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;
53 extern void __init inode_init(void);
54 extern void __init inode_init_early(void);
55 extern void __init files_init(unsigned long);
57 extern struct files_stat_struct files_stat;
58 extern unsigned long get_max_files(void);
59 extern int sysctl_nr_open;
60 extern struct inodes_stat_t inodes_stat;
61 extern int leases_enable, lease_break_time;
62 extern int sysctl_protected_symlinks;
63 extern int sysctl_protected_hardlinks;
66 typedef int (get_block_t)(struct inode *inode, sector_t iblock,
67 struct buffer_head *bh_result, int create);
68 typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
69 ssize_t bytes, void *private);
71 #define MAY_EXEC 0x00000001
72 #define MAY_WRITE 0x00000002
73 #define MAY_READ 0x00000004
74 #define MAY_APPEND 0x00000008
75 #define MAY_ACCESS 0x00000010
76 #define MAY_OPEN 0x00000020
77 #define MAY_CHDIR 0x00000040
78 /* called from RCU mode, don't block */
79 #define MAY_NOT_BLOCK 0x00000080
82 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
83 * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
86 /* file is open for reading */
87 #define FMODE_READ ((__force fmode_t)0x1)
88 /* file is open for writing */
89 #define FMODE_WRITE ((__force fmode_t)0x2)
90 /* file is seekable */
91 #define FMODE_LSEEK ((__force fmode_t)0x4)
92 /* file can be accessed using pread */
93 #define FMODE_PREAD ((__force fmode_t)0x8)
94 /* file can be accessed using pwrite */
95 #define FMODE_PWRITE ((__force fmode_t)0x10)
96 /* File is opened for execution with sys_execve / sys_uselib */
97 #define FMODE_EXEC ((__force fmode_t)0x20)
98 /* File is opened with O_NDELAY (only set for block devices) */
99 #define FMODE_NDELAY ((__force fmode_t)0x40)
100 /* File is opened with O_EXCL (only set for block devices) */
101 #define FMODE_EXCL ((__force fmode_t)0x80)
102 /* File is opened using open(.., 3, ..) and is writeable only for ioctls
103 (specialy hack for floppy.c) */
104 #define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
105 /* 32bit hashes as llseek() offset (for directories) */
106 #define FMODE_32BITHASH ((__force fmode_t)0x200)
107 /* 64bit hashes as llseek() offset (for directories) */
108 #define FMODE_64BITHASH ((__force fmode_t)0x400)
111 * Don't update ctime and mtime.
113 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
114 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
116 #define FMODE_NOCMTIME ((__force fmode_t)0x800)
118 /* Expect random access pattern */
119 #define FMODE_RANDOM ((__force fmode_t)0x1000)
121 /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
122 #define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
124 /* File is opened with O_PATH; almost nothing can be done with it */
125 #define FMODE_PATH ((__force fmode_t)0x4000)
127 /* File needs atomic accesses to f_pos */
128 #define FMODE_ATOMIC_POS ((__force fmode_t)0x8000)
129 /* Write access to underlying fs */
130 #define FMODE_WRITER ((__force fmode_t)0x10000)
131 /* Has read method(s) */
132 #define FMODE_CAN_READ ((__force fmode_t)0x20000)
133 /* Has write method(s) */
134 #define FMODE_CAN_WRITE ((__force fmode_t)0x40000)
136 /* File is stream-like */
137 #define FMODE_STREAM ((__force fmode_t)0x200000)
139 /* File was opened by fanotify and shouldn't generate fanotify events */
140 #define FMODE_NONOTIFY ((__force fmode_t)0x1000000)
143 * Flag for rw_copy_check_uvector and compat_rw_copy_check_uvector
144 * that indicates that they should check the contents of the iovec are
145 * valid, but not check the memory that the iovec elements
148 #define CHECK_IOVEC_ONLY -1
151 * The below are the various read and write types that we support. Some of
152 * them include behavioral modifiers that send information down to the
153 * block layer and IO scheduler. Terminology:
155 * The block layer uses device plugging to defer IO a little bit, in
156 * the hope that we will see more IO very shortly. This increases
157 * coalescing of adjacent IO and thus reduces the number of IOs we
158 * have to send to the device. It also allows for better queuing,
159 * if the IO isn't mergeable. If the caller is going to be waiting
160 * for the IO, then he must ensure that the device is unplugged so
161 * that the IO is dispatched to the driver.
163 * All IO is handled async in Linux. This is fine for background
164 * writes, but for reads or writes that someone waits for completion
165 * on, we want to notify the block layer and IO scheduler so that they
166 * know about it. That allows them to make better scheduling
167 * decisions. So when the below references 'sync' and 'async', it
168 * is referencing this priority hint.
170 * With that in mind, the available types are:
172 * READ A normal read operation. Device will be plugged.
