5 #include <linux/linkage.h>
6 #include <linux/wait.h>
7 #include <linux/kdev_t.h>
8 #include <linux/dcache.h>
9 #include <linux/path.h>
10 #include <linux/stat.h>
11 #include <linux/cache.h>
12 #include <linux/list.h>
13 #include <linux/list_lru.h>
14 #include <linux/llist.h>
15 #include <linux/radix-tree.h>
16 #include <linux/rbtree.h>
17 #include <linux/init.h>
18 #include <linux/pid.h>
19 #include <linux/bug.h>
20 #include <linux/mutex.h>
21 #include <linux/capability.h>
22 #include <linux/semaphore.h>
23 #include <linux/fiemap.h>
24 #include <linux/rculist_bl.h>
25 #include <linux/atomic.h>
26 #include <linux/shrinker.h>
27 #include <linux/migrate_mode.h>
28 #include <linux/uidgid.h>
29 #include <linux/lockdep.h>
30 #include <linux/percpu-rwsem.h>
31 #include <linux/blk_types.h>
33 #include <asm/byteorder.h>
34 #include <uapi/linux/fs.h>
36 struct export_operations;
42 struct pipe_inode_info;
43 struct poll_table_struct;
45 struct vm_area_struct;
48 struct swap_info_struct;
50 struct workqueue_struct;
52 extern void __init inode_init(void);
53 extern void __init inode_init_early(void);
54 extern void __init files_init(unsigned long);
56 extern struct files_stat_struct files_stat;
57 extern unsigned long get_max_files(void);
58 extern int sysctl_nr_open;
59 extern struct inodes_stat_t inodes_stat;
60 extern int leases_enable, lease_break_time;
61 extern int sysctl_protected_symlinks;
62 extern int sysctl_protected_hardlinks;
65 typedef int (get_block_t)(struct inode *inode, sector_t iblock,
66 struct buffer_head *bh_result, int create);
67 typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
68 ssize_t bytes, void *private);
70 #define MAY_EXEC 0x00000001
71 #define MAY_WRITE 0x00000002
72 #define MAY_READ 0x00000004
73 #define MAY_APPEND 0x00000008
74 #define MAY_ACCESS 0x00000010
75 #define MAY_OPEN 0x00000020
76 #define MAY_CHDIR 0x00000040
77 /* called from RCU mode, don't block */
78 #define MAY_NOT_BLOCK 0x00000080
81 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
82 * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
85 /* file is open for reading */
86 #define FMODE_READ ((__force fmode_t)0x1)
87 /* file is open for writing */
88 #define FMODE_WRITE ((__force fmode_t)0x2)
89 /* file is seekable */
90 #define FMODE_LSEEK ((__force fmode_t)0x4)
91 /* file can be accessed using pread */
92 #define FMODE_PREAD ((__force fmode_t)0x8)
93 /* file can be accessed using pwrite */
94 #define FMODE_PWRITE ((__force fmode_t)0x10)
95 /* File is opened for execution with sys_execve / sys_uselib */
96 #define FMODE_EXEC ((__force fmode_t)0x20)
97 /* File is opened with O_NDELAY (only set for block devices) */
98 #define FMODE_NDELAY ((__force fmode_t)0x40)
99 /* File is opened with O_EXCL (only set for block devices) */
100 #define FMODE_EXCL ((__force fmode_t)0x80)
101 /* File is opened using open(.., 3, ..) and is writeable only for ioctls
102 (specialy hack for floppy.c) */
103 #define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
104 /* 32bit hashes as llseek() offset (for directories) */
105 #define FMODE_32BITHASH ((__force fmode_t)0x200)
106 /* 64bit hashes as llseek() offset (for directories) */
107 #define FMODE_64BITHASH ((__force fmode_t)0x400)
110 * Don't update ctime and mtime.
112 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
113 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
115 #define FMODE_NOCMTIME ((__force fmode_t)0x800)
117 /* Expect random access pattern */
118 #define FMODE_RANDOM ((__force fmode_t)0x1000)
120 /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
121 #define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
123 /* File is opened with O_PATH; almost nothing can be done with it */
124 #define FMODE_PATH ((__force fmode_t)0x4000)
126 /* File was opened by fanotify and shouldn't generate fanotify events */
127 #define FMODE_NONOTIFY ((__force fmode_t)0x1000000)
130 * Flag for rw_copy_check_uvector and compat_rw_copy_check_uvector
131 * that indicates that they should check the contents of the iovec are
132 * valid, but not check the memory that the iovec elements
135 #define CHECK_IOVEC_ONLY -1
138 * The below are the various read and write types that we support. Some of
139 * them include behavioral modifiers that send information down to the
140 * block layer and IO scheduler. Terminology:
142 * The block layer uses device plugging to defer IO a little bit, in
143 * the hope that we will see more IO very shortly. This increases
144 * coalescing of adjacent IO and thus reduces the number of IOs we
145 * have to send to the device. It also allows for better queuing,
146 * if the IO isn't mergeable. If the caller is going to be waiting
147 * for the IO, then he must ensure that the device is unplugged so
148 * that the IO is dispatched to the driver.
150 * All IO is handled async in Linux. This is fine for background
151 * writes, but for reads or writes that someone waits for completion
152 * on, we want to notify the block layer and IO scheduler so that they
153 * know about it. That allows them to make better scheduling
154 * decisions. So when the below references 'sync' and 'async', it
155 * is referencing this priority hint.
157 * With that in mind, the available types are:
159 * READ A normal read operation. Device will be plugged.
160 * READ_SYNC A synchronous read. Device is not plugged, caller can
161 * immediately wait on this read without caring about
163 * READA Used for read-ahead operations. Lower priority, and the
164 * block layer could (in theory) choose to ignore this
165 * request if it runs into resource problems.
166 * WRITE A normal async write. Device will be plugged.
167 * WRITE_SYNC Synchronous write. Identical to WRITE, but passes down
168 * the hint that someone will be waiting on this IO
169 * shortly. The write equivalent of READ_SYNC.
170 * WRITE_ODIRECT Special case write for O_DIRECT only.
171 * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
172 * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
173 * non-volatile media on completion.
174 * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
175 * by a cache flush and data is guaranteed to be on
176 * non-volatile media on completion.
179 #define RW_MASK REQ_WRITE
180 #define RWA_MASK REQ_RAHEAD
183 #define WRITE RW_MASK
184 #define READA RWA_MASK
185 #define KERNEL_READ (READ|REQ_KERNEL)
186 #define KERNEL_WRITE (WRITE|REQ_KERNEL)
188 #define READ_SYNC (READ | REQ_SYNC)
189 #define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE)
190 #define WRITE_ODIRECT (WRITE | REQ_SYNC)
191 #define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
192 #define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
193 #define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
196 * Attribute flags. These should be or-ed together to figure out what
199 #define ATTR_MODE (1 << 0)
200 #define ATTR_UID (1 << 1)
201 #define ATTR_GID (1 << 2)
202 #define ATTR_SIZE (1 << 3)
203 #define ATTR_ATIME (1 << 4)
204 #define ATTR_MTIME (1 << 5)
205 #define ATTR_CTIME (1 << 6)
206 #define ATTR_ATIME_SET (1 << 7)
207 #define ATTR_MTIME_SET (1 << 8)
208 #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
209 #define ATTR_ATTR_FLAG (1 << 10)
210 #define ATTR_KILL_SUID (1 << 11)
211 #define ATTR_KILL_SGID (1 << 12)
212 #define ATTR_FILE (1 << 13)
213 #define ATTR_KILL_PRIV (1 << 14)
214 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
215 #define ATTR_TIMES_SET (1 << 16)
218 * This is the Inode Attributes structure, used for notify_change(). It
219 * uses the above definitions as flags, to know which values have changed.
220 * Also, in this manner, a Filesystem can look at only the values it cares
221 * about. Basically, these are the attributes that the VFS layer can
222 * request to change from the FS layer.
224 * Derek Atkins <warlord@MIT.EDU> 94-10-20
227 unsigned int ia_valid;
232 struct timespec ia_atime;
233 struct timespec ia_mtime;
234 struct timespec ia_ctime;
237 * Not an attribute, but an auxiliary info for filesystems wanting to
238 * implement an ftruncate() like method. NOTE: filesystem should
239 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
241 struct file *ia_file;
245 * Includes for diskquotas.
247 #include <linux/quota.h>
250 * enum positive_aop_returns - aop return codes with specific semantics
252 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
253 * completed, that the page is still locked, and
254 * should be considered active. The VM uses this hint
255 * to return the page to the active list -- it won't
256 * be a candidate for writeback again in the near
257 * future. Other callers must be careful to unlock
258 * the page if they get this return. Returned by
261 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
262 * unlocked it and the page might have been truncated.
263 * The caller should back up to acquiring a new page and
264 * trying again. The aop will be taking reasonable
265 * precautions not to livelock. If the caller held a page
266 * reference, it should drop it before retrying. Returned
269 * address_space_operation functions return these large constants to indicate
270 * special semantics to the caller. These are much larger than the bytes in a
271 * page to allow for functions that return the number of bytes operated on in a
275 enum positive_aop_returns {
276 AOP_WRITEPAGE_ACTIVATE = 0x80000,
277 AOP_TRUNCATED_PAGE = 0x80001,
280 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
281 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
282 #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
283 * helper code (eg buffer layer)
284 * to clear GFP_FS from alloc */
287 * oh the beauties of C type declarations.
