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 *);
315 static inline size_t iov_iter_copy_to_user_atomic(struct page *page,
316 struct iov_iter *i, unsigned long offset, size_t bytes)
318 return i->ops->ii_copy_to_user_atomic(page, i, offset, bytes);
320 static inline size_t __iov_iter_copy_to_user(struct page *page,
321 struct iov_iter *i, unsigned long offset, size_t bytes)
323 return i->ops->ii_copy_to_user(page, i, offset, bytes, 0);
325 static inline size_t iov_iter_copy_to_user(struct page *page,
326 struct iov_iter *i, unsigned long offset, size_t bytes)
328 return i->ops->ii_copy_to_user(page, i, offset, bytes, 1);
330 static inline size_t iov_iter_copy_from_user_atomic(struct page *page,
331 struct iov_iter *i, unsigned long offset, size_t bytes)
333 return i->ops->ii_copy_from_user_atomic(page, i, offset, bytes);
335 static inline size_t iov_iter_copy_from_user(struct page *page,
336 struct iov_iter *i, unsigned long offset, size_t bytes)
338 return i->ops->ii_copy_from_user(page, i, offset, bytes);
340 static inline void iov_iter_advance(struct iov_iter *i, size_t bytes)
342 return i->ops->ii_advance(i, bytes);
344 static inline int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes)
346 return i->ops->ii_fault_in_readable(i, bytes);
348 static inline size_t iov_iter_single_seg_count(const struct iov_iter *i)
350 return i->ops->ii_single_seg_count(i);
354 extern struct iov_iter_ops ii_bvec_ops;
357 static inline void iov_iter_init_bvec(struct iov_iter *i,
358 struct bio_vec *bvec,
359 unsigned long nr_segs,
360 size_t count, size_t written)
362 i->ops = &ii_bvec_ops;
363 i->data = (unsigned long)bvec;
364 i->nr_segs = nr_segs;
366 i->count = count + written;
368 iov_iter_advance(i, written);
372 extern struct iov_iter_ops ii_iovec_ops;
374 static inline void iov_iter_init(struct iov_iter *i,
375 const struct iovec *iov, unsigned long nr_segs,
376 size_t count, size_t written)
378 i->ops = &ii_iovec_ops;
379 i->data = (unsigned long)iov;
380 i->nr_segs = nr_segs;
382 i->count = count + written;
384 iov_iter_advance(i, written);
387 static inline struct iovec *iov_iter_iovec(struct iov_iter *i)
389 return (struct iovec *)i->data;
392 static inline size_t iov_iter_count(struct iov_iter *i)
398 * "descriptor" for what we're up to with a read.
399 * This allows us to use the same read code yet
400 * have multiple different users of the data that
401 * we read from a file.
403 * The simplest case just copies the data to user
416 typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
417 unsigned long, unsigned long);
419 struct address_space_operations {
420 int (*writepage)(struct page *page, struct writeback_control *wbc);
421 int (*readpage)(struct file *, struct page *);
423 /* Write back some dirty pages from this mapping. */
424 int (*writepages)(struct address_space *, struct writeback_control *);
426 /* Set a page dirty. Return true if this dirtied it */
427 int (*set_page_dirty)(struct page *page);
429 int (*readpages)(struct file *filp, struct address_space *mapping,
430 struct list_head *pages, unsigned nr_pages);
432 int (*write_begin)(struct file *, struct address_space *mapping,
433 loff_t pos, unsigned len, unsigned flags,
434 struct page **pagep, void **fsdata);
435 int (*write_end)(struct file *, struct address_space *mapping,
436 loff_t pos, unsigned len, unsigned copied,
437 struct page *page, void *fsdata);
439 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
440 sector_t (*bmap)(struct address_space *, sector_t);
441 void (*invalidatepage) (struct page *, unsigned int, unsigned int);
442 int (*releasepage) (struct page *, gfp_t);
443 void (*freepage)(struct page *);
444 ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
445 loff_t offset, unsigned long nr_segs);
446 int (*get_xip_mem)(struct address_space *, pgoff_t, int,
447 void **, unsigned long *);
449 * migrate the contents of a page to the specified target. If
450 * migrate_mode is MIGRATE_ASYNC, it must not block.
452 int (*migratepage) (struct address_space *,
453 struct page *, struct page *, enum migrate_mode);
454 int (*launder_page) (struct page *);
455 int (*is_partially_uptodate) (struct page *, read_descriptor_t *,
457 void (*is_dirty_writeback) (struct page *, bool *, bool *);
458 int (*error_remove_page)(struct address_space *, struct page *);
460 /* swapfile support */
461 int (*swap_activate)(struct swap_info_struct *sis, struct file *file,
463 void (*swap_deactivate)(struct file *file);
466 extern const struct address_space_operations empty_aops;
469 * pagecache_write_begin/pagecache_write_end must be used by general code
470 * to write into the pagecache.
472 int pagecache_write_begin(struct file *, struct address_space *mapping,
473 loff_t pos, unsigned len, unsigned flags,
474 struct page **pagep, void **fsdata);
476 int pagecache_write_end(struct file *, struct address_space *mapping,
477 loff_t pos, unsigned len, unsigned copied,
478 struct page *page, void *fsdata);
480 struct backing_dev_info;
481 struct address_space {
482 struct inode *host; /* owner: inode, block_device */
483 struct radix_tree_root page_tree; /* radix tree of all pages */
484 spinlock_t tree_lock; /* and lock protecting it */
485 unsigned int i_mmap_writable;/* count VM_SHARED mappings */
486 struct rb_root i_mmap; /* tree of private and shared mappings */
487 struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
488 struct mutex i_mmap_mutex; /* protect tree, count, list */
489 /* Protected by tree_lock together with the radix tree */
490 unsigned long nrpages; /* number of total pages */
491 pgoff_t writeback_index;/* writeback starts here */
492 const struct address_space_operations *a_ops; /* methods */
493 unsigned long flags; /* error bits/gfp mask */
494 struct backing_dev_info *backing_dev_info; /* device readahead, etc */
495 spinlock_t private_lock; /* for use by the address_space */
496 struct list_head private_list; /* ditto */
497 void *private_data; /* ditto */
498 } __attribute__((aligned(sizeof(long))));
500 * On most architectures that alignment is already the case; but
501 * must be enforced here for CRIS, to let the least significant bit
502 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
504 struct request_queue;
506 struct block_device {
507 dev_t bd_dev; /* not a kdev_t - it's a search key */
509 struct inode * bd_inode; /* will die */
510 struct super_block * bd_super;
511 struct mutex bd_mutex; /* open/close mutex */
512 struct list_head bd_inodes;
516 bool bd_write_holder;
518 struct list_head bd_holder_disks;
520 struct block_device * bd_contains;
521 unsigned bd_block_size;
522 struct hd_struct * bd_part;
523 /* number of times partitions within this device have been opened. */
524 unsigned bd_part_count;
526 struct gendisk * bd_disk;
527 struct request_queue * bd_queue;
528 struct list_head bd_list;
530 * Private data. You must have bd_claim'ed the block_device
531 * to use this. NOTE: bd_claim allows an owner to claim
532 * the same device multiple times, the owner must take special
533 * care to not mess up bd_private for that case.
535 unsigned long bd_private;
537 /* The counter of freeze processes */
538 int bd_fsfreeze_count;
539 /* Mutex for freeze */
540 struct mutex bd_fsfreeze_mutex;
544 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
547 #define PAGECACHE_TAG_DIRTY 0
548 #define PAGECACHE_TAG_WRITEBACK 1
549 #define PAGECACHE_TAG_TOWRITE 2
551 int mapping_tagged(struct address_space *mapping, int tag);
554 * Might pages of this file be mapped into userspace?
556 static inline int mapping_mapped(struct address_space *mapping)
558 return !RB_EMPTY_ROOT(&mapping->i_mmap) ||
559 !list_empty(&mapping->i_mmap_nonlinear);
563 * Might pages of this file have been modified in userspace?
564 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
565 * marks vma as VM_SHARED if it is shared, and the file was opened for
566 * writing i.e. vma may be mprotected writable even if now readonly.
568 static inline int mapping_writably_mapped(struct address_space *mapping)
570 return mapping->i_mmap_writable != 0;
574 * Use sequence counter to get consistent i_size on 32-bit processors.