173 * READ_SYNC A synchronous read. Device is not plugged, caller can
174 * immediately wait on this read without caring about
176 * READA Used for read-ahead operations. Lower priority, and the
177 * block layer could (in theory) choose to ignore this
178 * request if it runs into resource problems.
179 * WRITE A normal async write. Device will be plugged.
180 * WRITE_SYNC Synchronous write. Identical to WRITE, but passes down
181 * the hint that someone will be waiting on this IO
182 * shortly. The write equivalent of READ_SYNC.
183 * WRITE_ODIRECT Special case write for O_DIRECT only.
184 * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
185 * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
186 * non-volatile media on completion.
187 * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
188 * by a cache flush and data is guaranteed to be on
189 * non-volatile media on completion.
192 #define RW_MASK REQ_WRITE
193 #define RWA_MASK REQ_RAHEAD
196 #define WRITE RW_MASK
197 #define READA RWA_MASK
198 #define KERNEL_READ (READ|REQ_KERNEL)
199 #define KERNEL_WRITE (WRITE|REQ_KERNEL)
201 #define READ_SYNC (READ | REQ_SYNC)
202 #define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE)
203 #define WRITE_ODIRECT (WRITE | REQ_SYNC)
204 #define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
205 #define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
206 #define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
209 * Attribute flags. These should be or-ed together to figure out what
212 #define ATTR_MODE (1 << 0)
213 #define ATTR_UID (1 << 1)
214 #define ATTR_GID (1 << 2)
215 #define ATTR_SIZE (1 << 3)
216 #define ATTR_ATIME (1 << 4)
217 #define ATTR_MTIME (1 << 5)
218 #define ATTR_CTIME (1 << 6)
219 #define ATTR_ATIME_SET (1 << 7)
220 #define ATTR_MTIME_SET (1 << 8)
221 #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
222 #define ATTR_ATTR_FLAG (1 << 10)
223 #define ATTR_KILL_SUID (1 << 11)
224 #define ATTR_KILL_SGID (1 << 12)
225 #define ATTR_FILE (1 << 13)
226 #define ATTR_KILL_PRIV (1 << 14)
227 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
228 #define ATTR_TIMES_SET (1 << 16)
231 * This is the Inode Attributes structure, used for notify_change(). It
232 * uses the above definitions as flags, to know which values have changed.
233 * Also, in this manner, a Filesystem can look at only the values it cares
234 * about. Basically, these are the attributes that the VFS layer can
235 * request to change from the FS layer.
237 * Derek Atkins <warlord@MIT.EDU> 94-10-20
240 unsigned int ia_valid;
245 struct timespec ia_atime;
246 struct timespec ia_mtime;
247 struct timespec ia_ctime;
250 * Not an attribute, but an auxiliary info for filesystems wanting to
251 * implement an ftruncate() like method. NOTE: filesystem should
252 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
254 struct file *ia_file;
258 * Includes for diskquotas.
260 #include <linux/quota.h>
263 * Maximum number of layers of fs stack. Needs to be limited to
264 * prevent kernel stack overflow
266 #define FILESYSTEM_MAX_STACK_DEPTH 2
269 * enum positive_aop_returns - aop return codes with specific semantics
271 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
272 * completed, that the page is still locked, and
273 * should be considered active. The VM uses this hint
274 * to return the page to the active list -- it won't
275 * be a candidate for writeback again in the near
276 * future. Other callers must be careful to unlock
277 * the page if they get this return. Returned by
280 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
281 * unlocked it and the page might have been truncated.
282 * The caller should back up to acquiring a new page and
283 * trying again. The aop will be taking reasonable
284 * precautions not to livelock. If the caller held a page
285 * reference, it should drop it before retrying. Returned
288 * address_space_operation functions return these large constants to indicate
289 * special semantics to the caller. These are much larger than the bytes in a
290 * page to allow for functions that return the number of bytes operated on in a
294 enum positive_aop_returns {
295 AOP_WRITEPAGE_ACTIVATE = 0x80000,
296 AOP_TRUNCATED_PAGE = 0x80001,
299 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
300 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
301 #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
302 * helper code (eg buffer layer)
303 * to clear GFP_FS from alloc */
306 * oh the beauties of C type declarations.
309 struct address_space;
310 struct writeback_control;
313 * "descriptor" for what we're up to with a read.
314 * This allows us to use the same read code yet
315 * have multiple different users of the data that
316 * we read from a file.