290 struct address_space;
291 struct writeback_control;
294 struct iov_iter_ops *ops;
296 unsigned long nr_segs;
301 struct iov_iter_ops {
302 size_t (*ii_copy_to_user_atomic)(struct page *, struct iov_iter *,
303 unsigned long, size_t);
304 size_t (*ii_copy_to_user)(struct page *, struct iov_iter *,
305 unsigned long, size_t, int);
306 size_t (*ii_copy_from_user_atomic)(struct page *, struct iov_iter *,
307 unsigned long, size_t);
308 size_t (*ii_copy_from_user)(struct page *, struct iov_iter *,
309 unsigned long, size_t);
310 void (*ii_advance)(struct iov_iter *, size_t);
311 int (*ii_fault_in_readable)(struct iov_iter *, size_t);
312 size_t (*ii_single_seg_count)(const struct iov_iter *);
313 int (*ii_shorten)(struct iov_iter *, size_t);
316 static inline size_t iov_iter_copy_to_user_atomic(struct page *page,
317 struct iov_iter *i, unsigned long offset, size_t bytes)
319 return i->ops->ii_copy_to_user_atomic(page, i, offset, bytes);
321 static inline size_t __iov_iter_copy_to_user(struct page *page,
322 struct iov_iter *i, unsigned long offset, size_t bytes)
324 return i->ops->ii_copy_to_user(page, i, offset, bytes, 0);
326 static inline size_t iov_iter_copy_to_user(struct page *page,
327 struct iov_iter *i, unsigned long offset, size_t bytes)
329 return i->ops->ii_copy_to_user(page, i, offset, bytes, 1);
331 static inline size_t iov_iter_copy_from_user_atomic(struct page *page,
332 struct iov_iter *i, unsigned long offset, size_t bytes)
334 return i->ops->ii_copy_from_user_atomic(page, i, offset, bytes);
336 static inline size_t iov_iter_copy_from_user(struct page *page,
337 struct iov_iter *i, unsigned long offset, size_t bytes)
339 return i->ops->ii_copy_from_user(page, i, offset, bytes);
341 static inline void iov_iter_advance(struct iov_iter *i, size_t bytes)
343 return i->ops->ii_advance(i, bytes);
345 static inline int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes)
347 return i->ops->ii_fault_in_readable(i, bytes);
349 static inline size_t iov_iter_single_seg_count(const struct iov_iter *i)
351 return i->ops->ii_single_seg_count(i);
353 static inline int iov_iter_shorten(struct iov_iter *i, size_t count)
355 return i->ops->ii_shorten(i, count);
359 extern struct iov_iter_ops ii_bvec_ops;
362 static inline void iov_iter_init_bvec(struct iov_iter *i,
363 struct bio_vec *bvec,
364 unsigned long nr_segs,
365 size_t count, size_t written)
367 i->ops = &ii_bvec_ops;
368 i->data = (unsigned long)bvec;
369 i->nr_segs = nr_segs;
371 i->count = count + written;
373 iov_iter_advance(i, written);
376 static inline int iov_iter_has_bvec(struct iov_iter *i)
378 return i->ops == &ii_bvec_ops;
381 static inline struct bio_vec *iov_iter_bvec(struct iov_iter *i)
383 BUG_ON(!iov_iter_has_bvec(i));
384 return (struct bio_vec *)i->data;
388 extern struct iov_iter_ops ii_iovec_ops;
390 static inline void iov_iter_init(struct iov_iter *i,
391 const struct iovec *iov, unsigned long nr_segs,
392 size_t count, size_t written)
394 i->ops = &ii_iovec_ops;
395 i->data = (unsigned long)iov;
396 i->nr_segs = nr_segs;
398 i->count = count + written;
400 iov_iter_advance(i, written);
403 static inline int iov_iter_has_iovec(struct iov_iter *i)
405 return i->ops == &ii_iovec_ops;
408 static inline struct iovec *iov_iter_iovec(struct iov_iter *i)
410 BUG_ON(!iov_iter_has_iovec(i));
411 return (struct iovec *)i->data;
414 static inline size_t iov_iter_count(struct iov_iter *i)
420 * "descriptor" for what we're up to with a read.
421 * This allows us to use the same read code yet
422 * have multiple different users of the data that
423 * we read from a file.
425 * The simplest case just copies the data to user
438 typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
439 unsigned long, unsigned long);
441 struct address_space_operations {
442 int (*writepage)(struct page *page, struct writeback_control *wbc);
443 int (*readpage)(struct file *, struct page *);
445 /* Write back some dirty pages from this mapping. */
446 int (*writepages)(struct address_space *, struct writeback_control *);
448 /* Set a page dirty. Return true if this dirtied it */
449 int (*set_page_dirty)(struct page *page);
451 int (*readpages)(struct file *filp, struct address_space *mapping,
452 struct list_head *pages, unsigned nr_pages);
454 int (*write_begin)(struct file *, struct address_space *mapping,
455 loff_t pos, unsigned len, unsigned flags,
456 struct page **pagep, void **fsdata);
457 int (*write_end)(struct file *, struct address_space *mapping,
458 loff_t pos, unsigned len, unsigned copied,
459 struct page *page, void *fsdata);
461 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
462 sector_t (*bmap)(struct address_space *, sector_t);
463 void (*invalidatepage) (struct page *, unsigned int, unsigned int);
464 int (*releasepage) (struct page *, gfp_t);
465 void (*freepage)(struct page *);
466 ssize_t (*direct_IO)(int, struct kiocb *, struct iov_iter *iter,
468 int (*get_xip_mem)(struct address_space *, pgoff_t, int,
469 void **, unsigned long *);
471 * migrate the contents of a page to the specified target. If
472 * migrate_mode is MIGRATE_ASYNC, it must not block.
474 int (*migratepage) (struct address_space *,
475 struct page *, struct page *, enum migrate_mode);
476 int (*launder_page) (struct page *);
477 int (*is_partially_uptodate) (struct page *, read_descriptor_t *,
479 void (*is_dirty_writeback) (struct page *, bool *, bool *);
480 int (*error_remove_page)(struct address_space *, struct page *);
482 /* swapfile support */
483 int (*swap_activate)(struct swap_info_struct *sis, struct file *file,
485 void (*swap_deactivate)(struct file *file);
488 extern const struct address_space_operations empty_aops;
491 * pagecache_write_begin/pagecache_write_end must be used by general code
492 * to write into the pagecache.
494 int pagecache_write_begin(struct file *, struct address_space *mapping,
495 loff_t pos, unsigned len, unsigned flags,
496 struct page **pagep, void **fsdata);
498 int pagecache_write_end(struct file *, struct address_space *mapping,
499 loff_t pos, unsigned len, unsigned copied,
500 struct page *page, void *fsdata);
502 struct backing_dev_info;
503 struct address_space {
504 struct inode *host; /* owner: inode, block_device */
505 struct radix_tree_root page_tree; /* radix tree of all pages */
506 spinlock_t tree_lock; /* and lock protecting it */
507 unsigned int i_mmap_writable;/* count VM_SHARED mappings */
508 struct rb_root i_mmap; /* tree of private and shared mappings */
509 struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
510 struct mutex i_mmap_mutex; /* protect tree, count, list */
511 /* Protected by tree_lock together with the radix tree */
512 unsigned long nrpages; /* number of total pages */
513 pgoff_t writeback_index;/* writeback starts here */
514 const struct address_space_operations *a_ops; /* methods */
515 unsigned long flags; /* error bits/gfp mask */
516 struct backing_dev_info *backing_dev_info; /* device readahead, etc */
517 spinlock_t private_lock; /* for use by the address_space */
518 struct list_head private_list; /* ditto */
519 void *private_data; /* ditto */
520 } __attribute__((aligned(sizeof(long))));
522 * On most architectures that alignment is already the case; but
523 * must be enforced here for CRIS, to let the least significant bit
524 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
526 struct request_queue;
528 struct block_device {
529 dev_t bd_dev; /* not a kdev_t - it's a search key */
531 struct inode * bd_inode; /* will die */
532 struct super_block * bd_super;
533 struct mutex bd_mutex; /* open/close mutex */
534 struct list_head bd_inodes;
538 bool bd_write_holder;
540 struct list_head bd_holder_disks;
542 struct block_device * bd_contains;
543 unsigned bd_block_size;
544 struct hd_struct * bd_part;
545 /* number of times partitions within this device have been opened. */
546 unsigned bd_part_count;
548 struct gendisk * bd_disk;
549 struct request_queue * bd_queue;
550 struct list_head bd_list;
552 * Private data. You must have bd_claim'ed the block_device
553 * to use this. NOTE: bd_claim allows an owner to claim
554 * the same device multiple times, the owner must take special
555 * care to not mess up bd_private for that case.
557 unsigned long bd_private;
559 /* The counter of freeze processes */
560 int bd_fsfreeze_count;
561 /* Mutex for freeze */
562 struct mutex bd_fsfreeze_mutex;
566 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
569 #define PAGECACHE_TAG_DIRTY 0
570 #define PAGECACHE_TAG_WRITEBACK 1
571 #define PAGECACHE_TAG_TOWRITE 2
573 int mapping_tagged(struct address_space *mapping, int tag);
576 * Might pages of this file be mapped into userspace?
578 static inline int mapping_mapped(struct address_space *mapping)
580 return !RB_EMPTY_ROOT(&mapping->i_mmap) ||
581 !list_empty(&mapping->i_mmap_nonlinear);
585 * Might pages of this file have been modified in userspace?
586 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
587 * marks vma as VM_SHARED if it is shared, and the file was opened for
588 * writing i.e. vma may be mprotected writable even if now readonly.
590 static inline int mapping_writably_mapped(struct address_space *mapping)
592 return mapping->i_mmap_writable != 0;
596 * Use sequence counter to get consistent i_size on 32-bit processors.
598 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
599 #include <linux/seqlock.h>
600 #define __NEED_I_SIZE_ORDERED
601 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
603 #define i_size_ordered_init(inode) do { } while (0)
607 #define ACL_NOT_CACHED ((void *)(-1))
609 #define IOP_FASTPERM 0x0001
610 #define IOP_LOOKUP 0x0002
611 #define IOP_NOFOLLOW 0x0004
614 * Keep mostly read-only and often accessed (especially for
615 * the RCU path lookup and 'stat' data) fields at the beginning
616 * of the 'struct inode'
620 unsigned short i_opflags;
623 unsigned int i_flags;
625 #ifdef CONFIG_FS_POSIX_ACL
626 struct posix_acl *i_acl;
627 struct posix_acl *i_default_acl;
630 const struct inode_operations *i_op;
631 struct super_block *i_sb;
632 struct address_space *i_mapping;
634 #ifdef CONFIG_SECURITY
638 /* Stat data, not accessed from path walking */
641 * Filesystems may only read i_nlink directly. They shall use the
642 * following functions for modification:
644 * (set|clear|inc|drop)_nlink
645 * inode_(inc|dec)_link_count
648 const unsigned int i_nlink;
649 unsigned int __i_nlink;
653 struct timespec i_atime;
654 struct timespec i_mtime;
655 struct timespec i_ctime;
656 spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
657 unsigned short i_bytes;
658 unsigned int i_blkbits;
661 #ifdef __NEED_I_SIZE_ORDERED
662 seqcount_t i_size_seqcount;
666 unsigned long i_state;
667 struct mutex i_mutex;
669 unsigned long dirtied_when; /* jiffies of first dirtying */
671 struct hlist_node i_hash;
672 struct list_head i_wb_list; /* backing dev IO list */
673 struct list_head i_lru; /* inode LRU list */
674 struct list_head i_sb_list;
676 struct hlist_head i_dentry;
677 struct rcu_head i_rcu;
681 atomic_t i_dio_count;
682 atomic_t i_writecount;
683 const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
684 struct file_lock *i_flock;
685 struct address_space i_data;
687 struct dquot *i_dquot[MAXQUOTAS];
689 struct list_head i_devices;
691 struct pipe_inode_info *i_pipe;
692 struct block_device *i_bdev;
698 #ifdef CONFIG_FSNOTIFY
699 __u32 i_fsnotify_mask; /* all events this inode cares about */
700 struct hlist_head i_fsnotify_marks;
704 atomic_t i_readcount; /* struct files open RO */
706 void *i_private; /* fs or device private pointer */
709 static inline int inode_unhashed(struct inode *inode)
711 return hlist_unhashed(&inode->i_hash);
715 * inode->i_mutex nesting subclasses for the lock validator:
717 * 0: the object of the current VFS operation
722 * The locking order between these classes is
723 * parent -> child -> normal -> xattr -> quota
725 enum inode_i_mutex_lock_class
735 * NOTE: in a 32bit arch with a preemptable kernel and
736 * an UP compile the i_size_read/write must be atomic
737 * with respect to the local cpu (unlike with preempt disabled),
738 * but they don't need to be atomic with respect to other cpus like in
739 * true SMP (so they need either to either locally disable irq around
740 * the read or for example on x86 they can be still implemented as a
741 * cmpxchg8b without the need of the lock prefix). For SMP compiles
742 * and 64bit archs it makes no difference if preempt is enabled or not.