576 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
577 #include <linux/seqlock.h>
578 #define __NEED_I_SIZE_ORDERED
579 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
581 #define i_size_ordered_init(inode) do { } while (0)
585 #define ACL_NOT_CACHED ((void *)(-1))
587 #define IOP_FASTPERM 0x0001
588 #define IOP_LOOKUP 0x0002
589 #define IOP_NOFOLLOW 0x0004
592 * Keep mostly read-only and often accessed (especially for
593 * the RCU path lookup and 'stat' data) fields at the beginning
594 * of the 'struct inode'
598 unsigned short i_opflags;
601 unsigned int i_flags;
603 #ifdef CONFIG_FS_POSIX_ACL
604 struct posix_acl *i_acl;
605 struct posix_acl *i_default_acl;
608 const struct inode_operations *i_op;
609 struct super_block *i_sb;
610 struct address_space *i_mapping;
612 #ifdef CONFIG_SECURITY
616 /* Stat data, not accessed from path walking */
619 * Filesystems may only read i_nlink directly. They shall use the
620 * following functions for modification:
622 * (set|clear|inc|drop)_nlink
623 * inode_(inc|dec)_link_count
626 const unsigned int i_nlink;
627 unsigned int __i_nlink;
631 struct timespec i_atime;
632 struct timespec i_mtime;
633 struct timespec i_ctime;
634 spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
635 unsigned short i_bytes;
636 unsigned int i_blkbits;
639 #ifdef __NEED_I_SIZE_ORDERED
640 seqcount_t i_size_seqcount;
644 unsigned long i_state;
645 struct mutex i_mutex;
647 unsigned long dirtied_when; /* jiffies of first dirtying */
649 struct hlist_node i_hash;
650 struct list_head i_wb_list; /* backing dev IO list */
651 struct list_head i_lru; /* inode LRU list */
652 struct list_head i_sb_list;
654 struct hlist_head i_dentry;
655 struct rcu_head i_rcu;
659 atomic_t i_dio_count;
660 atomic_t i_writecount;
661 const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
662 struct file_lock *i_flock;
663 struct address_space i_data;
665 struct dquot *i_dquot[MAXQUOTAS];
667 struct list_head i_devices;
669 struct pipe_inode_info *i_pipe;
670 struct block_device *i_bdev;
676 #ifdef CONFIG_FSNOTIFY
677 __u32 i_fsnotify_mask; /* all events this inode cares about */
678 struct hlist_head i_fsnotify_marks;
682 atomic_t i_readcount; /* struct files open RO */
684 void *i_private; /* fs or device private pointer */
687 static inline int inode_unhashed(struct inode *inode)
689 return hlist_unhashed(&inode->i_hash);
693 * inode->i_mutex nesting subclasses for the lock validator:
695 * 0: the object of the current VFS operation
700 * The locking order between these classes is
701 * parent -> child -> normal -> xattr -> quota
703 enum inode_i_mutex_lock_class
713 * NOTE: in a 32bit arch with a preemptable kernel and
714 * an UP compile the i_size_read/write must be atomic
715 * with respect to the local cpu (unlike with preempt disabled),
716 * but they don't need to be atomic with respect to other cpus like in
717 * true SMP (so they need either to either locally disable irq around
718 * the read or for example on x86 they can be still implemented as a
719 * cmpxchg8b without the need of the lock prefix). For SMP compiles
720 * and 64bit archs it makes no difference if preempt is enabled or not.
722 static inline loff_t i_size_read(const struct inode *inode)
724 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
729 seq = read_seqcount_begin(&inode->i_size_seqcount);
730 i_size = inode->i_size;
731 } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
733 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
737 i_size = inode->i_size;
741 return inode->i_size;
746 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
747 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
748 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
750 static inline void i_size_write(struct inode *inode, loff_t i_size)
752 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
754 write_seqcount_begin(&inode->i_size_seqcount);
755 inode->i_size = i_size;
756 write_seqcount_end(&inode->i_size_seqcount);
758 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
760 inode->i_size = i_size;
763 inode->i_size = i_size;
767 /* Helper functions so that in most cases filesystems will
768 * not need to deal directly with kuid_t and kgid_t and can
769 * instead deal with the raw numeric values that are stored
772 static inline uid_t i_uid_read(const struct inode *inode)
774 return from_kuid(&init_user_ns, inode->i_uid);
777 static inline gid_t i_gid_read(const struct inode *inode)
779 return from_kgid(&init_user_ns, inode->i_gid);
782 static inline void i_uid_write(struct inode *inode, uid_t uid)
784 inode->i_uid = make_kuid(&init_user_ns, uid);
787 static inline void i_gid_write(struct inode *inode, gid_t gid)
789 inode->i_gid = make_kgid(&init_user_ns, gid);
792 static inline unsigned iminor(const struct inode *inode)
794 return MINOR(inode->i_rdev);
797 static inline unsigned imajor(const struct inode *inode)
799 return MAJOR(inode->i_rdev);
802 extern struct block_device *I_BDEV(struct inode *inode);
805 rwlock_t lock; /* protects pid, uid, euid fields */
806 struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
807 enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
808 kuid_t uid, euid; /* uid/euid of process setting the owner */
809 int signum; /* posix.1b rt signal to be delivered on IO */
813 * Track a single file's readahead state
815 struct file_ra_state {
816 pgoff_t start; /* where readahead started */
817 unsigned int size; /* # of readahead pages */
818 unsigned int async_size; /* do asynchronous readahead when
819 there are only # of pages ahead */
821 unsigned int ra_pages; /* Maximum readahead window */
822 unsigned int mmap_miss; /* Cache miss stat for mmap accesses */
823 loff_t prev_pos; /* Cache last read() position */
827 * Check if @index falls in the readahead windows.
829 static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
831 return (index >= ra->start &&
832 index < ra->start + ra->size);
835 #define FILE_MNT_WRITE_TAKEN 1
836 #define FILE_MNT_WRITE_RELEASED 2
840 * fu_list becomes invalid after file_free is called and queued via
841 * fu_rcuhead for RCU freeing
844 struct list_head fu_list;
845 struct llist_node fu_llist;
846 struct rcu_head fu_rcuhead;
849 #define f_dentry f_path.dentry
850 struct inode *f_inode; /* cached value */
851 const struct file_operations *f_op;
854 * Protects f_ep_links, f_flags, f_pos vs i_size in lseek SEEK_CUR.
855 * Must not be taken from IRQ context.
861 atomic_long_t f_count;
862 unsigned int f_flags;
865 struct fown_struct f_owner;
866 const struct cred *f_cred;
867 struct file_ra_state f_ra;
870 #ifdef CONFIG_SECURITY
873 /* needed for tty driver, and maybe others */
877 /* Used by fs/eventpoll.c to link all the hooks to this file */
878 struct list_head f_ep_links;
879 struct list_head f_tfile_llink;
880 #endif /* #ifdef CONFIG_EPOLL */
881 struct address_space *f_mapping;
882 #ifdef CONFIG_DEBUG_WRITECOUNT
883 unsigned long f_mnt_write_state;
890 /* file identifier */
891 unsigned char f_handle[0];
894 static inline struct file *get_file(struct file *f)
896 atomic_long_inc(&f->f_count);
899 #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
900 #define file_count(x) atomic_long_read(&(x)->f_count)
902 #ifdef CONFIG_DEBUG_WRITECOUNT
903 static inline void file_take_write(struct file *f)
905 WARN_ON(f->f_mnt_write_state != 0);
906 f->f_mnt_write_state = FILE_MNT_WRITE_TAKEN;
908 static inline void file_release_write(struct file *f)
910 f->f_mnt_write_state |= FILE_MNT_WRITE_RELEASED;
912 static inline void file_reset_write(struct file *f)
914 f->f_mnt_write_state = 0;
916 static inline void file_check_state(struct file *f)
919 * At this point, either both or neither of these bits
922 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN);
923 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_RELEASED);
925 static inline int file_check_writeable(struct file *f)
927 if (f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN)
929 printk(KERN_WARNING "writeable file with no "
930 "mnt_want_write()\n");
934 #else /* !CONFIG_DEBUG_WRITECOUNT */
935 static inline void file_take_write(struct file *filp) {}
936 static inline void file_release_write(struct file *filp) {}
937 static inline void file_reset_write(struct file *filp) {}
938 static inline void file_check_state(struct file *filp) {}
939 static inline int file_check_writeable(struct file *filp)
943 #endif /* CONFIG_DEBUG_WRITECOUNT */
945 #define MAX_NON_LFS ((1UL<<31) - 1)
947 /* Page cache limit. The filesystems should put that into their s_maxbytes
948 limits, otherwise bad things can happen in VM. */
949 #if BITS_PER_LONG==32
950 #define MAX_LFS_FILESIZE (((loff_t)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
951 #elif BITS_PER_LONG==64
952 #define MAX_LFS_FILESIZE ((loff_t)0x7fffffffffffffffLL)
957 #define FL_ACCESS 8 /* not trying to lock, just looking */
958 #define FL_EXISTS 16 /* when unlocking, test for existence */
959 #define FL_LEASE 32 /* lease held on this file */
960 #define FL_CLOSE 64 /* unlock on close */
961 #define FL_SLEEP 128 /* A blocking lock */
962 #define FL_DOWNGRADE_PENDING 256 /* Lease is being downgraded */
963 #define FL_UNLOCK_PENDING 512 /* Lease is being broken */
966 * Special return value from posix_lock_file() and vfs_lock_file() for
967 * asynchronous locking.
969 #define FILE_LOCK_DEFERRED 1
972 * The POSIX file lock owner is determined by
973 * the "struct files_struct" in the thread group
974 * (or NULL for no owner - BSD locks).
976 * Lockd stuffs a "host" pointer into this.
978 typedef struct files_struct *fl_owner_t;
980 struct file_lock_operations {
981 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
982 void (*fl_release_private)(struct file_lock *);
985 struct lock_manager_operations {
986 int (*lm_compare_owner)(struct file_lock *, struct file_lock *);
987 unsigned long (*lm_owner_key)(struct file_lock *);
988 void (*lm_notify)(struct file_lock *); /* unblock callback */
989 int (*lm_grant)(struct file_lock *, struct file_lock *, int);
990 void (*lm_break)(struct file_lock *);
991 int (*lm_change)(struct file_lock **, int);
994 struct lock_manager {
995 struct list_head list;
999 void locks_start_grace(struct net *, struct lock_manager *);
1000 void locks_end_grace(struct lock_manager *);
1001 int locks_in_grace(struct net *);
1003 /* that will die - we need it for nfs_lock_info */
1004 #include <linux/nfs_fs_i.h>
1007 * struct file_lock represents a generic "file lock". It's used to represent
1008 * POSIX byte range locks, BSD (flock) locks, and leases. It's important to
1009 * note that the same struct is used to represent both a request for a lock and
1010 * the lock itself, but the same object is never used for both.