318 * The simplest case just copies the data to user
331 typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
332 unsigned long, unsigned long);
334 struct address_space_operations {
335 int (*writepage)(struct page *page, struct writeback_control *wbc);
336 int (*readpage)(struct file *, struct page *);
338 /* Write back some dirty pages from this mapping. */
339 int (*writepages)(struct address_space *, struct writeback_control *);
341 /* Set a page dirty. Return true if this dirtied it */
342 int (*set_page_dirty)(struct page *page);
344 int (*readpages)(struct file *filp, struct address_space *mapping,
345 struct list_head *pages, unsigned nr_pages);
347 int (*write_begin)(struct file *, struct address_space *mapping,
348 loff_t pos, unsigned len, unsigned flags,
349 struct page **pagep, void **fsdata);
350 int (*write_end)(struct file *, struct address_space *mapping,
351 loff_t pos, unsigned len, unsigned copied,
352 struct page *page, void *fsdata);
354 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
355 sector_t (*bmap)(struct address_space *, sector_t);
356 void (*invalidatepage) (struct page *, unsigned int, unsigned int);
357 int (*releasepage) (struct page *, gfp_t);
358 void (*freepage)(struct page *);
359 ssize_t (*direct_IO)(int, struct kiocb *, struct iov_iter *iter, loff_t offset);
360 int (*get_xip_mem)(struct address_space *, pgoff_t, int,
361 void **, unsigned long *);
363 * migrate the contents of a page to the specified target. If
364 * migrate_mode is MIGRATE_ASYNC, it must not block.
366 int (*migratepage) (struct address_space *,
367 struct page *, struct page *, enum migrate_mode);
368 int (*launder_page) (struct page *);
369 int (*is_partially_uptodate) (struct page *, unsigned long,
371 void (*is_dirty_writeback) (struct page *, bool *, bool *);
372 int (*error_remove_page)(struct address_space *, struct page *);
374 /* swapfile support */
375 int (*swap_activate)(struct swap_info_struct *sis, struct file *file,
377 void (*swap_deactivate)(struct file *file);
380 extern const struct address_space_operations empty_aops;
383 * pagecache_write_begin/pagecache_write_end must be used by general code
384 * to write into the pagecache.
386 int pagecache_write_begin(struct file *, struct address_space *mapping,
387 loff_t pos, unsigned len, unsigned flags,
388 struct page **pagep, void **fsdata);
390 int pagecache_write_end(struct file *, struct address_space *mapping,
391 loff_t pos, unsigned len, unsigned copied,
392 struct page *page, void *fsdata);
394 struct backing_dev_info;
395 struct address_space {
396 struct inode *host; /* owner: inode, block_device */
397 struct radix_tree_root page_tree; /* radix tree of all pages */
398 spinlock_t tree_lock; /* and lock protecting it */
399 unsigned int i_mmap_writable;/* count VM_SHARED mappings */
400 struct rb_root i_mmap; /* tree of private and shared mappings */
401 struct mutex i_mmap_mutex; /* protect tree, count, list */
402 /* Protected by tree_lock together with the radix tree */
403 unsigned long nrpages; /* number of total pages */
404 unsigned long nrshadows; /* number of shadow entries */
405 pgoff_t writeback_index;/* writeback starts here */
406 const struct address_space_operations *a_ops; /* methods */
407 unsigned long flags; /* error bits/gfp mask */
408 struct backing_dev_info *backing_dev_info; /* device readahead, etc */
409 spinlock_t private_lock; /* for use by the address_space */
410 struct list_head private_list; /* ditto */
411 void *private_data; /* ditto */
412 } __attribute__((aligned(sizeof(long))));
414 * On most architectures that alignment is already the case; but
415 * must be enforced here for CRIS, to let the least significant bit
416 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
418 struct request_queue;
420 struct block_device {
421 dev_t bd_dev; /* not a kdev_t - it's a search key */
423 struct inode * bd_inode; /* will die */
424 struct super_block * bd_super;
425 struct mutex bd_mutex; /* open/close mutex */
426 struct list_head bd_inodes;
430 bool bd_write_holder;
432 struct list_head bd_holder_disks;
434 struct block_device * bd_contains;
435 unsigned bd_block_size;
436 struct hd_struct * bd_part;
437 /* number of times partitions within this device have been opened. */
438 unsigned bd_part_count;
440 struct gendisk * bd_disk;
441 struct request_queue * bd_queue;
442 struct list_head bd_list;
444 * Private data. You must have bd_claim'ed the block_device
445 * to use this. NOTE: bd_claim allows an owner to claim
446 * the same device multiple times, the owner must take special
447 * care to not mess up bd_private for that case.
449 unsigned long bd_private;
451 /* The counter of freeze processes */
452 int bd_fsfreeze_count;
453 /* Mutex for freeze */
454 struct mutex bd_fsfreeze_mutex;
458 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
461 #define PAGECACHE_TAG_DIRTY 0
462 #define PAGECACHE_TAG_WRITEBACK 1
463 #define PAGECACHE_TAG_TOWRITE 2
465 int mapping_tagged(struct address_space *mapping, int tag);
468 * Might pages of this file be mapped into userspace?