744 static inline loff_t i_size_read(const struct inode *inode)
746 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
751 seq = read_seqcount_begin(&inode->i_size_seqcount);
752 i_size = inode->i_size;
753 } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
755 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
759 i_size = inode->i_size;
763 return inode->i_size;
768 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
769 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
770 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
772 static inline void i_size_write(struct inode *inode, loff_t i_size)
774 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
776 write_seqcount_begin(&inode->i_size_seqcount);
777 inode->i_size = i_size;
778 write_seqcount_end(&inode->i_size_seqcount);
780 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
782 inode->i_size = i_size;
785 inode->i_size = i_size;
789 /* Helper functions so that in most cases filesystems will
790 * not need to deal directly with kuid_t and kgid_t and can
791 * instead deal with the raw numeric values that are stored
794 static inline uid_t i_uid_read(const struct inode *inode)
796 return from_kuid(&init_user_ns, inode->i_uid);
799 static inline gid_t i_gid_read(const struct inode *inode)
801 return from_kgid(&init_user_ns, inode->i_gid);
804 static inline void i_uid_write(struct inode *inode, uid_t uid)
806 inode->i_uid = make_kuid(&init_user_ns, uid);
809 static inline void i_gid_write(struct inode *inode, gid_t gid)
811 inode->i_gid = make_kgid(&init_user_ns, gid);
814 static inline unsigned iminor(const struct inode *inode)
816 return MINOR(inode->i_rdev);
819 static inline unsigned imajor(const struct inode *inode)
821 return MAJOR(inode->i_rdev);
824 extern struct block_device *I_BDEV(struct inode *inode);
827 rwlock_t lock; /* protects pid, uid, euid fields */
828 struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
829 enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
830 kuid_t uid, euid; /* uid/euid of process setting the owner */
831 int signum; /* posix.1b rt signal to be delivered on IO */
835 * Track a single file's readahead state
837 struct file_ra_state {
838 pgoff_t start; /* where readahead started */
839 unsigned int size; /* # of readahead pages */
840 unsigned int async_size; /* do asynchronous readahead when
841 there are only # of pages ahead */
843 unsigned int ra_pages; /* Maximum readahead window */
844 unsigned int mmap_miss; /* Cache miss stat for mmap accesses */
845 loff_t prev_pos; /* Cache last read() position */
849 * Check if @index falls in the readahead windows.
851 static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
853 return (index >= ra->start &&
854 index < ra->start + ra->size);
857 #define FILE_MNT_WRITE_TAKEN 1
858 #define FILE_MNT_WRITE_RELEASED 2
862 * fu_list becomes invalid after file_free is called and queued via
863 * fu_rcuhead for RCU freeing
866 struct list_head fu_list;
867 struct llist_node fu_llist;
868 struct rcu_head fu_rcuhead;
871 #define f_dentry f_path.dentry
872 struct inode *f_inode; /* cached value */
873 const struct file_operations *f_op;
876 * Protects f_ep_links, f_flags, f_pos vs i_size in lseek SEEK_CUR.
877 * Must not be taken from IRQ context.
883 atomic_long_t f_count;
884 unsigned int f_flags;
887 struct fown_struct f_owner;
888 const struct cred *f_cred;
889 struct file_ra_state f_ra;
892 #ifdef CONFIG_SECURITY
895 /* needed for tty driver, and maybe others */
899 /* Used by fs/eventpoll.c to link all the hooks to this file */
900 struct list_head f_ep_links;
901 struct list_head f_tfile_llink;
902 #endif /* #ifdef CONFIG_EPOLL */
903 struct address_space *f_mapping;
904 #ifdef CONFIG_DEBUG_WRITECOUNT
905 unsigned long f_mnt_write_state;
912 /* file identifier */
913 unsigned char f_handle[0];
916 static inline struct file *get_file(struct file *f)
918 atomic_long_inc(&f->f_count);
921 #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
922 #define file_count(x) atomic_long_read(&(x)->f_count)
924 #ifdef CONFIG_DEBUG_WRITECOUNT
925 static inline void file_take_write(struct file *f)
927 WARN_ON(f->f_mnt_write_state != 0);
928 f->f_mnt_write_state = FILE_MNT_WRITE_TAKEN;
930 static inline void file_release_write(struct file *f)
932 f->f_mnt_write_state |= FILE_MNT_WRITE_RELEASED;
934 static inline void file_reset_write(struct file *f)
936 f->f_mnt_write_state = 0;
938 static inline void file_check_state(struct file *f)
941 * At this point, either both or neither of these bits
944 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN);
945 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_RELEASED);
947 static inline int file_check_writeable(struct file *f)
949 if (f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN)
951 printk(KERN_WARNING "writeable file with no "
952 "mnt_want_write()\n");
956 #else /* !CONFIG_DEBUG_WRITECOUNT */
957 static inline void file_take_write(struct file *filp) {}
958 static inline void file_release_write(struct file *filp) {}
959 static inline void file_reset_write(struct file *filp) {}
960 static inline void file_check_state(struct file *filp) {}
961 static inline int file_check_writeable(struct file *filp)
965 #endif /* CONFIG_DEBUG_WRITECOUNT */
967 #define MAX_NON_LFS ((1UL<<31) - 1)
969 /* Page cache limit. The filesystems should put that into their s_maxbytes
970 limits, otherwise bad things can happen in VM. */
971 #if BITS_PER_LONG==32
972 #define MAX_LFS_FILESIZE (((loff_t)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
973 #elif BITS_PER_LONG==64
974 #define MAX_LFS_FILESIZE ((loff_t)0x7fffffffffffffffLL)
979 #define FL_ACCESS 8 /* not trying to lock, just looking */
980 #define FL_EXISTS 16 /* when unlocking, test for existence */
981 #define FL_LEASE 32 /* lease held on this file */
982 #define FL_CLOSE 64 /* unlock on close */
983 #define FL_SLEEP 128 /* A blocking lock */
984 #define FL_DOWNGRADE_PENDING 256 /* Lease is being downgraded */
985 #define FL_UNLOCK_PENDING 512 /* Lease is being broken */
988 * Special return value from posix_lock_file() and vfs_lock_file() for
989 * asynchronous locking.
991 #define FILE_LOCK_DEFERRED 1
994 * The POSIX file lock owner is determined by
995 * the "struct files_struct" in the thread group
996 * (or NULL for no owner - BSD locks).
998 * Lockd stuffs a "host" pointer into this.
1000 typedef struct files_struct *fl_owner_t;
1002 struct file_lock_operations {
1003 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
1004 void (*fl_release_private)(struct file_lock *);
1007 struct lock_manager_operations {
1008 int (*lm_compare_owner)(struct file_lock *, struct file_lock *);
1009 unsigned long (*lm_owner_key)(struct file_lock *);
1010 void (*lm_notify)(struct file_lock *); /* unblock callback */
1011 int (*lm_grant)(struct file_lock *, struct file_lock *, int);
1012 void (*lm_break)(struct file_lock *);
1013 int (*lm_change)(struct file_lock **, int);
1016 struct lock_manager {
1017 struct list_head list;
1021 void locks_start_grace(struct net *, struct lock_manager *);
1022 void locks_end_grace(struct lock_manager *);
1023 int locks_in_grace(struct net *);
1025 /* that will die - we need it for nfs_lock_info */
1026 #include <linux/nfs_fs_i.h>
1029 * struct file_lock represents a generic "file lock". It's used to represent
1030 * POSIX byte range locks, BSD (flock) locks, and leases. It's important to
1031 * note that the same struct is used to represent both a request for a lock and
1032 * the lock itself, but the same object is never used for both.
1034 * FIXME: should we create a separate "struct lock_request" to help distinguish
1037 * The i_flock list is ordered by:
1039 * 1) lock type -- FL_LEASEs first, then FL_FLOCK, and finally FL_POSIX
1041 * 3) lock range start
1044 * Obviously, the last two criteria only matter for POSIX locks.
