * then surely we'd have a deeper race in the caller?
2693 spin_lock(&dentry
->d_lock
);
2694 res
= dentry
->d_parent
->d_inode
->i_ino
;
2695 spin_unlock(&dentry
->d_lock
);
2699 /* Transaction based IO helpers */
2702 * An argresp is stored in an allocated page and holds the
2703 * size of the argument or response, along with its content
2705 struct simple_transaction_argresp
{
2710 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2712 char *simple_transaction_get(struct file
*file
, const char __user
*buf
,
2714 ssize_t
simple_transaction_read(struct file
*file
, char __user
*buf
,
2715 size_t size
, loff_t
*pos
);
2716 int simple_transaction_release(struct inode
*inode
, struct file
*file
);
2718 void simple_transaction_set(struct file
*file
, size_t n
);
2721 * simple attribute files
2723 * These attributes behave similar to those in sysfs:
2725 * Writing to an attribute immediately sets a value, an open file can be
2726 * written to multiple times.
2728 * Reading from an attribute creates a buffer from the value that might get
2729 * read with multiple read calls. When the attribute has been read
2730 * completely, no further read calls are possible until the file is opened
2733 * All attributes contain a text representation of a numeric value
2734 * that are accessed with the get() and set() functions.
2736 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2737 static int __fops ## _open(struct inode *inode, struct file *file) \
2739 __simple_attr_check_format(__fmt, 0ull); \
2740 return simple_attr_open(inode, file, __get, __set, __fmt); \
2742 static const struct file_operations __fops = { \
2743 .owner = THIS_MODULE, \
2744 .open = __fops ## _open, \
2745 .release = simple_attr_release, \
2746 .read = simple_attr_read, \
2747 .write = simple_attr_write, \
2748 .llseek = generic_file_llseek, \
2751 static inline __printf(1, 2)
2752 void __simple_attr_check_format(const char *fmt
, ...)
2754 /* don't do anything, just let the compiler check the arguments; */
2757 int simple_attr_open(struct inode
*inode
, struct file
*file
,
2758 int (*get
)(void *, u64
*), int (*set
)(void *, u64
),
2760 int simple_attr_release(struct inode
*inode
, struct file
*file
);
2761 ssize_t
simple_attr_read(struct file
*file
, char __user
*buf
,
2762 size_t len
, loff_t
*ppos
);
2763 ssize_t
simple_attr_write(struct file
*file
, const char __user
*buf
,
2764 size_t len
, loff_t
*ppos
);
2767 int proc_nr_files(struct ctl_table
*table
, int write
,
2768 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2769 int proc_nr_dentry(struct ctl_table
*table
, int write
,
2770 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2771 int proc_nr_inodes(struct ctl_table
*table
, int write
,
2772 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2773 int __init
get_filesystem_list(char *buf
);
2775 #define __FMODE_EXEC ((__force int) FMODE_EXEC)
2776 #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
2778 #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
2779 #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
2780 (flag & __FMODE_NONOTIFY)))
2782 static inline int is_sxid(umode_t mode
)
2784 return (mode
& S_ISUID
) || ((mode
& S_ISGID
) && (mode
& S_IXGRP
));
2787 static inline void inode_has_no_xattr(struct inode
*inode
)
2789 if (!is_sxid(inode
->i_mode
) && (inode
->i_sb
->s_flags
& MS_NOSEC
))
2790 inode
->i_flags
|= S_NOSEC
;
2793 #endif /* _LINUX_FS_H */