470 static inline int mapping_mapped(struct address_space *mapping)
472 return !RB_EMPTY_ROOT(&mapping->i_mmap);
476 * Might pages of this file have been modified in userspace?
477 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
478 * marks vma as VM_SHARED if it is shared, and the file was opened for
479 * writing i.e. vma may be mprotected writable even if now readonly.
481 static inline int mapping_writably_mapped(struct address_space *mapping)
483 return mapping->i_mmap_writable != 0;
487 * Use sequence counter to get consistent i_size on 32-bit processors.
489 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
490 #include <linux/seqlock.h>
491 #define __NEED_I_SIZE_ORDERED
492 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
494 #define i_size_ordered_init(inode) do { } while (0)
498 #define ACL_NOT_CACHED ((void *)(-1))
500 #define IOP_FASTPERM 0x0001
501 #define IOP_LOOKUP 0x0002
502 #define IOP_NOFOLLOW 0x0004
505 * Keep mostly read-only and often accessed (especially for
506 * the RCU path lookup and 'stat' data) fields at the beginning
507 * of the 'struct inode'
511 unsigned short i_opflags;
514 unsigned int i_flags;
516 #ifdef CONFIG_FS_POSIX_ACL
517 struct posix_acl *i_acl;
518 struct posix_acl *i_default_acl;
521 const struct inode_operations *i_op;
522 struct super_block *i_sb;
523 struct address_space *i_mapping;
525 #ifdef CONFIG_SECURITY
529 /* Stat data, not accessed from path walking */
532 * Filesystems may only read i_nlink directly. They shall use the
533 * following functions for modification:
535 * (set|clear|inc|drop)_nlink
536 * inode_(inc|dec)_link_count
539 const unsigned int i_nlink;
540 unsigned int __i_nlink;
544 struct timespec i_atime;
545 struct timespec i_mtime;
546 struct timespec i_ctime;
547 spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
548 unsigned short i_bytes;
549 unsigned int i_blkbits;
552 #ifdef __NEED_I_SIZE_ORDERED
553 seqcount_t i_size_seqcount;
557 unsigned long i_state;
558 struct mutex i_mutex;
560 unsigned long dirtied_when; /* jiffies of first dirtying */
562 struct hlist_node i_hash;
563 struct list_head i_wb_list; /* backing dev IO list */
564 struct list_head i_lru; /* inode LRU list */
565 struct list_head i_sb_list;
567 struct hlist_head i_dentry;
568 struct rcu_head i_rcu;
572 atomic_t i_dio_count;
573 atomic_t i_writecount;
575 atomic_t i_readcount; /* struct files open RO */
577 const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
578 struct file_lock *i_flock;
579 struct address_space i_data;
581 struct dquot *i_dquot[MAXQUOTAS];
583 struct list_head i_devices;
585 struct pipe_inode_info *i_pipe;
586 struct block_device *i_bdev;
592 #ifdef CONFIG_FSNOTIFY
593 __u32 i_fsnotify_mask; /* all events this inode cares about */
594 struct hlist_head i_fsnotify_marks;
597 void *i_private; /* fs or device private pointer */
600 static inline int inode_unhashed(struct inode *inode)
602 return hlist_unhashed(&inode->i_hash);
606 * inode->i_mutex nesting subclasses for the lock validator:
608 * 0: the object of the current VFS operation
612 * 4: second non-directory
613 * The last is for certain operations (such as rename) which lock two
614 * non-directories at once.
616 * The locking order between these classes is
617 * parent -> child -> normal -> xattr -> second non-directory
619 enum inode_i_mutex_lock_class
628 void lock_two_nondirectories(struct inode *, struct inode*);
629 void unlock_two_nondirectories(struct inode *, struct inode*);
632 * NOTE: in a 32bit arch with a preemptable kernel and
633 * an UP compile the i_size_read/write must be atomic
634 * with respect to the local cpu (unlike with preempt disabled),
635 * but they don't need to be atomic with respect to other cpus like in
636 * true SMP (so they need either to either locally disable irq around
637 * the read or for example on x86 they can be still implemented as a
638 * cmpxchg8b without the need of the lock prefix). For SMP compiles
639 * and 64bit archs it makes no difference if preempt is enabled or not.