1047 struct file_lock *fl_next; /* singly linked list for this inode */
1048 struct hlist_node fl_link; /* node in global lists */
1049 struct list_head fl_block; /* circular list of blocked processes */
1050 fl_owner_t fl_owner;
1051 unsigned int fl_flags;
1052 unsigned char fl_type;
1053 unsigned int fl_pid;
1054 int fl_link_cpu; /* what cpu's list is this on? */
1055 struct pid *fl_nspid;
1056 wait_queue_head_t fl_wait;
1057 struct file *fl_file;
1061 struct fasync_struct * fl_fasync; /* for lease break notifications */
1062 /* for lease breaks: */
1063 unsigned long fl_break_time;
1064 unsigned long fl_downgrade_time;
1066 const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
1067 const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
1069 struct nfs_lock_info nfs_fl;
1070 struct nfs4_lock_info nfs4_fl;
1072 struct list_head link; /* link in AFS vnode's pending_locks list */
1073 int state; /* state of grant or error if -ve */
1078 /* The following constant reflects the upper bound of the file/locking space */
1080 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
1081 #define OFFSET_MAX INT_LIMIT(loff_t)
1082 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
1085 #include <linux/fcntl.h>
1087 extern void send_sigio(struct fown_struct *fown, int fd, int band);
1089 #ifdef CONFIG_FILE_LOCKING
1090 extern int fcntl_getlk(struct file *, struct flock __user *);
1091 extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
1092 struct flock __user *);
1094 #if BITS_PER_LONG == 32
1095 extern int fcntl_getlk64(struct file *, struct flock64 __user *);
1096 extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
1097 struct flock64 __user *);
1100 extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
1101 extern int fcntl_getlease(struct file *filp);
1104 void locks_free_lock(struct file_lock *fl);
1105 extern void locks_init_lock(struct file_lock *);
1106 extern struct file_lock * locks_alloc_lock(void);
1107 extern void locks_copy_lock(struct file_lock *, struct file_lock *);
1108 extern void __locks_copy_lock(struct file_lock *, const struct file_lock *);
1109 extern void locks_remove_posix(struct file *, fl_owner_t);
1110 extern void locks_remove_flock(struct file *);
1111 extern void locks_release_private(struct file_lock *);
1112 extern void posix_test_lock(struct file *, struct file_lock *);
1113 extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
1114 extern int posix_lock_file_wait(struct file *, struct file_lock *);
1115 extern int posix_unblock_lock(struct file_lock *);
1116 extern int vfs_test_lock(struct file *, struct file_lock *);
1117 extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
1118 extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
1119 extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
1120 extern int __break_lease(struct inode *inode, unsigned int flags);
1121 extern void lease_get_mtime(struct inode *, struct timespec *time);
1122 extern int generic_setlease(struct file *, long, struct file_lock **);
1123 extern int vfs_setlease(struct file *, long, struct file_lock **);
1124 extern int lease_modify(struct file_lock **, int);
1125 extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
1126 extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
1127 #else /* !CONFIG_FILE_LOCKING */
1128 static inline int fcntl_getlk(struct file *file, struct flock __user *user)
1133 static inline int fcntl_setlk(unsigned int fd, struct file *file,
1134 unsigned int cmd, struct flock __user *user)
1139 #if BITS_PER_LONG == 32
1140 static inline int fcntl_getlk64(struct file *file, struct flock64 __user *user)
1145 static inline int fcntl_setlk64(unsigned int fd, struct file *file,
1146 unsigned int cmd, struct flock64 __user *user)
1151 static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
1156 static inline int fcntl_getlease(struct file *filp)
1161 static inline void locks_init_lock(struct file_lock *fl)
1166 static inline void __locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1171 static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1176 static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
1181 static inline void locks_remove_flock(struct file *filp)
1186 static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
1191 static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
1192 struct file_lock *conflock)
1197 static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
1202 static inline int posix_unblock_lock(struct file_lock *waiter)
1207 static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
1212 static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
1213 struct file_lock *fl, struct file_lock *conf)
1218 static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
1223 static inline int flock_lock_file_wait(struct file *filp,
1224 struct file_lock *request)
1229 static inline int __break_lease(struct inode *inode, unsigned int mode)
1234 static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
1239 static inline int generic_setlease(struct file *filp, long arg,
1240 struct file_lock **flp)
1245 static inline int vfs_setlease(struct file *filp, long arg,
1246 struct file_lock **lease)
1251 static inline int lease_modify(struct file_lock **before, int arg)
1256 static inline int lock_may_read(struct inode *inode, loff_t start,
1262 static inline int lock_may_write(struct inode *inode, loff_t start,
1267 #endif /* !CONFIG_FILE_LOCKING */
1270 struct fasync_struct {
1274 struct fasync_struct *fa_next; /* singly linked list */
1275 struct file *fa_file;
1276 struct rcu_head fa_rcu;
1279 #define FASYNC_MAGIC 0x4601
1281 /* SMP safe fasync helpers: */
1282 extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
1283 extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
1284 extern int fasync_remove_entry(struct file *, struct fasync_struct **);
1285 extern struct fasync_struct *fasync_alloc(void);
1286 extern void fasync_free(struct fasync_struct *);
1288 /* can be called from interrupts */
1289 extern void kill_fasync(struct fasync_struct **, int, int);
1291 extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
1292 extern int f_setown(struct file *filp, unsigned long arg, int force);
1293 extern void f_delown(struct file *filp);
1294 extern pid_t f_getown(struct file *filp);
1295 extern int send_sigurg(struct fown_struct *fown);
1303 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1304 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1305 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1306 #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1307 #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
1309 extern struct list_head super_blocks;
1310 extern spinlock_t sb_lock;
1312 /* Possible states of 'frozen' field */
1314 SB_UNFROZEN = 0, /* FS is unfrozen */
1315 SB_FREEZE_WRITE = 1, /* Writes, dir ops, ioctls frozen */
1316 SB_FREEZE_PAGEFAULT = 2, /* Page faults stopped as well */
1317 SB_FREEZE_FS = 3, /* For internal FS use (e.g. to stop
1318 * internal threads if needed) */
1319 SB_FREEZE_COMPLETE = 4, /* ->freeze_fs finished successfully */
1322 #define SB_FREEZE_LEVELS (SB_FREEZE_COMPLETE - 1)
1325 /* Counters for counting writers at each level */
1326 struct percpu_counter counter[SB_FREEZE_LEVELS];
1327 wait_queue_head_t wait; /* queue for waiting for
1328 writers / faults to finish */
1329 int frozen; /* Is sb frozen? */
1330 wait_queue_head_t wait_unfrozen; /* queue for waiting for
1332 #ifdef CONFIG_DEBUG_LOCK_ALLOC
1333 struct lockdep_map lock_map[SB_FREEZE_LEVELS];
1337 struct super_block {
1338 struct list_head s_list; /* Keep this first */
1339 dev_t s_dev; /* search index; _not_ kdev_t */
1340 unsigned char s_blocksize_bits;
1341 unsigned long s_blocksize;
1342 loff_t s_maxbytes; /* Max file size */
1343 struct file_system_type *s_type;
1344 const struct super_operations *s_op;
1345 const struct dquot_operations *dq_op;
1346 const struct quotactl_ops *s_qcop;
1347 const struct export_operations *s_export_op;
1348 unsigned long s_flags;
1349 unsigned long s_magic;
1350 struct dentry *s_root;
1351 struct rw_semaphore s_umount;
1354 #ifdef CONFIG_SECURITY
1357 const struct xattr_handler **s_xattr;
1359 struct list_head s_inodes; /* all inodes */
1360 struct hlist_bl_head s_anon; /* anonymous dentries for (nfs) exporting */
1362 struct list_head __percpu *s_files;
1364 struct list_head s_files;
1366 struct list_head s_mounts; /* list of mounts; _not_ for fs use */
1367 struct block_device *s_bdev;
1368 struct backing_dev_info *s_bdi;
1369 struct mtd_info *s_mtd;
1370 struct hlist_node s_instances;
1371 struct quota_info s_dquot; /* Diskquota specific options */
1373 struct sb_writers s_writers;
1375 char s_id[32]; /* Informational name */
1376 u8 s_uuid[16]; /* UUID */
1378 void *s_fs_info; /* Filesystem private info */
1379 unsigned int s_max_links;
1382 /* Granularity of c/m/atime in ns.
1383 Cannot be worse than a second */
1387 * The next field is for VFS *only*. No filesystems have any business
1388 * even looking at it. You had been warned.
1390 struct mutex s_vfs_rename_mutex; /* Kludge */
1393 * Filesystem subtype. If non-empty the filesystem type field
1394 * in /proc/mounts will be "type.subtype"
1399 * Saved mount options for lazy filesystems using
1400 * generic_show_options()
1402 char __rcu *s_options;
1403 const struct dentry_operations *s_d_op; /* default d_op for dentries */
1406 * Saved pool identifier for cleancache (-1 means none)
1408 int cleancache_poolid;
1410 struct shrinker s_shrink; /* per-sb shrinker handle */
1412 /* Number of inodes with nlink == 0 but still referenced */
1413 atomic_long_t s_remove_count;
1415 /* Being remounted read-only */
1416 int s_readonly_remount;
1418 /* AIO completions deferred from interrupt context */
1419 struct workqueue_struct *s_dio_done_wq;
1422 * Keep the lru lists last in the structure so they always sit on their
1423 * own individual cachelines.
1425 struct list_lru s_dentry_lru ____cacheline_aligned_in_smp;
1426 struct list_lru s_inode_lru ____cacheline_aligned_in_smp;
1429 extern struct timespec current_fs_time(struct super_block *sb);
1432 * Snapshotting support.
1435 void __sb_end_write(struct super_block *sb, int level);
1436 int __sb_start_write(struct super_block *sb, int level, bool wait);
1439 * sb_end_write - drop write access to a superblock
1440 * @sb: the super we wrote to
1442 * Decrement number of writers to the filesystem. Wake up possible waiters
1443 * wanting to freeze the filesystem.
1445 static inline void sb_end_write(struct super_block *sb)
1447 __sb_end_write(sb, SB_FREEZE_WRITE);
1451 * sb_end_pagefault - drop write access to a superblock from a page fault
1452 * @sb: the super we wrote to
1454 * Decrement number of processes handling write page fault to the filesystem.
1455 * Wake up possible waiters wanting to freeze the filesystem.
1457 static inline void sb_end_pagefault(struct super_block *sb)
1459 __sb_end_write(sb, SB_FREEZE_PAGEFAULT);
1463 * sb_end_intwrite - drop write access to a superblock for internal fs purposes
1464 * @sb: the super we wrote to
1466 * Decrement fs-internal number of writers to the filesystem. Wake up possible
1467 * waiters wanting to freeze the filesystem.
1469 static inline void sb_end_intwrite(struct super_block *sb)
1471 __sb_end_write(sb, SB_FREEZE_FS);
1475 * sb_start_write - get write access to a superblock
1476 * @sb: the super we write to
1478 * When a process wants to write data or metadata to a file system (i.e. dirty
1479 * a page or an inode), it should embed the operation in a sb_start_write() -
1480 * sb_end_write() pair to get exclusion against file system freezing. This
1481 * function increments number of writers preventing freezing. If the file
1482 * system is already frozen, the function waits until the file system is
1485 * Since freeze protection behaves as a lock, users have to preserve
1486 * ordering of freeze protection and other filesystem locks. Generally,
1487 * freeze protection should be the outermost lock. In particular, we have:
1490 * -> i_mutex (write path, truncate, directory ops, ...)
1491 * -> s_umount (freeze_super, thaw_super)
1493 static inline void sb_start_write(struct super_block *sb)
1495 __sb_start_write(sb, SB_FREEZE_WRITE, true);
1498 static inline int sb_start_write_trylock(struct super_block *sb)
1500 return __sb_start_write(sb, SB_FREEZE_WRITE, false);
1504 * sb_start_pagefault - get write access to a superblock from a page fault
1505 * @sb: the super we write to
1507 * When a process starts handling write page fault, it should embed the
1508 * operation into sb_start_pagefault() - sb_end_pagefault() pair to get
1509 * exclusion against file system freezing. This is needed since the page fault
1510 * is going to dirty a page. This function increments number of running page
1511 * faults preventing freezing. If the file system is already frozen, the
1512 * function waits until the file system is thawed.
1514 * Since page fault freeze protection behaves as a lock, users have to preserve
1515 * ordering of freeze protection and other filesystem locks. It is advised to
1516 * put sb_start_pagefault() close to mmap_sem in lock ordering. Page fault
1517 * handling code implies lock dependency:
1520 * -> sb_start_pagefault
1522 static inline void sb_start_pagefault(struct super_block *sb)
1524 __sb_start_write(sb, SB_FREEZE_PAGEFAULT, true);
1528 * sb_start_intwrite - get write access to a superblock for internal fs purposes
1529 * @sb: the super we write to
1531 * This is the third level of protection against filesystem freezing. It is
1532 * free for use by a filesystem. The only requirement is that it must rank
1533 * below sb_start_pagefault.
1535 * For example filesystem can call sb_start_intwrite() when starting a
1536 * transaction which somewhat eases handling of freezing for internal sources
1537 * of filesystem changes (internal fs threads, discarding preallocation on file
1540 static inline void sb_start_intwrite(struct super_block *sb)
1542 __sb_start_write(sb, SB_FREEZE_FS, true);
1546 extern bool inode_owner_or_capable(const struct inode *inode);
1549 * VFS helper functions..
1551 extern int vfs_create(struct inode *, struct dentry *, umode_t, bool);
1552 extern int vfs_mkdir(struct inode *, struct dentry *, umode_t);
1553 extern int vfs_mknod(struct inode *, struct dentry *, umode_t, dev_t);
1554 extern int vfs_symlink(struct inode *, struct dentry *, const char *);
1555 extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
1556 extern int vfs_rmdir(struct inode *, struct dentry *);
1557 extern int vfs_unlink(struct inode *, struct dentry *);
1558 extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
1561 * VFS dentry helper functions.