1012 * FIXME: should we create a separate "struct lock_request" to help distinguish
1015 * The i_flock list is ordered by:
1017 * 1) lock type -- FL_LEASEs first, then FL_FLOCK, and finally FL_POSIX
1019 * 3) lock range start
1022 * Obviously, the last two criteria only matter for POSIX locks.
1025 struct file_lock *fl_next; /* singly linked list for this inode */
1026 struct hlist_node fl_link; /* node in global lists */
1027 struct list_head fl_block; /* circular list of blocked processes */
1028 fl_owner_t fl_owner;
1029 unsigned int fl_flags;
1030 unsigned char fl_type;
1031 unsigned int fl_pid;
1032 int fl_link_cpu; /* what cpu's list is this on? */
1033 struct pid *fl_nspid;
1034 wait_queue_head_t fl_wait;
1035 struct file *fl_file;
1039 struct fasync_struct * fl_fasync; /* for lease break notifications */
1040 /* for lease breaks: */
1041 unsigned long fl_break_time;
1042 unsigned long fl_downgrade_time;
1044 const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
1045 const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
1047 struct nfs_lock_info nfs_fl;
1048 struct nfs4_lock_info nfs4_fl;
1050 struct list_head link; /* link in AFS vnode's pending_locks list */
1051 int state; /* state of grant or error if -ve */
1056 /* The following constant reflects the upper bound of the file/locking space */
1058 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
1059 #define OFFSET_MAX INT_LIMIT(loff_t)
1060 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
1063 #include <linux/fcntl.h>
1065 extern void send_sigio(struct fown_struct *fown, int fd, int band);
1067 #ifdef CONFIG_FILE_LOCKING
1068 extern int fcntl_getlk(struct file *, struct flock __user *);
1069 extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
1070 struct flock __user *);
1072 #if BITS_PER_LONG == 32
1073 extern int fcntl_getlk64(struct file *, struct flock64 __user *);
1074 extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
1075 struct flock64 __user *);
1078 extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
1079 extern int fcntl_getlease(struct file *filp);
1082 void locks_free_lock(struct file_lock *fl);
1083 extern void locks_init_lock(struct file_lock *);
1084 extern struct file_lock * locks_alloc_lock(void);
1085 extern void locks_copy_lock(struct file_lock *, struct file_lock *);
1086 extern void __locks_copy_lock(struct file_lock *, const struct file_lock *);
1087 extern void locks_remove_posix(struct file *, fl_owner_t);
1088 extern void locks_remove_flock(struct file *);
1089 extern void locks_release_private(struct file_lock *);
1090 extern void posix_test_lock(struct file *, struct file_lock *);
1091 extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
1092 extern int posix_lock_file_wait(struct file *, struct file_lock *);
1093 extern int posix_unblock_lock(struct file_lock *);
1094 extern int vfs_test_lock(struct file *, struct file_lock *);
1095 extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
1096 extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
1097 extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
1098 extern int __break_lease(struct inode *inode, unsigned int flags);
1099 extern void lease_get_mtime(struct inode *, struct timespec *time);
1100 extern int generic_setlease(struct file *, long, struct file_lock **);
1101 extern int vfs_setlease(struct file *, long, struct file_lock **);
1102 extern int lease_modify(struct file_lock **, int);
1103 extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
1104 extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
1105 #else /* !CONFIG_FILE_LOCKING */
1106 static inline int fcntl_getlk(struct file *file, struct flock __user *user)
1111 static inline int fcntl_setlk(unsigned int fd, struct file *file,
1112 unsigned int cmd, struct flock __user *user)
1117 #if BITS_PER_LONG == 32
1118 static inline int fcntl_getlk64(struct file *file, struct flock64 __user *user)
1123 static inline int fcntl_setlk64(unsigned int fd, struct file *file,
1124 unsigned int cmd, struct flock64 __user *user)
1129 static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
1134 static inline int fcntl_getlease(struct file *filp)
1139 static inline void locks_init_lock(struct file_lock *fl)
1144 static inline void __locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1149 static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1154 static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
1159 static inline void locks_remove_flock(struct file *filp)
1164 static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
1169 static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
1170 struct file_lock *conflock)
1175 static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
1180 static inline int posix_unblock_lock(struct file_lock *waiter)
1185 static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
1190 static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
1191 struct file_lock *fl, struct file_lock *conf)
1196 static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
1201 static inline int flock_lock_file_wait(struct file *filp,
1202 struct file_lock *request)
1207 static inline int __break_lease(struct inode *inode, unsigned int mode)
1212 static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
1217 static inline int generic_setlease(struct file *filp, long arg,
1218 struct file_lock **flp)
1223 static inline int vfs_setlease(struct file *filp, long arg,
1224 struct file_lock **lease)
1229 static inline int lease_modify(struct file_lock **before, int arg)
1234 static inline int lock_may_read(struct inode *inode, loff_t start,
1240 static inline int lock_may_write(struct inode *inode, loff_t start,
1245 #endif /* !CONFIG_FILE_LOCKING */
1248 struct fasync_struct {
1252 struct fasync_struct *fa_next; /* singly linked list */
1253 struct file *fa_file;
1254 struct rcu_head fa_rcu;
1257 #define FASYNC_MAGIC 0x4601
1259 /* SMP safe fasync helpers: */
1260 extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
1261 extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
1262 extern int fasync_remove_entry(struct file *, struct fasync_struct **);
1263 extern struct fasync_struct *fasync_alloc(void);
1264 extern void fasync_free(struct fasync_struct *);
1266 /* can be called from interrupts */
1267 extern void kill_fasync(struct fasync_struct **, int, int);
1269 extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
1270 extern int f_setown(struct file *filp, unsigned long arg, int force);
1271 extern void f_delown(struct file *filp);
1272 extern pid_t f_getown(struct file *filp);
1273 extern int send_sigurg(struct fown_struct *fown);
1281 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1282 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1283 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1284 #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1285 #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
1287 extern struct list_head super_blocks;
1288 extern spinlock_t sb_lock;
1290 /* Possible states of 'frozen' field */
1292 SB_UNFROZEN = 0, /* FS is unfrozen */
1293 SB_FREEZE_WRITE = 1, /* Writes, dir ops, ioctls frozen */
1294 SB_FREEZE_PAGEFAULT = 2, /* Page faults stopped as well */
1295 SB_FREEZE_FS = 3, /* For internal FS use (e.g. to stop
1296 * internal threads if needed) */
1297 SB_FREEZE_COMPLETE = 4, /* ->freeze_fs finished successfully */
1300 #define SB_FREEZE_LEVELS (SB_FREEZE_COMPLETE - 1)
1303 /* Counters for counting writers at each level */
1304 struct percpu_counter counter[SB_FREEZE_LEVELS];
1305 wait_queue_head_t wait; /* queue for waiting for
1306 writers / faults to finish */
1307 int frozen; /* Is sb frozen? */
1308 wait_queue_head_t wait_unfrozen; /* queue for waiting for
1310 #ifdef CONFIG_DEBUG_LOCK_ALLOC
1311 struct lockdep_map lock_map[SB_FREEZE_LEVELS];
1315 struct super_block {
1316 struct list_head s_list; /* Keep this first */
1317 dev_t s_dev; /* search index; _not_ kdev_t */
1318 unsigned char s_blocksize_bits;
1319 unsigned long s_blocksize;
1320 loff_t s_maxbytes; /* Max file size */
1321 struct file_system_type *s_type;
1322 const struct super_operations *s_op;
1323 const struct dquot_operations *dq_op;
1324 const struct quotactl_ops *s_qcop;
1325 const struct export_operations *s_export_op;
1326 unsigned long s_flags;
1327 unsigned long s_magic;
1328 struct dentry *s_root;
1329 struct rw_semaphore s_umount;
1332 #ifdef CONFIG_SECURITY
1335 const struct xattr_handler **s_xattr;
1337 struct list_head s_inodes; /* all inodes */
1338 struct hlist_bl_head s_anon; /* anonymous dentries for (nfs) exporting */
1340 struct list_head __percpu *s_files;
1342 struct list_head s_files;
1344 struct list_head s_mounts; /* list of mounts; _not_ for fs use */
1345 struct block_device *s_bdev;
1346 struct backing_dev_info *s_bdi;
1347 struct mtd_info *s_mtd;
1348 struct hlist_node s_instances;
1349 struct quota_info s_dquot; /* Diskquota specific options */
1351 struct sb_writers s_writers;
1353 char s_id[32]; /* Informational name */
1354 u8 s_uuid[16]; /* UUID */
1356 void *s_fs_info; /* Filesystem private info */
1357 unsigned int s_max_links;
1360 /* Granularity of c/m/atime in ns.
1361 Cannot be worse than a second */
1365 * The next field is for VFS *only*. No filesystems have any business
1366 * even looking at it. You had been warned.
1368 struct mutex s_vfs_rename_mutex; /* Kludge */
1371 * Filesystem subtype. If non-empty the filesystem type field
1372 * in /proc/mounts will be "type.subtype"
1377 * Saved mount options for lazy filesystems using
1378 * generic_show_options()
1380 char __rcu *s_options;
1381 const struct dentry_operations *s_d_op; /* default d_op for dentries */
1384 * Saved pool identifier for cleancache (-1 means none)
1386 int cleancache_poolid;
1388 struct shrinker s_shrink; /* per-sb shrinker handle */
1390 /* Number of inodes with nlink == 0 but still referenced */
1391 atomic_long_t s_remove_count;
1393 /* Being remounted read-only */
1394 int s_readonly_remount;
1396 /* AIO completions deferred from interrupt context */
1397 struct workqueue_struct *s_dio_done_wq;
1400 * Keep the lru lists last in the structure so they always sit on their
1401 * own individual cachelines.