641 static inline loff_t i_size_read(const struct inode *inode)
643 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
648 seq = read_seqcount_begin(&inode->i_size_seqcount);
649 i_size = inode->i_size;
650 } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
652 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
656 i_size = inode->i_size;
660 return inode->i_size;
665 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
666 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
667 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
669 static inline void i_size_write(struct inode *inode, loff_t i_size)
671 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
673 write_seqcount_begin(&inode->i_size_seqcount);
674 inode->i_size = i_size;
675 write_seqcount_end(&inode->i_size_seqcount);
677 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
679 inode->i_size = i_size;
682 inode->i_size = i_size;
686 /* Helper functions so that in most cases filesystems will
687 * not need to deal directly with kuid_t and kgid_t and can
688 * instead deal with the raw numeric values that are stored
691 static inline uid_t i_uid_read(const struct inode *inode)
693 return from_kuid(&init_user_ns, inode->i_uid);
696 static inline gid_t i_gid_read(const struct inode *inode)
698 return from_kgid(&init_user_ns, inode->i_gid);
701 static inline void i_uid_write(struct inode *inode, uid_t uid)
703 inode->i_uid = make_kuid(&init_user_ns, uid);
706 static inline void i_gid_write(struct inode *inode, gid_t gid)
708 inode->i_gid = make_kgid(&init_user_ns, gid);
711 static inline unsigned iminor(const struct inode *inode)
713 return MINOR(inode->i_rdev);
716 static inline unsigned imajor(const struct inode *inode)
718 return MAJOR(inode->i_rdev);
721 extern struct block_device *I_BDEV(struct inode *inode);
724 rwlock_t lock; /* protects pid, uid, euid fields */
725 struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
726 enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
727 kuid_t uid, euid; /* uid/euid of process setting the owner */
728 int signum; /* posix.1b rt signal to be delivered on IO */
732 * Track a single file's readahead state
734 struct file_ra_state {
735 pgoff_t start; /* where readahead started */
736 unsigned int size; /* # of readahead pages */
737 unsigned int async_size; /* do asynchronous readahead when
738 there are only # of pages ahead */
740 unsigned int ra_pages; /* Maximum readahead window */
741 unsigned int mmap_miss; /* Cache miss stat for mmap accesses */
742 loff_t prev_pos; /* Cache last read() position */
746 * Check if @index falls in the readahead windows.
748 static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
750 return (index >= ra->start &&
751 index < ra->start + ra->size);
756 struct llist_node fu_llist;
757 struct rcu_head fu_rcuhead;
760 #define f_dentry f_path.dentry
761 struct inode *f_inode; /* cached value */
762 const struct file_operations *f_op;
765 * Protects f_ep_links, f_flags.
766 * Must not be taken from IRQ context.
769 atomic_long_t f_count;
770 unsigned int f_flags;
772 struct mutex f_pos_lock;
774 struct fown_struct f_owner;
775 const struct cred *f_cred;
776 struct file_ra_state f_ra;
779 #ifdef CONFIG_SECURITY
782 /* needed for tty driver, and maybe others */
786 /* Used by fs/eventpoll.c to link all the hooks to this file */
787 struct list_head f_ep_links;
788 struct list_head f_tfile_llink;
789 #endif /* #ifdef CONFIG_EPOLL */
790 struct address_space *f_mapping;
791 } __attribute__((aligned(4))); /* lest something weird decides that 2 is OK */
796 /* file identifier */
797 unsigned char f_handle[0];
800 static inline struct file *get_file(struct file *f)
802 atomic_long_inc(&f->f_count);
805 #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
806 #define file_count(x) atomic_long_read(&(x)->f_count)
808 #define MAX_NON_LFS ((1UL<<31) - 1)
810 /* Page cache limit. The filesystems should put that into their s_maxbytes
811 limits, otherwise bad things can happen in VM. */
812 #if BITS_PER_LONG==32
813 #define MAX_LFS_FILESIZE ((loff_t)ULONG_MAX << PAGE_SHIFT)
814 #elif BITS_PER_LONG==64
815 #define MAX_LFS_FILESIZE ((loff_t)LLONG_MAX)
820 #define FL_DELEG 4 /* NFSv4 delegation */
821 #define FL_ACCESS 8 /* not trying to lock, just looking */
822 #define FL_EXISTS 16 /* when unlocking, test for existence */
823 #define FL_LEASE 32 /* lease held on this file */
824 #define FL_CLOSE 64 /* unlock on close */
825 #define FL_SLEEP 128 /* A blocking lock */
826 #define FL_DOWNGRADE_PENDING 256 /* Lease is being downgraded */
827 #define FL_UNLOCK_PENDING 512 /* Lease is being broken */
828 #define FL_OFDLCK 1024 /* lock is "owned" by struct file */
831 * Special return value from posix_lock_file() and vfs_lock_file() for
832 * asynchronous locking.
834 #define FILE_LOCK_DEFERRED 1
837 * The POSIX file lock owner is determined by
838 * the "struct files_struct" in the thread group
839 * (or NULL for no owner - BSD locks).
841 * Lockd stuffs a "host" pointer into this.