1563 extern void dentry_unhash(struct dentry *dentry);
1566 * VFS file helper functions.
1568 extern void inode_init_owner(struct inode *inode, const struct inode *dir,
1571 * VFS FS_IOC_FIEMAP helper definitions.
1573 struct fiemap_extent_info {
1574 unsigned int fi_flags; /* Flags as passed from user */
1575 unsigned int fi_extents_mapped; /* Number of mapped extents */
1576 unsigned int fi_extents_max; /* Size of fiemap_extent array */
1577 struct fiemap_extent __user *fi_extents_start; /* Start of
1578 fiemap_extent array */
1580 int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
1581 u64 phys, u64 len, u32 flags);
1582 int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
1587 * NOTE! These match bits 12..15 of stat.st_mode
1588 * (ie "(i_mode >> 12) & 15").
1590 #define DT_UNKNOWN 0
1601 * This is the "filldir" function type, used by readdir() to let
1602 * the kernel specify what kind of dirent layout it wants to have.
1603 * This allows the kernel to read directories into kernel space or
1604 * to have different dirent layouts depending on the binary type.
1606 typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned);
1607 struct dir_context {
1608 const filldir_t actor;
1612 struct block_device_operations;
1614 /* These macros are for out of kernel modules to test that
1615 * the kernel supports the unlocked_ioctl and compat_ioctl
1616 * fields in struct file_operations. */
1617 #define HAVE_COMPAT_IOCTL 1
1618 #define HAVE_UNLOCKED_IOCTL 1
1620 struct file_operations {
1621 struct module *owner;
1622 loff_t (*llseek) (struct file *, loff_t, int);
1623 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1624 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
1625 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1626 ssize_t (*read_iter) (struct kiocb *, struct iov_iter *, loff_t);
1627 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1628 ssize_t (*write_iter) (struct kiocb *, struct iov_iter *, loff_t);
1629 int (*iterate) (struct file *, struct dir_context *);
1630 unsigned int (*poll) (struct file *, struct poll_table_struct *);
1631 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1632 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1633 int (*mmap) (struct file *, struct vm_area_struct *);
1634 int (*open) (struct inode *, struct file *);
1635 int (*flush) (struct file *, fl_owner_t id);
1636 int (*release) (struct inode *, struct file *);
1637 int (*fsync) (struct file *, loff_t, loff_t, int datasync);
1638 int (*aio_fsync) (struct kiocb *, int datasync);
1639 int (*fasync) (int, struct file *, int);
1640 int (*lock) (struct file *, int, struct file_lock *);
1641 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1642 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1643 int (*check_flags)(int);
1644 int (*flock) (struct file *, int, struct file_lock *);
1645 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1646 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
1647 int (*setlease)(struct file *, long, struct file_lock **);
1648 long (*fallocate)(struct file *file, int mode, loff_t offset,
1650 int (*show_fdinfo)(struct seq_file *m, struct file *f);
1653 struct inode_operations {
1654 struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int);
1655 void * (*follow_link) (struct dentry *, struct nameidata *);
1656 int (*permission) (struct inode *, int);
1657 struct posix_acl * (*get_acl)(struct inode *, int);
1659 int (*readlink) (struct dentry *, char __user *,int);
1660 void (*put_link) (struct dentry *, struct nameidata *, void *);
1662 int (*create) (struct inode *,struct dentry *, umode_t, bool);
1663 int (*link) (struct dentry *,struct inode *,struct dentry *);
1664 int (*unlink) (struct inode *,struct dentry *);
1665 int (*symlink) (struct inode *,struct dentry *,const char *);
1666 int (*mkdir) (struct inode *,struct dentry *,umode_t);
1667 int (*rmdir) (struct inode *,struct dentry *);
1668 int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t);
1669 int (*rename) (struct inode *, struct dentry *,
1670 struct inode *, struct dentry *);
1671 int (*setattr) (struct dentry *, struct iattr *);
1672 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1673 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1674 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1675 ssize_t (*listxattr) (struct dentry *, char *, size_t);
1676 int (*removexattr) (struct dentry *, const char *);
1677 int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
1679 int (*update_time)(struct inode *, struct timespec *, int);
1680 int (*atomic_open)(struct inode *, struct dentry *,
1681 struct file *, unsigned open_flag,
1682 umode_t create_mode, int *opened);
1683 int (*tmpfile) (struct inode *, struct dentry *, umode_t);
1684 } ____cacheline_aligned;
1686 ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
1687 unsigned long nr_segs, unsigned long fast_segs,
1688 struct iovec *fast_pointer,
1689 struct iovec **ret_pointer);
1691 extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1692 extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1693 extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1694 unsigned long, loff_t *);
1695 extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1696 unsigned long, loff_t *);
1698 struct super_operations {
1699 struct inode *(*alloc_inode)(struct super_block *sb);
1700 void (*destroy_inode)(struct inode *);
1702 void (*dirty_inode) (struct inode *, int flags);
1703 int (*write_inode) (struct inode *, struct writeback_control *wbc);
1704 int (*drop_inode) (struct inode *);
1705 void (*evict_inode) (struct inode *);
1706 void (*put_super) (struct super_block *);
1707 int (*sync_fs)(struct super_block *sb, int wait);
1708 int (*freeze_fs) (struct super_block *);
1709 int (*unfreeze_fs) (struct super_block *);
1710 int (*statfs) (struct dentry *, struct kstatfs *);
1711 int (*remount_fs) (struct super_block *, int *, char *);
1712 void (*umount_begin) (struct super_block *);
1714 int (*show_options)(struct seq_file *, struct dentry *);
1715 int (*show_devname)(struct seq_file *, struct dentry *);
1716 int (*show_path)(struct seq_file *, struct dentry *);
1717 int (*show_stats)(struct seq_file *, struct dentry *);
1719 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1720 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
1722 int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
1723 long (*nr_cached_objects)(struct super_block *, int);
1724 long (*free_cached_objects)(struct super_block *, long, int);
1728 * Inode flags - they have no relation to superblock flags now
1730 #define S_SYNC 1 /* Writes are synced at once */
1731 #define S_NOATIME 2 /* Do not update access times */
1732 #define S_APPEND 4 /* Append-only file */
1733 #define S_IMMUTABLE 8 /* Immutable file */
1734 #define S_DEAD 16 /* removed, but still open directory */
1735 #define S_NOQUOTA 32 /* Inode is not counted to quota */
1736 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
1737 #define S_NOCMTIME 128 /* Do not update file c/mtime */
1738 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
1739 #define S_PRIVATE 512 /* Inode is fs-internal */
1740 #define S_IMA 1024 /* Inode has an associated IMA struct */
1741 #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
1742 #define S_NOSEC 4096 /* no suid or xattr security attributes */
1745 * Note that nosuid etc flags are inode-specific: setting some file-system
1746 * flags just means all the inodes inherit those flags by default. It might be
1747 * possible to override it selectively if you really wanted to with some
1748 * ioctl() that is not currently implemented.
1750 * Exception: MS_RDONLY is always applied to the entire file system.
1752 * Unfortunately, it is possible to change a filesystems flags with it mounted
1753 * with files in use. This means that all of the inodes will not have their
1754 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
1755 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
1757 #define __IS_FLG(inode, flg) ((inode)->i_sb->s_flags & (flg))
1759 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
1760 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
1761 ((inode)->i_flags & S_SYNC))
1762 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
1763 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
1764 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
1765 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
1766 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
1768 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
1769 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
1770 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
1771 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
1773 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
1774 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
1775 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
1776 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
1777 #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
1778 #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
1779 #define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
1782 * Inode state bits. Protected by inode->i_lock
1784 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1785 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1787 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1788 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1789 * various stages of removing an inode.
1791 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1793 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1794 * fdatasync(). i_atime is the usual cause.
1795 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1796 * these changes separately from I_DIRTY_SYNC so that we
1797 * don't have to write inode on fdatasync() when only
1798 * mtime has changed in it.
1799 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1800 * I_NEW Serves as both a mutex and completion notification.
1801 * New inodes set I_NEW. If two processes both create
1802 * the same inode, one of them will release its inode and
1803 * wait for I_NEW to be released before returning.
1804 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1805 * also cause waiting on I_NEW, without I_NEW actually
1806 * being set. find_inode() uses this to prevent returning
1807 * nearly-dead inodes.
1808 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1809 * is zero. I_FREEING must be set when I_WILL_FREE is
1811 * I_FREEING Set when inode is about to be freed but still has dirty
1812 * pages or buffers attached or the inode itself is still
1814 * I_CLEAR Added by clear_inode(). In this state the inode is
1815 * clean and can be destroyed. Inode keeps I_FREEING.
1817 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1818 * prohibited for many purposes. iget() must wait for
1819 * the inode to be completely released, then create it
1820 * anew. Other functions will just ignore such inodes,
1821 * if appropriate. I_NEW is used for waiting.
1823 * I_SYNC Writeback of inode is running. The bit is set during
1824 * data writeback, and cleared with a wakeup on the bit
1825 * address once it is done. The bit is also used to pin
1826 * the inode in memory for flusher thread.
1828 * I_REFERENCED Marks the inode as recently references on the LRU list.
1830 * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit().
1832 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1834 #define I_DIRTY_SYNC (1 << 0)
1835 #define I_DIRTY_DATASYNC (1 << 1)
1836 #define I_DIRTY_PAGES (1 << 2)
1838 #define I_NEW (1 << __I_NEW)
1839 #define I_WILL_FREE (1 << 4)
1840 #define I_FREEING (1 << 5)
1841 #define I_CLEAR (1 << 6)
1843 #define I_SYNC (1 << __I_SYNC)
1844 #define I_REFERENCED (1 << 8)
1845 #define __I_DIO_WAKEUP 9
1846 #define I_DIO_WAKEUP (1 << I_DIO_WAKEUP)
1847 #define I_LINKABLE (1 << 10)
1849 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1851 extern void __mark_inode_dirty(struct inode *, int);
1852 static inline void mark_inode_dirty(struct inode *inode)
1854 __mark_inode_dirty(inode, I_DIRTY);
1857 static inline void mark_inode_dirty_sync(struct inode *inode)
1859 __mark_inode_dirty(inode, I_DIRTY_SYNC);
1862 extern void inc_nlink(struct inode *inode);
1863 extern void drop_nlink(struct inode *inode);
1864 extern void clear_nlink(struct inode *inode);
1865 extern void set_nlink(struct inode *inode, unsigned int nlink);
1867 static inline void inode_inc_link_count(struct inode *inode)
1870 mark_inode_dirty(inode);
1873 static inline void inode_dec_link_count(struct inode *inode)
1876 mark_inode_dirty(inode);
1880 * inode_inc_iversion - increments i_version
1881 * @inode: inode that need to be updated
1883 * Every time the inode is modified, the i_version field will be incremented.