1403 struct list_lru s_dentry_lru ____cacheline_aligned_in_smp;
1404 struct list_lru s_inode_lru ____cacheline_aligned_in_smp;
1407 extern struct timespec current_fs_time(struct super_block *sb);
1410 * Snapshotting support.
1413 void __sb_end_write(struct super_block *sb, int level);
1414 int __sb_start_write(struct super_block *sb, int level, bool wait);
1417 * sb_end_write - drop write access to a superblock
1418 * @sb: the super we wrote to
1420 * Decrement number of writers to the filesystem. Wake up possible waiters
1421 * wanting to freeze the filesystem.
1423 static inline void sb_end_write(struct super_block *sb)
1425 __sb_end_write(sb, SB_FREEZE_WRITE);
1429 * sb_end_pagefault - drop write access to a superblock from a page fault
1430 * @sb: the super we wrote to
1432 * Decrement number of processes handling write page fault to the filesystem.
1433 * Wake up possible waiters wanting to freeze the filesystem.
1435 static inline void sb_end_pagefault(struct super_block *sb)
1437 __sb_end_write(sb, SB_FREEZE_PAGEFAULT);
1441 * sb_end_intwrite - drop write access to a superblock for internal fs purposes
1442 * @sb: the super we wrote to
1444 * Decrement fs-internal number of writers to the filesystem. Wake up possible
1445 * waiters wanting to freeze the filesystem.
1447 static inline void sb_end_intwrite(struct super_block *sb)
1449 __sb_end_write(sb, SB_FREEZE_FS);
1453 * sb_start_write - get write access to a superblock
1454 * @sb: the super we write to
1456 * When a process wants to write data or metadata to a file system (i.e. dirty
1457 * a page or an inode), it should embed the operation in a sb_start_write() -
1458 * sb_end_write() pair to get exclusion against file system freezing. This
1459 * function increments number of writers preventing freezing. If the file
1460 * system is already frozen, the function waits until the file system is
1463 * Since freeze protection behaves as a lock, users have to preserve
1464 * ordering of freeze protection and other filesystem locks. Generally,
1465 * freeze protection should be the outermost lock. In particular, we have:
1468 * -> i_mutex (write path, truncate, directory ops, ...)
1469 * -> s_umount (freeze_super, thaw_super)
1471 static inline void sb_start_write(struct super_block *sb)
1473 __sb_start_write(sb, SB_FREEZE_WRITE, true);
1476 static inline int sb_start_write_trylock(struct super_block *sb)
1478 return __sb_start_write(sb, SB_FREEZE_WRITE, false);
1482 * sb_start_pagefault - get write access to a superblock from a page fault
1483 * @sb: the super we write to
1485 * When a process starts handling write page fault, it should embed the
1486 * operation into sb_start_pagefault() - sb_end_pagefault() pair to get
1487 * exclusion against file system freezing. This is needed since the page fault
1488 * is going to dirty a page. This function increments number of running page
1489 * faults preventing freezing. If the file system is already frozen, the
1490 * function waits until the file system is thawed.
1492 * Since page fault freeze protection behaves as a lock, users have to preserve
1493 * ordering of freeze protection and other filesystem locks. It is advised to
1494 * put sb_start_pagefault() close to mmap_sem in lock ordering. Page fault
1495 * handling code implies lock dependency:
1498 * -> sb_start_pagefault
1500 static inline void sb_start_pagefault(struct super_block *sb)
1502 __sb_start_write(sb, SB_FREEZE_PAGEFAULT, true);
1506 * sb_start_intwrite - get write access to a superblock for internal fs purposes
1507 * @sb: the super we write to
1509 * This is the third level of protection against filesystem freezing. It is
1510 * free for use by a filesystem. The only requirement is that it must rank
1511 * below sb_start_pagefault.
1513 * For example filesystem can call sb_start_intwrite() when starting a
1514 * transaction which somewhat eases handling of freezing for internal sources
1515 * of filesystem changes (internal fs threads, discarding preallocation on file
1518 static inline void sb_start_intwrite(struct super_block *sb)
1520 __sb_start_write(sb, SB_FREEZE_FS, true);
1524 extern bool inode_owner_or_capable(const struct inode *inode);
1527 * VFS helper functions..
1529 extern int vfs_create(struct inode *, struct dentry *, umode_t, bool);
1530 extern int vfs_mkdir(struct inode *, struct dentry *, umode_t);
1531 extern int vfs_mknod(struct inode *, struct dentry *, umode_t, dev_t);
1532 extern int vfs_symlink(struct inode *, struct dentry *, const char *);
1533 extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
1534 extern int vfs_rmdir(struct inode *, struct dentry *);
1535 extern int vfs_unlink(struct inode *, struct dentry *);
1536 extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
1539 * VFS dentry helper functions.
1541 extern void dentry_unhash(struct dentry *dentry);
1544 * VFS file helper functions.
1546 extern void inode_init_owner(struct inode *inode, const struct inode *dir,
1549 * VFS FS_IOC_FIEMAP helper definitions.
1551 struct fiemap_extent_info {
1552 unsigned int fi_flags; /* Flags as passed from user */
1553 unsigned int fi_extents_mapped; /* Number of mapped extents */
1554 unsigned int fi_extents_max; /* Size of fiemap_extent array */
1555 struct fiemap_extent __user *fi_extents_start; /* Start of
1556 fiemap_extent array */
1558 int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
1559 u64 phys, u64 len, u32 flags);
1560 int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
1565 * NOTE! These match bits 12..15 of stat.st_mode
1566 * (ie "(i_mode >> 12) & 15").
1568 #define DT_UNKNOWN 0
1579 * This is the "filldir" function type, used by readdir() to let
1580 * the kernel specify what kind of dirent layout it wants to have.
1581 * This allows the kernel to read directories into kernel space or
1582 * to have different dirent layouts depending on the binary type.
1584 typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned);
1585 struct dir_context {
1586 const filldir_t actor;
1590 struct block_device_operations;
1592 /* These macros are for out of kernel modules to test that
1593 * the kernel supports the unlocked_ioctl and compat_ioctl
1594 * fields in struct file_operations. */
1595 #define HAVE_COMPAT_IOCTL 1
1596 #define HAVE_UNLOCKED_IOCTL 1
1598 struct file_operations {
1599 struct module *owner;
1600 loff_t (*llseek) (struct file *, loff_t, int);
1601 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1602 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
1603 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1604 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1605 int (*iterate) (struct file *, struct dir_context *);
1606 unsigned int (*poll) (struct file *, struct poll_table_struct *);
1607 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1608 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1609 int (*mmap) (struct file *, struct vm_area_struct *);
1610 int (*open) (struct inode *, struct file *);
1611 int (*flush) (struct file *, fl_owner_t id);
1612 int (*release) (struct inode *, struct file *);
1613 int (*fsync) (struct file *, loff_t, loff_t, int datasync);
1614 int (*aio_fsync) (struct kiocb *, int datasync);
1615 int (*fasync) (int, struct file *, int);
1616 int (*lock) (struct file *, int, struct file_lock *);
1617 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1618 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1619 int (*check_flags)(int);
1620 int (*flock) (struct file *, int, struct file_lock *);
1621 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1622 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
1623 int (*setlease)(struct file *, long, struct file_lock **);
1624 long (*fallocate)(struct file *file, int mode, loff_t offset,
1626 int (*show_fdinfo)(struct seq_file *m, struct file *f);
1629 struct inode_operations {
1630 struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int);
1631 void * (*follow_link) (struct dentry *, struct nameidata *);
1632 int (*permission) (struct inode *, int);
1633 struct posix_acl * (*get_acl)(struct inode *, int);
1635 int (*readlink) (struct dentry *, char __user *,int);
1636 void (*put_link) (struct dentry *, struct nameidata *, void *);
1638 int (*create) (struct inode *,struct dentry *, umode_t, bool);
1639 int (*link) (struct dentry *,struct inode *,struct dentry *);
1640 int (*unlink) (struct inode *,struct dentry *);
1641 int (*symlink) (struct inode *,struct dentry *,const char *);
1642 int (*mkdir) (struct inode *,struct dentry *,umode_t);
1643 int (*rmdir) (struct inode *,struct dentry *);
1644 int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t);
1645 int (*rename) (struct inode *, struct dentry *,
1646 struct inode *, struct dentry *);
1647 int (*setattr) (struct dentry *, struct iattr *);
1648 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1649 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1650 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1651 ssize_t (*listxattr) (struct dentry *, char *, size_t);
1652 int (*removexattr) (struct dentry *, const char *);
1653 int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
1655 int (*update_time)(struct inode *, struct timespec *, int);
1656 int (*atomic_open)(struct inode *, struct dentry *,
1657 struct file *, unsigned open_flag,
1658 umode_t create_mode, int *opened);
1659 int (*tmpfile) (struct inode *, struct dentry *, umode_t);
1660 } ____cacheline_aligned;
1662 ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
1663 unsigned long nr_segs, unsigned long fast_segs,
1664 struct iovec *fast_pointer,
1665 struct iovec **ret_pointer);
1667 extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1668 extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1669 extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1670 unsigned long, loff_t *);
1671 extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1672 unsigned long, loff_t *);
1674 struct super_operations {
1675 struct inode *(*alloc_inode)(struct super_block *sb);
1676 void (*destroy_inode)(struct inode *);
1678 void (*dirty_inode) (struct inode *, int flags);
1679 int (*write_inode) (struct inode *, struct writeback_control *wbc);
1680 int (*drop_inode) (struct inode *);
1681 void (*evict_inode) (struct inode *);
1682 void (*put_super) (struct super_block *);
1683 int (*sync_fs)(struct super_block *sb, int wait);
1684 int (*freeze_fs) (struct super_block *);
1685 int (*unfreeze_fs) (struct super_block *);
1686 int (*statfs) (struct dentry *, struct kstatfs *);
1687 int (*remount_fs) (struct super_block *, int *, char *);
1688 void (*umount_begin) (struct super_block *);
1690 int (*show_options)(struct seq_file *, struct dentry *);
1691 int (*show_devname)(struct seq_file *, struct dentry *);
1692 int (*show_path)(struct seq_file *, struct dentry *);
1693 int (*show_stats)(struct seq_file *, struct dentry *);
1695 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1696 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
1698 int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
1699 long (*nr_cached_objects)(struct super_block *, int);
1700 long (*free_cached_objects)(struct super_block *, long, int);
1704 * Inode flags - they have no relation to superblock flags now
1706 #define S_SYNC 1 /* Writes are synced at once */
1707 #define S_NOATIME 2 /* Do not update access times */
1708 #define S_APPEND 4 /* Append-only file */
1709 #define S_IMMUTABLE 8 /* Immutable file */
1710 #define S_DEAD 16 /* removed, but still open directory */
1711 #define S_NOQUOTA 32 /* Inode is not counted to quota */
1712 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
1713 #define S_NOCMTIME 128 /* Do not update file c/mtime */
1714 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
1715 #define S_PRIVATE 512 /* Inode is fs-internal */
1716 #define S_IMA 1024 /* Inode has an associated IMA struct */
1717 #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
1718 #define S_NOSEC 4096 /* no suid or xattr security attributes */
1721 * Note that nosuid etc flags are inode-specific: setting some file-system
1722 * flags just means all the inodes inherit those flags by default. It might be
1723 * possible to override it selectively if you really wanted to with some
1724 * ioctl() that is not currently implemented.