843 typedef struct files_struct *fl_owner_t;
845 struct file_lock_operations {
846 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
847 void (*fl_release_private)(struct file_lock *);
850 struct lock_manager_operations {
851 int (*lm_compare_owner)(struct file_lock *, struct file_lock *);
852 unsigned long (*lm_owner_key)(struct file_lock *);
853 void (*lm_notify)(struct file_lock *); /* unblock callback */
854 int (*lm_grant)(struct file_lock *, struct file_lock *, int);
855 void (*lm_break)(struct file_lock *);
856 int (*lm_change)(struct file_lock **, int);
859 struct lock_manager {
860 struct list_head list;
864 void locks_start_grace(struct net *, struct lock_manager *);
865 void locks_end_grace(struct lock_manager *);
866 int locks_in_grace(struct net *);
868 /* that will die - we need it for nfs_lock_info */
869 #include <linux/nfs_fs_i.h>
872 * struct file_lock represents a generic "file lock". It's used to represent
873 * POSIX byte range locks, BSD (flock) locks, and leases. It's important to
874 * note that the same struct is used to represent both a request for a lock and
875 * the lock itself, but the same object is never used for both.
877 * FIXME: should we create a separate "struct lock_request" to help distinguish
880 * The i_flock list is ordered by:
882 * 1) lock type -- FL_LEASEs first, then FL_FLOCK, and finally FL_POSIX
884 * 3) lock range start
887 * Obviously, the last two criteria only matter for POSIX locks.
890 struct file_lock *fl_next; /* singly linked list for this inode */
891 struct hlist_node fl_link; /* node in global lists */
892 struct list_head fl_block; /* circular list of blocked processes */
894 unsigned int fl_flags;
895 unsigned char fl_type;
897 int fl_link_cpu; /* what cpu's list is this on? */
898 struct pid *fl_nspid;
899 wait_queue_head_t fl_wait;
900 struct file *fl_file;
904 struct fasync_struct * fl_fasync; /* for lease break notifications */
905 /* for lease breaks: */
906 unsigned long fl_break_time;
907 unsigned long fl_downgrade_time;
909 const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
910 const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
912 struct nfs_lock_info nfs_fl;
913 struct nfs4_lock_info nfs4_fl;
915 struct list_head link; /* link in AFS vnode's pending_locks list */
916 int state; /* state of grant or error if -ve */
921 /* The following constant reflects the upper bound of the file/locking space */
923 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
924 #define OFFSET_MAX INT_LIMIT(loff_t)
925 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
928 #include <linux/fcntl.h>
930 extern void send_sigio(struct fown_struct *fown, int fd, int band);
932 #ifdef CONFIG_FILE_LOCKING
933 extern int fcntl_getlk(struct file *, unsigned int, struct flock __user *);
934 extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
935 struct flock __user *);
937 #if BITS_PER_LONG == 32
938 extern int fcntl_getlk64(struct file *, unsigned int, struct flock64 __user *);
939 extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
940 struct flock64 __user *);
943 extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
944 extern int fcntl_getlease(struct file *filp);
947 void locks_free_lock(struct file_lock *fl);
948 extern void locks_init_lock(struct file_lock *);
949 extern struct file_lock * locks_alloc_lock(void);
950 extern void locks_copy_lock(struct file_lock *, struct file_lock *);
951 extern void __locks_copy_lock(struct file_lock *, const struct file_lock *);
952 extern void locks_remove_posix(struct file *, fl_owner_t);
953 extern void locks_remove_file(struct file *);
954 extern void locks_release_private(struct file_lock *);
955 extern void posix_test_lock(struct file *, struct file_lock *);
956 extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
957 extern int posix_lock_file_wait(struct file *, struct file_lock *);
958 extern int posix_unblock_lock(struct file_lock *);
959 extern int vfs_test_lock(struct file *, struct file_lock *);
960 extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
961 extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
962 extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
963 extern int __break_lease(struct inode *inode, unsigned int flags, unsigned int type);
964 extern void lease_get_mtime(struct inode *, struct timespec *time);
965 extern int generic_setlease(struct file *, long, struct file_lock **);
966 extern int vfs_setlease(struct file *, long, struct file_lock **);
967 extern int lease_modify(struct file_lock **, int);
968 extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
969 extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
970 #else /* !CONFIG_FILE_LOCKING */
971 static inline int fcntl_getlk(struct file *file, unsigned int cmd,
972 struct flock __user *user)
977 static inline int fcntl_setlk(unsigned int fd, struct file *file,
978 unsigned int cmd, struct flock __user *user)
983 #if BITS_PER_LONG == 32
984 static inline int fcntl_getlk64(struct file *file, unsigned int cmd,
985 struct flock64 __user *user)
990 static inline int fcntl_setlk64(unsigned int fd, struct file *file,
991 unsigned int cmd, struct flock64 __user *user)