1884 * The filesystem has to be mounted with i_version flag
1887 static inline void inode_inc_iversion(struct inode *inode)
1889 spin_lock(&inode->i_lock);
1891 spin_unlock(&inode->i_lock);
1894 enum file_time_flags {
1901 extern void touch_atime(const struct path *);
1902 static inline void file_accessed(struct file *file)
1904 if (!(file->f_flags & O_NOATIME))
1905 touch_atime(&file->f_path);
1908 int sync_inode(struct inode *inode, struct writeback_control *wbc);
1909 int sync_inode_metadata(struct inode *inode, int wait);
1911 struct file_system_type {
1914 #define FS_REQUIRES_DEV 1
1915 #define FS_BINARY_MOUNTDATA 2
1916 #define FS_HAS_SUBTYPE 4
1917 #define FS_USERNS_MOUNT 8 /* Can be mounted by userns root */
1918 #define FS_USERNS_DEV_MOUNT 16 /* A userns mount does not imply MNT_NODEV */
1919 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() during rename() internally. */
1920 struct dentry *(*mount) (struct file_system_type *, int,
1921 const char *, void *);
1922 void (*kill_sb) (struct super_block *);
1923 struct module *owner;
1924 struct file_system_type * next;
1925 struct hlist_head fs_supers;
1927 struct lock_class_key s_lock_key;
1928 struct lock_class_key s_umount_key;
1929 struct lock_class_key s_vfs_rename_key;
1930 struct lock_class_key s_writers_key[SB_FREEZE_LEVELS];
1932 struct lock_class_key i_lock_key;
1933 struct lock_class_key i_mutex_key;
1934 struct lock_class_key i_mutex_dir_key;
1937 #define MODULE_ALIAS_FS(NAME) MODULE_ALIAS("fs-" NAME)
1939 extern struct dentry *mount_ns(struct file_system_type *fs_type, int flags,
1940 void *data, int (*fill_super)(struct super_block *, void *, int));
1941 extern struct dentry *mount_bdev(struct file_system_type *fs_type,
1942 int flags, const char *dev_name, void *data,
1943 int (*fill_super)(struct super_block *, void *, int));
1944 extern struct dentry *mount_single(struct file_system_type *fs_type,
1945 int flags, void *data,
1946 int (*fill_super)(struct super_block *, void *, int));
1947 extern struct dentry *mount_nodev(struct file_system_type *fs_type,
1948 int flags, void *data,
1949 int (*fill_super)(struct super_block *, void *, int));
1950 extern struct dentry *mount_subtree(struct vfsmount *mnt, const char *path);
1951 void generic_shutdown_super(struct super_block *sb);
1952 void kill_block_super(struct super_block *sb);
1953 void kill_anon_super(struct super_block *sb);
1954 void kill_litter_super(struct super_block *sb);
1955 void deactivate_super(struct super_block *sb);
1956 void deactivate_locked_super(struct super_block *sb);
1957 int set_anon_super(struct super_block *s, void *data);
1958 int get_anon_bdev(dev_t *);
1959 void free_anon_bdev(dev_t);
1960 struct super_block *sget(struct file_system_type *type,
1961 int (*test)(struct super_block *,void *),
1962 int (*set)(struct super_block *,void *),
1963 int flags, void *data);
1964 extern struct dentry *mount_pseudo(struct file_system_type *, char *,
1965 const struct super_operations *ops,
1966 const struct dentry_operations *dops,
1969 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1970 #define fops_get(fops) \
1971 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1972 #define fops_put(fops) \
1973 do { if (fops) module_put((fops)->owner); } while(0)
1975 extern int register_filesystem(struct file_system_type *);
1976 extern int unregister_filesystem(struct file_system_type *);
1977 extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
1978 #define kern_mount(type) kern_mount_data(type, NULL)
1979 extern void kern_unmount(struct vfsmount *mnt);
1980 extern int may_umount_tree(struct vfsmount *);
1981 extern int may_umount(struct vfsmount *);
1982 extern long do_mount(const char *, const char *, const char *, unsigned long, void *);
1983 extern struct vfsmount *collect_mounts(struct path *);
1984 extern void drop_collected_mounts(struct vfsmount *);
1985 extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
1987 extern int vfs_statfs(struct path *, struct kstatfs *);
1988 extern int user_statfs(const char __user *, struct kstatfs *);
1989 extern int fd_statfs(int, struct kstatfs *);
1990 extern int vfs_ustat(dev_t, struct kstatfs *);
1991 extern int freeze_super(struct super_block *super);
1992 extern int thaw_super(struct super_block *super);
1993 extern bool our_mnt(struct vfsmount *mnt);
1994 extern bool fs_fully_visible(struct file_system_type *);
1996 extern int current_umask(void);
1999 extern struct kobject *fs_kobj;
2001 #define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
2003 #define FLOCK_VERIFY_READ 1
2004 #define FLOCK_VERIFY_WRITE 2
2006 #ifdef CONFIG_FILE_LOCKING
2007 extern int locks_mandatory_locked(struct inode *);
2008 extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
2011 * Candidates for mandatory locking have the setgid bit set
2012 * but no group execute bit - an otherwise meaningless combination.
2015 static inline int __mandatory_lock(struct inode *ino)
2017 return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
2021 * ... and these candidates should be on MS_MANDLOCK mounted fs,
2022 * otherwise these will be advisory locks
2025 static inline int mandatory_lock(struct inode *ino)
2027 return IS_MANDLOCK(ino) && __mandatory_lock(ino);
2030 static inline int locks_verify_locked(struct inode *inode)
2032 if (mandatory_lock(inode))
2033 return locks_mandatory_locked(inode);
2037 static inline int locks_verify_truncate(struct inode *inode,
2041 if (inode->i_flock && mandatory_lock(inode))
2042 return locks_mandatory_area(
2043 FLOCK_VERIFY_WRITE, inode, filp,
2044 size < inode->i_size ? size : inode->i_size,
2045 (size < inode->i_size ? inode->i_size - size
2046 : size - inode->i_size)
2051 static inline int break_lease(struct inode *inode, unsigned int mode)
2054 return __break_lease(inode, mode);
2057 #else /* !CONFIG_FILE_LOCKING */
2058 static inline int locks_mandatory_locked(struct inode *inode)
2063 static inline int locks_mandatory_area(int rw, struct inode *inode,
2064 struct file *filp, loff_t offset,
2070 static inline int __mandatory_lock(struct inode *inode)
2075 static inline int mandatory_lock(struct inode *inode)
2080 static inline int locks_verify_locked(struct inode *inode)
2085 static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
2091 static inline int break_lease(struct inode *inode, unsigned int mode)
2096 #endif /* CONFIG_FILE_LOCKING */
2101 const char *name; /* pointer to actual string */
2102 const __user char *uptr; /* original userland pointer */
2103 struct audit_names *aname;
2104 bool separate; /* should "name" be freed? */
2107 extern long vfs_truncate(struct path *, loff_t);
2108 extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
2110 extern int do_fallocate(struct file *file, int mode, loff_t offset,
2112 extern long do_sys_open(int dfd, const char __user *filename, int flags,
2114 extern struct file *file_open_name(struct filename *, int, umode_t);
2115 extern struct file *filp_open(const char *, int, umode_t);
2116 extern struct file *file_open_root(struct dentry *, struct vfsmount *,
2118 extern struct file * dentry_open(const struct path *, int, const struct cred *);
2119 extern int filp_close(struct file *, fl_owner_t id);
2121 extern struct filename *getname(const char __user *);
2127 extern int finish_open(struct file *file, struct dentry *dentry,
2128 int (*open)(struct inode *, struct file *),
2130 extern int finish_no_open(struct file *file, struct dentry *dentry);
2134 extern int ioctl_preallocate(struct file *filp, void __user *argp);
2137 extern void __init vfs_caches_init_early(void);
2138 extern void __init vfs_caches_init(unsigned long);
2140 extern struct kmem_cache *names_cachep;
2142 extern void final_putname(struct filename *name);
2144 #define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL)
2145 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
2146 #ifndef CONFIG_AUDITSYSCALL
2147 #define putname(name) final_putname(name)
2149 extern void putname(struct filename *name);
2153 extern int register_blkdev(unsigned int, const char *);
2154 extern void unregister_blkdev(unsigned int, const char *);
2155 extern struct block_device *bdget(dev_t);
2156 extern struct block_device *bdgrab(struct block_device *bdev);
2157 extern void bd_set_size(struct block_device *, loff_t size);
2158 extern void bd_forget(struct inode *inode);
2159 extern void bdput(struct block_device *);
2160 extern void invalidate_bdev(struct block_device *);
2161 extern void iterate_bdevs(void (*)(struct block_device *, void *), void *);
2162 extern int sync_blockdev(struct block_device *bdev);
2163 extern void kill_bdev(struct block_device *);
2164 extern struct super_block *freeze_bdev(struct block_device *);
2165 extern void emergency_thaw_all(void);
2166 extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
2167 extern int fsync_bdev(struct block_device *);
2168 extern int sb_is_blkdev_sb(struct super_block *sb);
2170 static inline void bd_forget(struct inode *inode) {}
2171 static inline int sync_blockdev(struct block_device *bdev) { return 0; }
2172 static inline void kill_bdev(struct block_device *bdev) {}
2173 static inline void invalidate_bdev(struct block_device *bdev) {}
2175 static inline struct super_block *freeze_bdev(struct block_device *sb)
2180 static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
2185 static inline void iterate_bdevs(void (*f)(struct block_device *, void *), void *arg)
2189 static inline int sb_is_blkdev_sb(struct super_block *sb)
2194 extern int sync_filesystem(struct super_block *);
2195 extern const struct file_operations def_blk_fops;
2196 extern const struct file_operations def_chr_fops;
2197 extern const struct file_operations bad_sock_fops;
2199 extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
2200 extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
2201 extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
2202 extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
2203 extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
2205 extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
2207 extern void blkdev_put(struct block_device *bdev, fmode_t mode);
2209 extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
2210 extern void bd_unlink_disk_holder(struct block_device *bdev,
2211 struct gendisk *disk);
2213 static inline int bd_link_disk_holder(struct block_device *bdev,
2214 struct gendisk *disk)
2218 static inline void bd_unlink_disk_holder(struct block_device *bdev,
2219 struct gendisk *disk)
2226 #define CHRDEV_MAJOR_HASH_SIZE 255
2227 extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
2228 extern int register_chrdev_region(dev_t, unsigned, const char *);
2229 extern int __register_chrdev(unsigned int major, unsigned int baseminor,
2230 unsigned int count, const char *name,
2231 const struct file_operations *fops);
2232 extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
2233 unsigned int count, const char *name);
2234 extern void unregister_chrdev_region(dev_t, unsigned);
2235 extern void chrdev_show(struct seq_file *,off_t);
2237 static inline int register_chrdev(unsigned int major, const char *name,
2238 const struct file_operations *fops)
2240 return __register_chrdev(major, 0, 256, name, fops);
2243 static inline void unregister_chrdev(unsigned int major, const char *name)
2245 __unregister_chrdev(major, 0, 256, name);
2248 /* fs/block_dev.c */
2249 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