1726 * Exception: MS_RDONLY is always applied to the entire file system.
1728 * Unfortunately, it is possible to change a filesystems flags with it mounted
1729 * with files in use. This means that all of the inodes will not have their
1730 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
1731 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
1733 #define __IS_FLG(inode, flg) ((inode)->i_sb->s_flags & (flg))
1735 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
1736 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
1737 ((inode)->i_flags & S_SYNC))
1738 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
1739 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
1740 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
1741 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
1742 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
1744 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
1745 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
1746 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
1747 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
1749 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
1750 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
1751 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
1752 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
1753 #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
1754 #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
1755 #define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
1758 * Inode state bits. Protected by inode->i_lock
1760 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1761 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1763 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1764 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1765 * various stages of removing an inode.
1767 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1769 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1770 * fdatasync(). i_atime is the usual cause.
1771 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1772 * these changes separately from I_DIRTY_SYNC so that we
1773 * don't have to write inode on fdatasync() when only
1774 * mtime has changed in it.
1775 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1776 * I_NEW Serves as both a mutex and completion notification.
1777 * New inodes set I_NEW. If two processes both create
1778 * the same inode, one of them will release its inode and
1779 * wait for I_NEW to be released before returning.
1780 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1781 * also cause waiting on I_NEW, without I_NEW actually
1782 * being set. find_inode() uses this to prevent returning
1783 * nearly-dead inodes.
1784 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1785 * is zero. I_FREEING must be set when I_WILL_FREE is
1787 * I_FREEING Set when inode is about to be freed but still has dirty
1788 * pages or buffers attached or the inode itself is still
1790 * I_CLEAR Added by clear_inode(). In this state the inode is
1791 * clean and can be destroyed. Inode keeps I_FREEING.
1793 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1794 * prohibited for many purposes. iget() must wait for
1795 * the inode to be completely released, then create it
1796 * anew. Other functions will just ignore such inodes,
1797 * if appropriate. I_NEW is used for waiting.
1799 * I_SYNC Writeback of inode is running. The bit is set during
1800 * data writeback, and cleared with a wakeup on the bit
1801 * address once it is done. The bit is also used to pin
1802 * the inode in memory for flusher thread.
1804 * I_REFERENCED Marks the inode as recently references on the LRU list.
1806 * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit().
1808 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1810 #define I_DIRTY_SYNC (1 << 0)
1811 #define I_DIRTY_DATASYNC (1 << 1)
1812 #define I_DIRTY_PAGES (1 << 2)
1814 #define I_NEW (1 << __I_NEW)
1815 #define I_WILL_FREE (1 << 4)
1816 #define I_FREEING (1 << 5)
1817 #define I_CLEAR (1 << 6)
1819 #define I_SYNC (1 << __I_SYNC)
1820 #define I_REFERENCED (1 << 8)
1821 #define __I_DIO_WAKEUP 9
1822 #define I_DIO_WAKEUP (1 << I_DIO_WAKEUP)
1823 #define I_LINKABLE (1 << 10)
1825 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1827 extern void __mark_inode_dirty(struct inode *, int);
1828 static inline void mark_inode_dirty(struct inode *inode)
1830 __mark_inode_dirty(inode, I_DIRTY);
1833 static inline void mark_inode_dirty_sync(struct inode *inode)
1835 __mark_inode_dirty(inode, I_DIRTY_SYNC);
1838 extern void inc_nlink(struct inode *inode);
1839 extern void drop_nlink(struct inode *inode);
1840 extern void clear_nlink(struct inode *inode);
1841 extern void set_nlink(struct inode *inode, unsigned int nlink);
1843 static inline void inode_inc_link_count(struct inode *inode)
1846 mark_inode_dirty(inode);
1849 static inline void inode_dec_link_count(struct inode *inode)
1852 mark_inode_dirty(inode);
1856 * inode_inc_iversion - increments i_version
1857 * @inode: inode that need to be updated
1859 * Every time the inode is modified, the i_version field will be incremented.
1860 * The filesystem has to be mounted with i_version flag
1863 static inline void inode_inc_iversion(struct inode *inode)
1865 spin_lock(&inode->i_lock);
1867 spin_unlock(&inode->i_lock);
1870 enum file_time_flags {
1877 extern void touch_atime(const struct path *);
1878 static inline void file_accessed(struct file *file)
1880 if (!(file->f_flags & O_NOATIME))
1881 touch_atime(&file->f_path);
1884 int sync_inode(struct inode *inode, struct writeback_control *wbc);
1885 int sync_inode_metadata(struct inode *inode, int wait);
1887 struct file_system_type {
1890 #define FS_REQUIRES_DEV 1
1891 #define FS_BINARY_MOUNTDATA 2
1892 #define FS_HAS_SUBTYPE 4
1893 #define FS_USERNS_MOUNT 8 /* Can be mounted by userns root */
1894 #define FS_USERNS_DEV_MOUNT 16 /* A userns mount does not imply MNT_NODEV */
1895 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() during rename() internally. */
1896 struct dentry *(*mount) (struct file_system_type *, int,
1897 const char *, void *);
1898 void (*kill_sb) (struct super_block *);
1899 struct module *owner;
1900 struct file_system_type * next;
1901 struct hlist_head fs_supers;
1903 struct lock_class_key s_lock_key;
1904 struct lock_class_key s_umount_key;
1905 struct lock_class_key s_vfs_rename_key;
1906 struct lock_class_key s_writers_key[SB_FREEZE_LEVELS];
1908 struct lock_class_key i_lock_key;
1909 struct lock_class_key i_mutex_key;
1910 struct lock_class_key i_mutex_dir_key;
1913 #define MODULE_ALIAS_FS(NAME) MODULE_ALIAS("fs-" NAME)
1915 extern struct dentry *mount_ns(struct file_system_type *fs_type, int flags,
1916 void *data, int (*fill_super)(struct super_block *, void *, int));
1917 extern struct dentry *mount_bdev(struct file_system_type *fs_type,
1918 int flags, const char *dev_name, void *data,
1919 int (*fill_super)(struct super_block *, void *, int));
1920 extern struct dentry *mount_single(struct file_system_type *fs_type,
1921 int flags, void *data,
1922 int (*fill_super)(struct super_block *, void *, int));
1923 extern struct dentry *mount_nodev(struct file_system_type *fs_type,
1924 int flags, void *data,
1925 int (*fill_super)(struct super_block *, void *, int));
1926 extern struct dentry *mount_subtree(struct vfsmount *mnt, const char *path);
1927 void generic_shutdown_super(struct super_block *sb);
1928 void kill_block_super(struct super_block *sb);
1929 void kill_anon_super(struct super_block *sb);
1930 void kill_litter_super(struct super_block *sb);
1931 void deactivate_super(struct super_block *sb);
1932 void deactivate_locked_super(struct super_block *sb);
1933 int set_anon_super(struct super_block *s, void *data);
1934 int get_anon_bdev(dev_t *);
1935 void free_anon_bdev(dev_t);
1936 struct super_block *sget(struct file_system_type *type,
1937 int (*test)(struct super_block *,void *),
1938 int (*set)(struct super_block *,void *),
1939 int flags, void *data);
1940 extern struct dentry *mount_pseudo(struct file_system_type *, char *,
1941 const struct super_operations *ops,
1942 const struct dentry_operations *dops,
1945 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1946 #define fops_get(fops) \
1947 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1948 #define fops_put(fops) \
1949 do { if (fops) module_put((fops)->owner); } while(0)
1951 extern int register_filesystem(struct file_system_type *);
1952 extern int unregister_filesystem(struct file_system_type *);
1953 extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
1954 #define kern_mount(type) kern_mount_data(type, NULL)
1955 extern void kern_unmount(struct vfsmount *mnt);
1956 extern int may_umount_tree(struct vfsmount *);
1957 extern int may_umount(struct vfsmount *);
1958 extern long do_mount(const char *, const char *, const char *, unsigned long, void *);
1959 extern struct vfsmount *collect_mounts(struct path *);
1960 extern void drop_collected_mounts(struct vfsmount *);
1961 extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
1963 extern int vfs_statfs(struct path *, struct kstatfs *);
1964 extern int user_statfs(const char __user *, struct kstatfs *);
1965 extern int fd_statfs(int, struct kstatfs *);
1966 extern int vfs_ustat(dev_t, struct kstatfs *);
1967 extern int freeze_super(struct super_block *super);
1968 extern int thaw_super(struct super_block *super);
1969 extern bool our_mnt(struct vfsmount *mnt);
1970 extern bool fs_fully_visible(struct file_system_type *);
1972 extern int current_umask(void);
1975 extern struct kobject *fs_kobj;
1977 #define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
1979 #define FLOCK_VERIFY_READ 1
1980 #define FLOCK_VERIFY_WRITE 2
1982 #ifdef CONFIG_FILE_LOCKING
1983 extern int locks_mandatory_locked(struct inode *);
1984 extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
1987 * Candidates for mandatory locking have the setgid bit set
1988 * but no group execute bit - an otherwise meaningless combination.