996 static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
1001 static inline int fcntl_getlease(struct file *filp)
1006 static inline void locks_init_lock(struct file_lock *fl)
1011 static inline void __locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1016 static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1021 static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
1026 static inline void locks_remove_file(struct file *filp)
1031 static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
1036 static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
1037 struct file_lock *conflock)
1042 static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
1047 static inline int posix_unblock_lock(struct file_lock *waiter)
1052 static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
1057 static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
1058 struct file_lock *fl, struct file_lock *conf)
1063 static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
1068 static inline int flock_lock_file_wait(struct file *filp,
1069 struct file_lock *request)
1074 static inline int __break_lease(struct inode *inode, unsigned int mode, unsigned int type)
1079 static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
1084 static inline int generic_setlease(struct file *filp, long arg,
1085 struct file_lock **flp)
1090 static inline int vfs_setlease(struct file *filp, long arg,
1091 struct file_lock **lease)
1096 static inline int lease_modify(struct file_lock **before, int arg)
1101 static inline int lock_may_read(struct inode *inode, loff_t start,
1107 static inline int lock_may_write(struct inode *inode, loff_t start,
1112 #endif /* !CONFIG_FILE_LOCKING */
1115 struct fasync_struct {
1119 struct fasync_struct *fa_next; /* singly linked list */
1120 struct file *fa_file;
1121 struct rcu_head fa_rcu;
1124 #define FASYNC_MAGIC 0x4601
1126 /* SMP safe fasync helpers: */
1127 extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
1128 extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
1129 extern int fasync_remove_entry(struct file *, struct fasync_struct **);
1130 extern struct fasync_struct *fasync_alloc(void);
1131 extern void fasync_free(struct fasync_struct *);
1133 /* can be called from interrupts */
1134 extern void kill_fasync(struct fasync_struct **, int, int);
1136 extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
1137 extern int f_setown(struct file *filp, unsigned long arg, int force);
1138 extern void f_delown(struct file *filp);
1139 extern pid_t f_getown(struct file *filp);
1140 extern int send_sigurg(struct fown_struct *fown);
1148 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1149 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1150 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1151 #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1152 #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
1155 #define SB_I_MULTIROOT 0x00000008 /* Multiple roots to the dentry tree */
1157 extern struct list_head super_blocks;
1158 extern spinlock_t sb_lock;
1160 /* Possible states of 'frozen' field */
1162 SB_UNFROZEN = 0, /* FS is unfrozen */
1163 SB_FREEZE_WRITE = 1, /* Writes, dir ops, ioctls frozen */
1164 SB_FREEZE_PAGEFAULT = 2, /* Page faults stopped as well */
1165 SB_FREEZE_FS = 3, /* For internal FS use (e.g. to stop
1166 * internal threads if needed) */
1167 SB_FREEZE_COMPLETE = 4, /* ->freeze_fs finished successfully */
1170 #define SB_FREEZE_LEVELS (SB_FREEZE_COMPLETE - 1)
1173 /* Counters for counting writers at each level */
1174 struct percpu_counter counter[SB_FREEZE_LEVELS];
1175 wait_queue_head_t wait; /* queue for waiting for
1176 writers / faults to finish */
1177 int frozen; /* Is sb frozen? */
1178 wait_queue_head_t wait_unfrozen; /* queue for waiting for
1180 #ifdef CONFIG_DEBUG_LOCK_ALLOC
1181 struct lockdep_map lock_map[SB_FREEZE_LEVELS];
1185 struct super_block {
1186 struct list_head s_list; /* Keep this first */
1187 dev_t s_dev; /* search index; _not_ kdev_t */
1188 unsigned char s_blocksize_bits;
1189 unsigned long s_blocksize;
1190 loff_t s_maxbytes; /* Max file size */
1191 struct file_system_type *s_type;
1192 const struct super_operations *s_op;
1193 const struct dquot_operations *dq_op;
1194 const struct quotactl_ops *s_qcop;
1195 const struct export_operations *s_export_op;
1196 unsigned long s_flags;
1197 unsigned long s_iflags; /* internal SB_I_* flags */
1198 unsigned long s_magic;
1199 struct dentry *s_root;
1200 struct rw_semaphore s_umount;
1203 #ifdef CONFIG_SECURITY
1206 const struct xattr_handler **s_xattr;
1208 struct list_head s_inodes; /* all inodes */
1209 struct hlist_bl_head s_anon; /* anonymous dentries for (nfs) exporting */
1210 struct list_head s_mounts; /* list of mounts; _not_ for fs use */
1211 struct block_device *s_bdev;
1212 struct backing_dev_info *s_bdi;
1213 struct mtd_info *s_mtd;
1214 struct hlist_node s_instances;
1215 struct quota_info s_dquot; /* Diskquota specific options */
1217 struct sb_writers s_writers;
1219 char s_id[32]; /* Informational name */
1220 u8 s_uuid[16]; /* UUID */
1222 void *s_fs_info; /* Filesystem private info */
1223 unsigned int s_max_links;
1226 /* Granularity of c/m/atime in ns.