2250 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
2253 #define BLKDEV_MAJOR_HASH_SIZE 255
2254 extern const char *__bdevname(dev_t, char *buffer);
2255 extern const char *bdevname(struct block_device *bdev, char *buffer);
2256 extern struct block_device *lookup_bdev(const char *);
2257 extern void blkdev_show(struct seq_file *,off_t);
2260 #define BLKDEV_MAJOR_HASH_SIZE 0
2263 extern void init_special_inode(struct inode *, umode_t, dev_t);
2265 /* Invalid inode operations -- fs/bad_inode.c */
2266 extern void make_bad_inode(struct inode *);
2267 extern int is_bad_inode(struct inode *);
2271 * return READ, READA, or WRITE
2273 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2276 * return data direction, READ or WRITE
2278 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
2280 extern void check_disk_size_change(struct gendisk *disk,
2281 struct block_device *bdev);
2282 extern int revalidate_disk(struct gendisk *);
2283 extern int check_disk_change(struct block_device *);
2284 extern int __invalidate_device(struct block_device *, bool);
2285 extern int invalidate_partition(struct gendisk *, int);
2287 unsigned long invalidate_mapping_pages(struct address_space *mapping,
2288 pgoff_t start, pgoff_t end);
2290 static inline void invalidate_remote_inode(struct inode *inode)
2292 if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
2293 S_ISLNK(inode->i_mode))
2294 invalidate_mapping_pages(inode->i_mapping, 0, -1);
2296 extern int invalidate_inode_pages2(struct address_space *mapping);
2297 extern int invalidate_inode_pages2_range(struct address_space *mapping,
2298 pgoff_t start, pgoff_t end);
2299 extern int write_inode_now(struct inode *, int);
2300 extern int filemap_fdatawrite(struct address_space *);
2301 extern int filemap_flush(struct address_space *);
2302 extern int filemap_fdatawait(struct address_space *);
2303 extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
2305 extern int filemap_write_and_wait(struct address_space *mapping);
2306 extern int filemap_write_and_wait_range(struct address_space *mapping,
2307 loff_t lstart, loff_t lend);
2308 extern int __filemap_fdatawrite_range(struct address_space *mapping,
2309 loff_t start, loff_t end, int sync_mode);
2310 extern int filemap_fdatawrite_range(struct address_space *mapping,
2311 loff_t start, loff_t end);
2313 extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
2315 extern int vfs_fsync(struct file *file, int datasync);
2316 extern int generic_write_sync(struct file *file, loff_t pos, loff_t count);
2317 extern void emergency_sync(void);
2318 extern void emergency_remount(void);
2320 extern sector_t bmap(struct inode *, sector_t);
2322 extern int notify_change(struct dentry *, struct iattr *);
2323 extern int inode_permission(struct inode *, int);
2324 extern int generic_permission(struct inode *, int);
2326 static inline bool execute_ok(struct inode *inode)
2328 return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
2331 static inline struct inode *file_inode(struct file *f)
2336 static inline void file_start_write(struct file *file)
2338 if (!S_ISREG(file_inode(file)->i_mode))
2340 __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, true);
2343 static inline bool file_start_write_trylock(struct file *file)
2345 if (!S_ISREG(file_inode(file)->i_mode))
2347 return __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, false);
2350 static inline void file_end_write(struct file *file)
2352 if (!S_ISREG(file_inode(file)->i_mode))
2354 __sb_end_write(file_inode(file)->i_sb, SB_FREEZE_WRITE);
2358 * get_write_access() gets write permission for a file.
2359 * put_write_access() releases this write permission.
2360 * This is used for regular files.
2361 * We cannot support write (and maybe mmap read-write shared) accesses and
2362 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2363 * can have the following values:
2364 * 0: no writers, no VM_DENYWRITE mappings
2365 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2366 * > 0: (i_writecount) users are writing to the file.
2368 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2369 * except for the cases where we don't hold i_writecount yet. Then we need to
2370 * use {get,deny}_write_access() - these functions check the sign and refuse
2371 * to do the change if sign is wrong.
2373 static inline int get_write_access(struct inode *inode)
2375 return atomic_inc_unless_negative(&inode->i_writecount) ? 0 : -ETXTBSY;
2377 static inline int deny_write_access(struct file *file)
2379 struct inode *inode = file_inode(file);
2380 return atomic_dec_unless_positive(&inode->i_writecount) ? 0 : -ETXTBSY;
2382 static inline void put_write_access(struct inode * inode)
2384 atomic_dec(&inode->i_writecount);
2386 static inline void allow_write_access(struct file *file)
2389 atomic_inc(&file_inode(file)->i_writecount);
2392 static inline void i_readcount_dec(struct inode *inode)
2394 BUG_ON(!atomic_read(&inode->i_readcount));
2395 atomic_dec(&inode->i_readcount);
2397 static inline void i_readcount_inc(struct inode *inode)
2399 atomic_inc(&inode->i_readcount);
2402 static inline void i_readcount_dec(struct inode *inode)
2406 static inline void i_readcount_inc(struct inode *inode)
2411 extern int do_pipe_flags(int *, int);
2413 extern int kernel_read(struct file *, loff_t, char *, unsigned long);
2414 extern ssize_t kernel_write(struct file *, const char *, size_t, loff_t);
2415 extern struct file * open_exec(const char *);
2417 /* fs/dcache.c -- generic fs support functions */
2418 extern int is_subdir(struct dentry *, struct dentry *);
2419 extern int path_is_under(struct path *, struct path *);
2421 #include <linux/err.h>
2423 /* needed for stackable file system support */
2424 extern loff_t default_llseek(struct file *file, loff_t offset, int whence);
2426 extern loff_t vfs_llseek(struct file *file, loff_t offset, int whence);
2428 extern int inode_init_always(struct super_block *, struct inode *);
2429 extern void inode_init_once(struct inode *);
2430 extern void address_space_init_once(struct address_space *mapping);
2431 extern void ihold(struct inode * inode);
2432 extern void iput(struct inode *);
2433 extern struct inode * igrab(struct inode *);
2434 extern ino_t iunique(struct super_block *, ino_t);
2435 extern int inode_needs_sync(struct inode *inode);
2436 extern int generic_delete_inode(struct inode *inode);
2437 static inline int generic_drop_inode(struct inode *inode)
2439 return !inode->i_nlink || inode_unhashed(inode);
2442 extern struct inode *ilookup5_nowait(struct super_block *sb,
2443 unsigned long hashval, int (*test)(struct inode *, void *),
2445 extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
2446 int (*test)(struct inode *, void *), void *data);
2447 extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
2449 extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
2450 extern struct inode * iget_locked(struct super_block *, unsigned long);
2451 extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
2452 extern int insert_inode_locked(struct inode *);
2453 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2454 extern void lockdep_annotate_inode_mutex_key(struct inode *inode);
2456 static inline void lockdep_annotate_inode_mutex_key(struct inode *inode) { };
2458 extern void unlock_new_inode(struct inode *);
2459 extern unsigned int get_next_ino(void);
2461 extern void __iget(struct inode * inode);
2462 extern void iget_failed(struct inode *);
2463 extern void clear_inode(struct inode *);
2464 extern void __destroy_inode(struct inode *);
2465 extern struct inode *new_inode_pseudo(struct super_block *sb);
2466 extern struct inode *new_inode(struct super_block *sb);
2467 extern void free_inode_nonrcu(struct inode *inode);
2468 extern int should_remove_suid(struct dentry *);
2469 extern int file_remove_suid(struct file *);
2471 extern void __insert_inode_hash(struct inode *, unsigned long hashval);
2472 static inline void insert_inode_hash(struct inode *inode)
2474 __insert_inode_hash(inode, inode->i_ino);
2477 extern void __remove_inode_hash(struct inode *);
2478 static inline void remove_inode_hash(struct inode *inode)
2480 if (!inode_unhashed(inode))
2481 __remove_inode_hash(inode);
2484 extern void inode_sb_list_add(struct inode *inode);
2487 extern void submit_bio(int, struct bio *);
2488 extern int bdev_read_only(struct block_device *);
2490 extern int set_blocksize(struct block_device *, int);
2491 extern int sb_set_blocksize(struct super_block *, int);
2492 extern int sb_min_blocksize(struct super_block *, int);
2494 extern int generic_file_mmap(struct file *, struct vm_area_struct *);
2495 extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
2496 extern int generic_file_remap_pages(struct vm_area_struct *, unsigned long addr,
2497 unsigned long size, pgoff_t pgoff);
2498 extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
2499 int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
2500 extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2501 extern ssize_t generic_file_read_iter(struct kiocb *, struct iov_iter *,
2503 extern ssize_t __generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long,
2505 extern ssize_t __generic_file_write_iter(struct kiocb *, struct iov_iter *,
2507 extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2508 extern ssize_t generic_file_write_iter(struct kiocb *, struct iov_iter *,
2510 extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
2511 unsigned long *, loff_t, loff_t *, size_t, size_t);
2512 extern ssize_t generic_file_direct_write_iter(struct kiocb *, struct iov_iter *,
2513 loff_t, loff_t *, size_t);
2514 extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
2515 unsigned long, loff_t, loff_t *, size_t, ssize_t);
2516 extern ssize_t generic_file_buffered_write_iter(struct kiocb *,
2517 struct iov_iter *, loff_t, loff_t *, size_t, ssize_t);
2518 extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
2519 extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
2520 extern int generic_segment_checks(const struct iovec *iov,
2521 unsigned long *nr_segs, size_t *count, int access_flags);
2523 /* fs/block_dev.c */
2524 extern ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
2525 unsigned long nr_segs, loff_t pos);
2526 extern int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
2528 extern void block_sync_page(struct page *page);
2531 extern ssize_t generic_file_splice_read(struct file *, loff_t *,
2532 struct pipe_inode_info *, size_t, unsigned int);
2533 extern ssize_t default_file_splice_read(struct file *, loff_t *,
2534 struct pipe_inode_info *, size_t, unsigned int);
2535 extern ssize_t generic_file_splice_write(struct pipe_inode_info *,
2536 struct file *, loff_t *, size_t, unsigned int);
2537 extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
2538 struct file *out, loff_t *, size_t len, unsigned int flags);
2541 file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
2542 extern loff_t noop_llseek(struct file *file, loff_t offset, int whence);
2543 extern loff_t no_llseek(struct file *file, loff_t offset, int whence);
2544 extern loff_t vfs_setpos(struct file *file, loff_t offset, loff_t maxsize);
2545 extern loff_t generic_file_llseek(struct file *file, loff_t offset, int whence);
2546 extern loff_t generic_file_llseek_size(struct file *file, loff_t offset,
2547 int whence, loff_t maxsize, loff_t eof);
2548 extern loff_t fixed_size_llseek(struct file *file, loff_t offset,
2549 int whence, loff_t size);