1991 static inline int __mandatory_lock(struct inode *ino)
1993 return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
1997 * ... and these candidates should be on MS_MANDLOCK mounted fs,
1998 * otherwise these will be advisory locks
2001 static inline int mandatory_lock(struct inode *ino)
2003 return IS_MANDLOCK(ino) && __mandatory_lock(ino);
2006 static inline int locks_verify_locked(struct inode *inode)
2008 if (mandatory_lock(inode))
2009 return locks_mandatory_locked(inode);
2013 static inline int locks_verify_truncate(struct inode *inode,
2017 if (inode->i_flock && mandatory_lock(inode))
2018 return locks_mandatory_area(
2019 FLOCK_VERIFY_WRITE, inode, filp,
2020 size < inode->i_size ? size : inode->i_size,
2021 (size < inode->i_size ? inode->i_size - size
2022 : size - inode->i_size)
2027 static inline int break_lease(struct inode *inode, unsigned int mode)
2030 return __break_lease(inode, mode);
2033 #else /* !CONFIG_FILE_LOCKING */
2034 static inline int locks_mandatory_locked(struct inode *inode)
2039 static inline int locks_mandatory_area(int rw, struct inode *inode,
2040 struct file *filp, loff_t offset,
2046 static inline int __mandatory_lock(struct inode *inode)
2051 static inline int mandatory_lock(struct inode *inode)
2056 static inline int locks_verify_locked(struct inode *inode)
2061 static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
2067 static inline int break_lease(struct inode *inode, unsigned int mode)
2072 #endif /* CONFIG_FILE_LOCKING */
2077 const char *name; /* pointer to actual string */
2078 const __user char *uptr; /* original userland pointer */
2079 struct audit_names *aname;
2080 bool separate; /* should "name" be freed? */
2083 extern long vfs_truncate(struct path *, loff_t);
2084 extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
2086 extern int do_fallocate(struct file *file, int mode, loff_t offset,
2088 extern long do_sys_open(int dfd, const char __user *filename, int flags,
2090 extern struct file *file_open_name(struct filename *, int, umode_t);
2091 extern struct file *filp_open(const char *, int, umode_t);
2092 extern struct file *file_open_root(struct dentry *, struct vfsmount *,
2094 extern struct file * dentry_open(const struct path *, int, const struct cred *);
2095 extern int filp_close(struct file *, fl_owner_t id);
2097 extern struct filename *getname(const char __user *);
2103 extern int finish_open(struct file *file, struct dentry *dentry,
2104 int (*open)(struct inode *, struct file *),
2106 extern int finish_no_open(struct file *file, struct dentry *dentry);
2110 extern int ioctl_preallocate(struct file *filp, void __user *argp);
2113 extern void __init vfs_caches_init_early(void);
2114 extern void __init vfs_caches_init(unsigned long);
2116 extern struct kmem_cache *names_cachep;
2118 extern void final_putname(struct filename *name);
2120 #define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL)
2121 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
2122 #ifndef CONFIG_AUDITSYSCALL
2123 #define putname(name) final_putname(name)
2125 extern void putname(struct filename *name);
2129 extern int register_blkdev(unsigned int, const char *);
2130 extern void unregister_blkdev(unsigned int, const char *);
2131 extern struct block_device *bdget(dev_t);
2132 extern struct block_device *bdgrab(struct block_device *bdev);
2133 extern void bd_set_size(struct block_device *, loff_t size);
2134 extern void bd_forget(struct inode *inode);
2135 extern void bdput(struct block_device *);
2136 extern void invalidate_bdev(struct block_device *);
2137 extern void iterate_bdevs(void (*)(struct block_device *, void *), void *);
2138 extern int sync_blockdev(struct block_device *bdev);
2139 extern void kill_bdev(struct block_device *);
2140 extern struct super_block *freeze_bdev(struct block_device *);
2141 extern void emergency_thaw_all(void);
2142 extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
2143 extern int fsync_bdev(struct block_device *);
2144 extern int sb_is_blkdev_sb(struct super_block *sb);
2146 static inline void bd_forget(struct inode *inode) {}
2147 static inline int sync_blockdev(struct block_device *bdev) { return 0; }
2148 static inline void kill_bdev(struct block_device *bdev) {}
2149 static inline void invalidate_bdev(struct block_device *bdev) {}
2151 static inline struct super_block *freeze_bdev(struct block_device *sb)
2156 static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
2161 static inline void iterate_bdevs(void (*f)(struct block_device *, void *), void *arg)
2165 static inline int sb_is_blkdev_sb(struct super_block *sb)
2170 extern int sync_filesystem(struct super_block *);
2171 extern const struct file_operations def_blk_fops;
2172 extern const struct file_operations def_chr_fops;
2173 extern const struct file_operations bad_sock_fops;
2175 extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
2176 extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
2177 extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
2178 extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
2179 extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
2181 extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
2183 extern void blkdev_put(struct block_device *bdev, fmode_t mode);
2185 extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
2186 extern void bd_unlink_disk_holder(struct block_device *bdev,
2187 struct gendisk *disk);
2189 static inline int bd_link_disk_holder(struct block_device *bdev,
2190 struct gendisk *disk)
2194 static inline void bd_unlink_disk_holder(struct block_device *bdev,
2195 struct gendisk *disk)
2202 #define CHRDEV_MAJOR_HASH_SIZE 255
2203 extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
2204 extern int register_chrdev_region(dev_t, unsigned, const char *);
2205 extern int __register_chrdev(unsigned int major, unsigned int baseminor,
2206 unsigned int count, const char *name,
2207 const struct file_operations *fops);
2208 extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
2209 unsigned int count, const char *name);
2210 extern void unregister_chrdev_region(dev_t, unsigned);
2211 extern void chrdev_show(struct seq_file *,off_t);
2213 static inline int register_chrdev(unsigned int major, const char *name,
2214 const struct file_operations *fops)
2216 return __register_chrdev(major, 0, 256, name, fops);
2219 static inline void unregister_chrdev(unsigned int major, const char *name)
2221 __unregister_chrdev(major, 0, 256, name);
2224 /* fs/block_dev.c */
2225 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
2226 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
2229 #define BLKDEV_MAJOR_HASH_SIZE 255
2230 extern const char *__bdevname(dev_t, char *buffer);
2231 extern const char *bdevname(struct block_device *bdev, char *buffer);
2232 extern struct block_device *lookup_bdev(const char *);
2233 extern void blkdev_show(struct seq_file *,off_t);
2236 #define BLKDEV_MAJOR_HASH_SIZE 0
2239 extern void init_special_inode(struct inode *, umode_t, dev_t);
2241 /* Invalid inode operations -- fs/bad_inode.c */
2242 extern void make_bad_inode(struct inode *);
2243 extern int is_bad_inode(struct inode *);
2247 * return READ, READA, or WRITE
2249 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2252 * return data direction, READ or WRITE
2254 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
2256 extern void check_disk_size_change(struct gendisk *disk,
2257 struct block_device *bdev);
2258 extern int revalidate_disk(struct gendisk *);
2259 extern int check_disk_change(struct block_device *);
2260 extern int __invalidate_device(struct block_device *, bool);
2261 extern int invalidate_partition(struct gendisk *, int);
2263 unsigned long invalidate_mapping_pages(struct address_space *mapping,
2264 pgoff_t start, pgoff_t end);
2266 static inline void invalidate_remote_inode(struct inode *inode)
2268 if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
2269 S_ISLNK(inode->i_mode))
2270 invalidate_mapping_pages(inode->i_mapping, 0, -1);
2272 extern int invalidate_inode_pages2(struct address_space *mapping);
2273 extern int invalidate_inode_pages2_range(struct address_space *mapping,
2274 pgoff_t start, pgoff_t end);
2275 extern int write_inode_now(struct inode *, int);
2276 extern int filemap_fdatawrite(struct address_space *);
2277 extern int filemap_flush(struct address_space *);
2278 extern int filemap_fdatawait(struct address_space *);
2279 extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
2281 extern int filemap_write_and_wait(struct address_space *mapping);
2282 extern int filemap_write_and_wait_range(struct address_space *mapping,
2283 loff_t lstart, loff_t lend);
2284 extern int __filemap_fdatawrite_range(struct address_space *mapping,
2285 loff_t start, loff_t end, int sync_mode);
2286 extern int filemap_fdatawrite_range(struct address_space *mapping,
2287 loff_t start, loff_t end);
2289 extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
2291 extern int vfs_fsync(struct file *file, int datasync);
2292 extern int generic_write_sync(struct file *file, loff_t pos, loff_t count);
2293 extern void emergency_sync(void);
2294 extern void emergency_remount(void);
2296 extern sector_t bmap(struct inode *, sector_t);
2298 extern int notify_change(struct dentry *, struct iattr *);
2299 extern int inode_permission(struct inode *, int);
2300 extern int generic_permission(struct inode *, int);
2302 static inline bool execute_ok(struct inode *inode)
2304 return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
2307 static inline struct inode *file_inode(struct file *f)
2312 static inline void file_start_write(struct file *file)
2314 if (!S_ISREG(file_inode(file)->i_mode))
2316 __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, true);
2319 static inline bool file_start_write_trylock(struct file *file)
2321 if (!S_ISREG(file_inode(file)->i_mode))
2323 return __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, false);
2326 static inline void file_end_write(struct file *file)
2328 if (!S_ISREG(file_inode(file)->i_mode))
2330 __sb_end_write(file_inode(file)->i_sb, SB_FREEZE_WRITE);
2334 * get_write_access() gets write permission for a file.
2335 * put_write_access() releases this write permission.
2336 * This is used for regular files.
2337 * We cannot support write (and maybe mmap read-write shared) accesses and
2338 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2339 * can have the following values:
2340 * 0: no writers, no VM_DENYWRITE mappings
2341 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2342 * > 0: (i_writecount) users are writing to the file.
2344 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2345 * except for the cases where we don't hold i_writecount yet. Then we need to
2346 * use {get,deny}_write_access() - these functions check the sign and refuse
2347 * to do the change if sign is wrong.
2349 static inline int get_write_access(struct inode *inode)
2351 return atomic_inc_unless_negative(&inode->i_writecount) ? 0 : -ETXTBSY;
2353 static inline int deny_write_access(struct file *file)
2355 struct inode *inode = file_inode(file);
2356 return atomic_dec_unless_positive(&inode->i_writecount) ? 0 : -ETXTBSY;
2358 static inline void put_write_access(struct inode * inode)
2360 atomic_dec(&inode->i_writecount);
2362 static inline void allow_write_access(struct file *file)
2365 atomic_inc(&file_inode(file)->i_writecount);
2368 static inline void i_readcount_dec(struct inode *inode)
2370 BUG_ON(!atomic_read(&inode->i_readcount));
2371 atomic_dec(&inode->i_readcount);
2373 static inline void i_readcount_inc(struct inode *inode)
2375 atomic_inc(&inode->i_readcount);
2378 static inline void i_readcount_dec(struct inode *inode)
2382 static inline void i_readcount_inc(struct inode *inode)
2387 extern int do_pipe_flags(int *, int);
2389 extern int kernel_read(struct file *, loff_t, char *, unsigned long);
2390 extern ssize_t kernel_write(struct file *, const char *, size_t, loff_t);
2391 extern struct file * open_exec(const char *);
2393 /* fs/dcache.c -- generic fs support functions */
2394 extern int is_subdir(struct dentry *, struct dentry *);
2395 extern int path_is_under(struct path *, struct path *);
2397 #include <linux/err.h>
2399 /* needed for stackable file system support */
2400 extern loff_t default_llseek(struct file *file, loff_t offset, int whence);
2402 extern loff_t vfs_llseek(struct file *file, loff_t offset, int whence);
2404 extern int inode_init_always(struct super_block *, struct inode *);
2405 extern void inode_init_once(struct inode *);
2406 extern void address_space_init_once(struct address_space *mapping);
2407 extern void ihold(struct inode * inode);
2408 extern void iput(struct inode *);
2409 extern struct inode * igrab(struct inode *);
2410 extern ino_t iunique(struct super_block *, ino_t);
2411 extern int inode_needs_sync(struct inode *inode);
2412 extern int generic_delete_inode(struct inode *inode);
2413 static inline int generic_drop_inode(struct inode *inode)
2415 return !inode->i_nlink || inode_unhashed(inode);
2418 extern struct inode *ilookup5_nowait(struct super_block *sb,
2419 unsigned long hashval, int (*test)(struct inode *, void *),
2421 extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
2422 int (*test)(struct inode *, void *), void *data);
2423 extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
2425 extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
2426 extern struct inode * iget_locked(struct super_block *, unsigned long);
2427 extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
2428 extern int insert_inode_locked(struct inode *);
2429 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2430 extern void lockdep_annotate_inode_mutex_key(struct inode *inode);
2432 static inline void lockdep_annotate_inode_mutex_key(struct inode *inode) { };
2434 extern void unlock_new_inode(struct inode *);
2435 extern unsigned int get_next_ino(void);
2437 extern void __iget(struct inode * inode);
2438 extern void iget_failed(struct inode *);
2439 extern void clear_inode(struct inode *);
2440 extern void __destroy_inode(struct inode *);
2441 extern struct inode *new_inode_pseudo(struct super_block *sb);
2442 extern struct inode *new_inode(struct super_block *sb);
2443 extern void free_inode_nonrcu(struct inode *inode);
2444 extern int should_remove_suid(struct dentry *);
2445 extern int file_remove_suid(struct file *);
2447 extern void __insert_inode_hash(struct inode *, unsigned long hashval);
2448 static inline void insert_inode_hash(struct inode *inode)
2450 __insert_inode_hash(inode, inode->i_ino);
2453 extern void __remove_inode_hash(struct inode *);
2454 static inline void remove_inode_hash(struct inode *inode)
2456 if (!inode_unhashed(inode))
2457 __remove_inode_hash(inode);
2460 extern void inode_sb_list_add(struct inode *inode);
2463 extern void submit_bio(int, struct bio *);
2464 extern int bdev_read_only(struct block_device *);
2466 extern int set_blocksize(struct block_device *, int);
2467 extern int sb_set_blocksize(struct super_block *, int);
2468 extern int sb_min_blocksize(struct super_block *, int);
2470 extern int generic_file_mmap(struct file *, struct vm_area_struct *);
2471 extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
2472 extern int generic_file_remap_pages(struct vm_area_struct *, unsigned long addr,
2473 unsigned long size, pgoff_t pgoff);
2474 extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
2475 int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
2476 extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2477 extern ssize_t __generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long,
2479 extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2480 extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
2481 unsigned long *, loff_t, loff_t *, size_t, size_t);
2482 extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
2483 unsigned long, loff_t, loff_t *, size_t, ssize_t);
2484 extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
2485 extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
2486 extern int generic_segment_checks(const struct iovec *iov,
2487 unsigned long *nr_segs, size_t *count, int access_flags);
2489 /* fs/block_dev.c */
2490 extern ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
2491 unsigned long nr_segs, loff_t pos);
2492 extern int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
2494 extern void block_sync_page(struct page *page);
2497 extern ssize_t generic_file_splice_read(struct file *, loff_t *,
2498 struct pipe_inode_info *, size_t, unsigned int);
2499 extern ssize_t default_file_splice_read(struct file *, loff_t *,
2500 struct pipe_inode_info *, size_t, unsigned int);
2501 extern ssize_t generic_file_splice_write(struct pipe_inode_info *,
2502 struct file *, loff_t *, size_t, unsigned int);
2503 extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
2504 struct file *out, loff_t *, size_t len, unsigned int flags);
2507 file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
2508 extern loff_t noop_llseek(struct file *file, loff_t offset, int whence);
2509 extern loff_t no_llseek(struct file *file, loff_t offset, int whence);
2510 extern loff_t vfs_setpos(struct file *file, loff_t offset, loff_t maxsize);
2511 extern loff_t generic_file_llseek(struct file *file, loff_t offset, int whence);
2512 extern loff_t generic_file_llseek_size(struct file *file, loff_t offset,
2513 int whence, loff_t maxsize, loff_t eof);
2514 extern loff_t fixed_size_llseek(struct file *file, loff_t offset,
2515 int whence, loff_t size);
2516 extern int generic_file_open(struct inode * inode, struct file * filp);
2517 extern int nonseekable_open(struct inode * inode, struct file * filp);
2519 #ifdef CONFIG_FS_XIP
2520 extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
2522 extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
2523 extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
2524 size_t len, loff_t *ppos);
2525 extern int xip_truncate_page(struct address_space *mapping, loff_t from);
2527 static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
2534 typedef void (dio_submit_t)(int rw, struct bio *bio, struct inode *inode,
2535 loff_t file_offset);
2538 /* need locking between buffered and direct access */
2541 /* filesystem does not support filling holes */
2542 DIO_SKIP_HOLES = 0x02,
2545 void dio_end_io(struct bio *bio, int error);
2547 ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
2548 struct block_device *bdev, const struct iovec *iov, loff_t offset,
2549 unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
2550 dio_submit_t submit_io, int flags);
2552 static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
2553 struct inode *inode, const struct iovec *iov, loff_t offset,
2554 unsigned long nr_segs, get_block_t get_block)
2556 return __blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
2557 offset, nr_segs, get_block, NULL, NULL,
2558 DIO_LOCKING | DIO_SKIP_HOLES);
2562 void inode_dio_wait(struct inode *inode);
2563 void inode_dio_done(struct inode *inode);
2565 extern const struct file_operations generic_ro_fops;
2567 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2569 extern int vfs_readlink(struct dentry *, char __user *, int, const char *);
2570 extern int page_readlink(struct dentry *, char __user *, int);
2571 extern void *page_follow_link_light(struct dentry *, struct nameidata *);
2572 extern void page_put_link(struct dentry *, struct nameidata *, void *);
2573 extern int __page_symlink(struct inode *inode, const char *symname, int len,
2575 extern int page_symlink(struct inode *inode, const char *symname, int len);