2550 extern int generic_file_open(struct inode * inode, struct file * filp);
2551 extern int nonseekable_open(struct inode * inode, struct file * filp);
2553 #ifdef CONFIG_FS_XIP
2554 extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
2556 extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
2557 extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
2558 size_t len, loff_t *ppos);
2559 extern int xip_truncate_page(struct address_space *mapping, loff_t from);
2561 static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
2568 typedef void (dio_submit_t)(int rw, struct bio *bio, struct inode *inode,
2569 loff_t file_offset);
2572 /* need locking between buffered and direct access */
2575 /* filesystem does not support filling holes */
2576 DIO_SKIP_HOLES = 0x02,
2579 void dio_end_io(struct bio *bio, int error);
2581 ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
2582 struct block_device *bdev, struct iov_iter *iter, loff_t offset,
2583 get_block_t get_block, dio_iodone_t end_io, dio_submit_t submit_io,
2586 static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
2587 struct inode *inode, struct iov_iter *iter, loff_t offset,
2588 get_block_t get_block)
2590 return __blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iter,
2591 offset, get_block, NULL, NULL,
2592 DIO_LOCKING | DIO_SKIP_HOLES);
2596 void inode_dio_wait(struct inode *inode);
2597 void inode_dio_done(struct inode *inode);
2599 extern const struct file_operations generic_ro_fops;
2601 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2603 extern int vfs_readlink(struct dentry *, char __user *, int, const char *);
2604 extern int page_readlink(struct dentry *, char __user *, int);
2605 extern void *page_follow_link_light(struct dentry *, struct nameidata *);
2606 extern void page_put_link(struct dentry *, struct nameidata *, void *);
2607 extern int __page_symlink(struct inode *inode, const char *symname, int len,
2609 extern int page_symlink(struct inode *inode, const char *symname, int len);
2610 extern const struct inode_operations page_symlink_inode_operations;
2611 extern int generic_readlink(struct dentry *, char __user *, int);
2612 extern void generic_fillattr(struct inode *, struct kstat *);
2613 extern int vfs_getattr(struct path *, struct kstat *);
2614 void __inode_add_bytes(struct inode *inode, loff_t bytes);
2615 void inode_add_bytes(struct inode *inode, loff_t bytes);
2616 void __inode_sub_bytes(struct inode *inode, loff_t bytes);
2617 void inode_sub_bytes(struct inode *inode, loff_t bytes);
2618 loff_t inode_get_bytes(struct inode *inode);
2619 void inode_set_bytes(struct inode *inode, loff_t bytes);
2621 extern int vfs_readdir(struct file *, filldir_t, void *);
2622 extern int iterate_dir(struct file *, struct dir_context *);
2624 extern int vfs_stat(const char __user *, struct kstat *);
2625 extern int vfs_lstat(const char __user *, struct kstat *);
2626 extern int vfs_fstat(unsigned int, struct kstat *);
2627 extern int vfs_fstatat(int , const char __user *, struct kstat *, int);
2629 extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
2631 extern int __generic_block_fiemap(struct inode *inode,
2632 struct fiemap_extent_info *fieinfo,
2633 loff_t start, loff_t len,
2634 get_block_t *get_block);
2635 extern int generic_block_fiemap(struct inode *inode,
2636 struct fiemap_extent_info *fieinfo, u64 start,
2637 u64 len, get_block_t *get_block);
2639 extern void get_filesystem(struct file_system_type *fs);
2640 extern void put_filesystem(struct file_system_type *fs);
2641 extern struct file_system_type *get_fs_type(const char *name);
2642 extern struct super_block *get_super(struct block_device *);
2643 extern struct super_block *get_super_thawed(struct block_device *);
2644 extern struct super_block *get_active_super(struct block_device *bdev);
2645 extern void drop_super(struct super_block *sb);
2646 extern void iterate_supers(void (*)(struct super_block *, void *), void *);
2647 extern void iterate_supers_type(struct file_system_type *,
2648 void (*)(struct super_block *, void *), void *);
2650 extern int dcache_dir_open(struct inode *, struct file *);
2651 extern int dcache_dir_close(struct inode *, struct file *);
2652 extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
2653 extern int dcache_readdir(struct file *, struct dir_context *);
2654 extern int simple_setattr(struct dentry *, struct iattr *);
2655 extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2656 extern int simple_statfs(struct dentry *, struct kstatfs *);
2657 extern int simple_open(struct inode *inode, struct file *file);
2658 extern int simple_link(struct dentry *, struct inode *, struct dentry *);
2659 extern int simple_unlink(struct inode *, struct dentry *);
2660 extern int simple_rmdir(struct inode *, struct dentry *);
2661 extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
2662 extern int noop_fsync(struct file *, loff_t, loff_t, int);
2663 extern int simple_empty(struct dentry *);
2664 extern int simple_readpage(struct file *file, struct page *page);
2665 extern int simple_write_begin(struct file *file, struct address_space *mapping,
2666 loff_t pos, unsigned len, unsigned flags,
2667 struct page **pagep, void **fsdata);
2668 extern int simple_write_end(struct file *file, struct address_space *mapping,
2669 loff_t pos, unsigned len, unsigned copied,
2670 struct page *page, void *fsdata);
2672 extern struct dentry *simple_lookup(struct inode *, struct dentry *, unsigned int flags);
2673 extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
2674 extern const struct file_operations simple_dir_operations;
2675 extern const struct inode_operations simple_dir_inode_operations;
2676 struct tree_descr { char *name; const struct file_operations *ops; int mode; };
2677 struct dentry *d_alloc_name(struct dentry *, const char *);
2678 extern int simple_fill_super(struct super_block *, unsigned long, struct tree_descr *);
2679 extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
2680 extern void simple_release_fs(struct vfsmount **mount, int *count);
2682 extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
2683 loff_t *ppos, const void *from, size_t available);
2684 extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
2685 const void __user *from, size_t count);
2687 extern int generic_file_fsync(struct file *, loff_t, loff_t, int);
2689 extern int generic_check_addressable(unsigned, u64);
2691 #ifdef CONFIG_MIGRATION
2692 extern int buffer_migrate_page(struct address_space *,
2693 struct page *, struct page *,
2696 #define buffer_migrate_page NULL
2699 extern int inode_change_ok(const struct inode *, struct iattr *);
2700 extern int inode_newsize_ok(const struct inode *, loff_t offset);
2701 extern void setattr_copy(struct inode *inode, const struct iattr *attr);
2703 extern int file_update_time(struct file *file);
2705 extern int generic_show_options(struct seq_file *m, struct dentry *root);
2706 extern void save_mount_options(struct super_block *sb, char *options);
2707 extern void replace_mount_options(struct super_block *sb, char *options);
2709 static inline ino_t parent_ino(struct dentry *dentry)
2714 * Don't strictly need d_lock here? If the parent ino could change
2715 * then surely we'd have a deeper race in the caller?
2717 spin_lock(&dentry->d_lock);
2718 res = dentry->d_parent->d_inode->i_ino;
2719 spin_unlock(&dentry->d_lock);
2723 /* Transaction based IO helpers */
2726 * An argresp is stored in an allocated page and holds the
2727 * size of the argument or response, along with its content
2729 struct simple_transaction_argresp {
2734 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2736 char *simple_transaction_get(struct file *file, const char __user *buf,
2738 ssize_t simple_transaction_read(struct file *file, char __user *buf,
2739 size_t size, loff_t *pos);
2740 int simple_transaction_release(struct inode *inode, struct file *file);
2742 void simple_transaction_set(struct file *file, size_t n);
2745 * simple attribute files
2747 * These attributes behave similar to those in sysfs:
2749 * Writing to an attribute immediately sets a value, an open file can be
2750 * written to multiple times.
2752 * Reading from an attribute creates a buffer from the value that might get
2753 * read with multiple read calls. When the attribute has been read
2754 * completely, no further read calls are possible until the file is opened
2757 * All attributes contain a text representation of a numeric value
2758 * that are accessed with the get() and set() functions.
2760 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2761 static int __fops ## _open(struct inode *inode, struct file *file) \
2763 __simple_attr_check_format(__fmt, 0ull); \
2764 return simple_attr_open(inode, file, __get, __set, __fmt); \
2766 static const struct file_operations __fops = { \
2767 .owner = THIS_MODULE, \
2768 .open = __fops ## _open, \
2769 .release = simple_attr_release, \
2770 .read = simple_attr_read, \
2771 .write = simple_attr_write, \
2772 .llseek = generic_file_llseek, \
2775 static inline __printf(1, 2)
2776 void __simple_attr_check_format(const char *fmt, ...)
2778 /* don't do anything, just let the compiler check the arguments; */
2781 int simple_attr_open(struct inode *inode, struct file *file,
2782 int (*get)(void *, u64 *), int (*set)(void *, u64),
2784 int simple_attr_release(struct inode *inode, struct file *file);
2785 ssize_t simple_attr_read(struct file *file, char __user *buf,
2786 size_t len, loff_t *ppos);
2787 ssize_t simple_attr_write(struct file *file, const char __user *buf,
2788 size_t len, loff_t *ppos);
2791 int proc_nr_files(struct ctl_table *table, int write,
2792 void __user *buffer, size_t *lenp, loff_t *ppos);
2793 int proc_nr_dentry(struct ctl_table *table, int write,
2794 void __user *buffer, size_t *lenp, loff_t *ppos);
2795 int proc_nr_inodes(struct ctl_table *table, int write,
2796 void __user *buffer, size_t *lenp, loff_t *ppos);
2797 int __init get_filesystem_list(char *buf);
2799 #define __FMODE_EXEC ((__force int) FMODE_EXEC)
2800 #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
2802 #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
2803 #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
2804 (flag & __FMODE_NONOTIFY)))
2806 static inline int is_sxid(umode_t mode)
2808 return (mode & S_ISUID) || ((mode & S_ISGID) && (mode & S_IXGRP));
2811 static inline void inode_has_no_xattr(struct inode *inode)
2813 if (!is_sxid(inode->i_mode) && (inode->i_sb->s_flags & MS_NOSEC))
2814 inode->i_flags |= S_NOSEC;
2817 static inline bool dir_emit(struct dir_context *ctx,
2818 const char *name, int namelen,
2819 u64 ino, unsigned type)
2821 return ctx->actor(ctx, name, namelen, ctx->pos, ino, type) == 0;
2823 static inline bool dir_emit_dot(struct file *file, struct dir_context *ctx)
2825 return ctx->actor(ctx, ".", 1, ctx->pos,
2826 file->f_path.dentry->d_inode->i_ino, DT_DIR) == 0;
2828 static inline bool dir_emit_dotdot(struct file *file, struct dir_context *ctx)
2830 return ctx->actor(ctx, "..", 2, ctx->pos,
2831 parent_ino(file->f_path.dentry), DT_DIR) == 0;
2833 static inline bool dir_emit_dots(struct file *file, struct dir_context *ctx)
2835 if (ctx->pos == 0) {
2836 if (!dir_emit_dot(file, ctx))
2840 if (ctx->pos == 1) {
2841 if (!dir_emit_dotdot(file, ctx))
2847 static inline bool dir_relax(struct inode *inode)
2849 mutex_unlock(&inode->i_mutex);
2850 mutex_lock(&inode->i_mutex);
2851 return !IS_DEADDIR(inode);
2854 #endif /* _LINUX_FS_H */