2576 extern const struct inode_operations page_symlink_inode_operations;
2577 extern int generic_readlink(struct dentry *, char __user *, int);
2578 extern void generic_fillattr(struct inode *, struct kstat *);
2579 extern int vfs_getattr(struct path *, struct kstat *);
2580 void __inode_add_bytes(struct inode *inode, loff_t bytes);
2581 void inode_add_bytes(struct inode *inode, loff_t bytes);
2582 void __inode_sub_bytes(struct inode *inode, loff_t bytes);
2583 void inode_sub_bytes(struct inode *inode, loff_t bytes);
2584 loff_t inode_get_bytes(struct inode *inode);
2585 void inode_set_bytes(struct inode *inode, loff_t bytes);
2587 extern int vfs_readdir(struct file *, filldir_t, void *);
2588 extern int iterate_dir(struct file *, struct dir_context *);
2590 extern int vfs_stat(const char __user *, struct kstat *);
2591 extern int vfs_lstat(const char __user *, struct kstat *);
2592 extern int vfs_fstat(unsigned int, struct kstat *);
2593 extern int vfs_fstatat(int , const char __user *, struct kstat *, int);
2595 extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
2597 extern int __generic_block_fiemap(struct inode *inode,
2598 struct fiemap_extent_info *fieinfo,
2599 loff_t start, loff_t len,
2600 get_block_t *get_block);
2601 extern int generic_block_fiemap(struct inode *inode,
2602 struct fiemap_extent_info *fieinfo, u64 start,
2603 u64 len, get_block_t *get_block);
2605 extern void get_filesystem(struct file_system_type *fs);
2606 extern void put_filesystem(struct file_system_type *fs);
2607 extern struct file_system_type *get_fs_type(const char *name);
2608 extern struct super_block *get_super(struct block_device *);
2609 extern struct super_block *get_super_thawed(struct block_device *);
2610 extern struct super_block *get_active_super(struct block_device *bdev);
2611 extern void drop_super(struct super_block *sb);
2612 extern void iterate_supers(void (*)(struct super_block *, void *), void *);
2613 extern void iterate_supers_type(struct file_system_type *,
2614 void (*)(struct super_block *, void *), void *);
2616 extern int dcache_dir_open(struct inode *, struct file *);
2617 extern int dcache_dir_close(struct inode *, struct file *);
2618 extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
2619 extern int dcache_readdir(struct file *, struct dir_context *);
2620 extern int simple_setattr(struct dentry *, struct iattr *);
2621 extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2622 extern int simple_statfs(struct dentry *, struct kstatfs *);
2623 extern int simple_open(struct inode *inode, struct file *file);
2624 extern int simple_link(struct dentry *, struct inode *, struct dentry *);
2625 extern int simple_unlink(struct inode *, struct dentry *);
2626 extern int simple_rmdir(struct inode *, struct dentry *);
2627 extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
2628 extern int noop_fsync(struct file *, loff_t, loff_t, int);
2629 extern int simple_empty(struct dentry *);
2630 extern int simple_readpage(struct file *file, struct page *page);
2631 extern int simple_write_begin(struct file *file, struct address_space *mapping,
2632 loff_t pos, unsigned len, unsigned flags,
2633 struct page **pagep, void **fsdata);
2634 extern int simple_write_end(struct file *file, struct address_space *mapping,
2635 loff_t pos, unsigned len, unsigned copied,
2636 struct page *page, void *fsdata);
2638 extern struct dentry *simple_lookup(struct inode *, struct dentry *, unsigned int flags);
2639 extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
2640 extern const struct file_operations simple_dir_operations;
2641 extern const struct inode_operations simple_dir_inode_operations;
2642 struct tree_descr { char *name; const struct file_operations *ops; int mode; };
2643 struct dentry *d_alloc_name(struct dentry *, const char *);
2644 extern int simple_fill_super(struct super_block *, unsigned long, struct tree_descr *);
2645 extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
2646 extern void simple_release_fs(struct vfsmount **mount, int *count);
2648 extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
2649 loff_t *ppos, const void *from, size_t available);
2650 extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
2651 const void __user *from, size_t count);
2653 extern int generic_file_fsync(struct file *, loff_t, loff_t, int);
2655 extern int generic_check_addressable(unsigned, u64);
2657 #ifdef CONFIG_MIGRATION
2658 extern int buffer_migrate_page(struct address_space *,
2659 struct page *, struct page *,
2662 #define buffer_migrate_page NULL
2665 extern int inode_change_ok(const struct inode *, struct iattr *);
2666 extern int inode_newsize_ok(const struct inode *, loff_t offset);
2667 extern void setattr_copy(struct inode *inode, const struct iattr *attr);
2669 extern int file_update_time(struct file *file);
2671 extern int generic_show_options(struct seq_file *m, struct dentry *root);
2672 extern void save_mount_options(struct super_block *sb, char *options);
2673 extern void replace_mount_options(struct super_block *sb, char *options);
2675 static inline ino_t parent_ino(struct dentry *dentry)
2680 * Don't strictly need d_lock here? If the parent ino could change
2681 * then surely we'd have a deeper race in the caller?
2683 spin_lock(&dentry->d_lock);
2684 res = dentry->d_parent->d_inode->i_ino;
2685 spin_unlock(&dentry->d_lock);
2689 /* Transaction based IO helpers */
2692 * An argresp is stored in an allocated page and holds the
2693 * size of the argument or response, along with its content
2695 struct simple_transaction_argresp {
2700 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2702 char *simple_transaction_get(struct file *file, const char __user *buf,
2704 ssize_t simple_transaction_read(struct file *file, char __user *buf,
2705 size_t size, loff_t *pos);
2706 int simple_transaction_release(struct inode *inode, struct file *file);
2708 void simple_transaction_set(struct file *file, size_t n);
2711 * simple attribute files
2713 * These attributes behave similar to those in sysfs:
2715 * Writing to an attribute immediately sets a value, an open file can be
2716 * written to multiple times.
2718 * Reading from an attribute creates a buffer from the value that might get
2719 * read with multiple read calls. When the attribute has been read
2720 * completely, no further read calls are possible until the file is opened
2723 * All attributes contain a text representation of a numeric value
2724 * that are accessed with the get() and set() functions.
2726 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2727 static int __fops ## _open(struct inode *inode, struct file *file) \
2729 __simple_attr_check_format(__fmt, 0ull); \
2730 return simple_attr_open(inode, file, __get, __set, __fmt); \
2732 static const struct file_operations __fops = { \
2733 .owner = THIS_MODULE, \
2734 .open = __fops ## _open, \
2735 .release = simple_attr_release, \
2736 .read = simple_attr_read, \
2737 .write = simple_attr_write, \
2738 .llseek = generic_file_llseek, \
2741 static inline __printf(1, 2)
2742 void __simple_attr_check_format(const char *fmt, ...)
2744 /* don't do anything, just let the compiler check the arguments; */
2747 int simple_attr_open(struct inode *inode, struct file *file,
2748 int (*get)(void *, u64 *), int (*set)(void *, u64),
2750 int simple_attr_release(struct inode *inode, struct file *file);
2751 ssize_t simple_attr_read(struct file *file, char __user *buf,
2752 size_t len, loff_t *ppos);
2753 ssize_t simple_attr_write(struct file *file, const char __user *buf,
2754 size_t len, loff_t *ppos);
2757 int proc_nr_files(struct ctl_table *table, int write,
2758 void __user *buffer, size_t *lenp, loff_t *ppos);
2759 int proc_nr_dentry(struct ctl_table *table, int write,
2760 void __user *buffer, size_t *lenp, loff_t *ppos);
2761 int proc_nr_inodes(struct ctl_table *table, int write,
2762 void __user *buffer, size_t *lenp, loff_t *ppos);
2763 int __init get_filesystem_list(char *buf);
2765 #define __FMODE_EXEC ((__force int) FMODE_EXEC)
2766 #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
2768 #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
2769 #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
2770 (flag & __FMODE_NONOTIFY)))
2772 static inline int is_sxid(umode_t mode)
2774 return (mode & S_ISUID) || ((mode & S_ISGID) && (mode & S_IXGRP));
2777 static inline void inode_has_no_xattr(struct inode *inode)
2779 if (!is_sxid(inode->i_mode) && (inode->i_sb->s_flags & MS_NOSEC))
2780 inode->i_flags |= S_NOSEC;
2783 static inline bool dir_emit(struct dir_context *ctx,
2784 const char *name, int namelen,
2785 u64 ino, unsigned type)
2787 return ctx->actor(ctx, name, namelen, ctx->pos, ino, type) == 0;
2789 static inline bool dir_emit_dot(struct file *file, struct dir_context *ctx)
2791 return ctx->actor(ctx, ".", 1, ctx->pos,
2792 file->f_path.dentry->d_inode->i_ino, DT_DIR) == 0;
2794 static inline bool dir_emit_dotdot(struct file *file, struct dir_context *ctx)
2796 return ctx->actor(ctx, "..", 2, ctx->pos,
2797 parent_ino(file->f_path.dentry), DT_DIR) == 0;
2799 static inline bool dir_emit_dots(struct file *file, struct dir_context *ctx)
2801 if (ctx->pos == 0) {
2802 if (!dir_emit_dot(file, ctx))
2806 if (ctx->pos == 1) {
2807 if (!dir_emit_dotdot(file, ctx))
2813 static inline bool dir_relax(struct inode *inode)
2815 mutex_unlock(&inode->i_mutex);
2816 mutex_lock(&inode->i_mutex);
2817 return !IS_DEADDIR(inode);
2820 #endif /* _LINUX_FS_H */