1227 Cannot be worse than a second */
1231 * The next field is for VFS *only*. No filesystems have any business
1232 * even looking at it. You had been warned.
1234 struct mutex s_vfs_rename_mutex; /* Kludge */
1237 * Filesystem subtype. If non-empty the filesystem type field
1238 * in /proc/mounts will be "type.subtype"
1243 * Saved mount options for lazy filesystems using
1244 * generic_show_options()
1246 char __rcu *s_options;
1247 const struct dentry_operations *s_d_op; /* default d_op for dentries */
1250 * Saved pool identifier for cleancache (-1 means none)
1252 int cleancache_poolid;
1254 struct shrinker s_shrink; /* per-sb shrinker handle */
1256 /* Number of inodes with nlink == 0 but still referenced */
1257 atomic_long_t s_remove_count;
1259 /* Being remounted read-only */
1260 int s_readonly_remount;
1262 /* AIO completions deferred from interrupt context */
1263 struct workqueue_struct *s_dio_done_wq;
1266 * Keep the lru lists last in the structure so they always sit on their
1267 * own individual cachelines.
1269 struct list_lru s_dentry_lru ____cacheline_aligned_in_smp;
1270 struct list_lru s_inode_lru ____cacheline_aligned_in_smp;
1271 struct rcu_head rcu;
1274 * Indicates how deep in a filesystem stack this SB is
1279 extern struct timespec current_fs_time(struct super_block *sb);
1282 * Snapshotting support.
1285 void __sb_end_write(struct super_block *sb, int level);
1286 int __sb_start_write(struct super_block *sb, int level, bool wait);
1289 * sb_end_write - drop write access to a superblock
1290 * @sb: the super we wrote to
1292 * Decrement number of writers to the filesystem. Wake up possible waiters
1293 * wanting to freeze the filesystem.
1295 static inline void sb_end_write(struct super_block *sb)
1297 __sb_end_write(sb, SB_FREEZE_WRITE);
1301 * sb_end_pagefault - drop write access to a superblock from a page fault
1302 * @sb: the super we wrote to
1304 * Decrement number of processes handling write page fault to the filesystem.
1305 * Wake up possible waiters wanting to freeze the filesystem.
1307 static inline void sb_end_pagefault(struct super_block *sb)
1309 __sb_end_write(sb, SB_FREEZE_PAGEFAULT);
1313 * sb_end_intwrite - drop write access to a superblock for internal fs purposes
1314 * @sb: the super we wrote to
1316 * Decrement fs-internal number of writers to the filesystem. Wake up possible
1317 * waiters wanting to freeze the filesystem.
1319 static inline void sb_end_intwrite(struct super_block *sb)
1321 __sb_end_write(sb, SB_FREEZE_FS);
1325 * sb_start_write - get write access to a superblock
1326 * @sb: the super we write to
1328 * When a process wants to write data or metadata to a file system (i.e. dirty
1329 * a page or an inode), it should embed the operation in a sb_start_write() -
1330 * sb_end_write() pair to get exclusion against file system freezing. This
1331 * function increments number of writers preventing freezing. If the file
1332 * system is already frozen, the function waits until the file system is
1335 * Since freeze protection behaves as a lock, users have to preserve
1336 * ordering of freeze protection and other filesystem locks. Generally,
1337 * freeze protection should be the outermost lock. In particular, we have:
1340 * -> i_mutex (write path, truncate, directory ops, ...)
1341 * -> s_umount (freeze_super, thaw_super)
1343 static inline void sb_start_write(struct super_block *sb)
1345 __sb_start_write(sb, SB_FREEZE_WRITE, true);
1348 static inline int sb_start_write_trylock(struct super_block *sb)
1350 return __sb_start_write(sb, SB_FREEZE_WRITE, false);
1354 * sb_start_pagefault - get write access to a superblock from a page fault
1355 * @sb: the super we write to
1357 * When a process starts handling write page fault, it should embed the
1358 * operation into sb_start_pagefault() - sb_end_pagefault() pair to get
1359 * exclusion against file system freezing. This is needed since the page fault
1360 * is going to dirty a page. This function increments number of running page
1361 * faults preventing freezing. If the file system is already frozen, the
1362 * function waits until the file system is thawed.
1364 * Since page fault freeze protection behaves as a lock, users have to preserve
1365 * ordering of freeze protection and other filesystem locks. It is advised to
1366 * put sb_start_pagefault() close to mmap_sem in lock ordering. Page fault
1367 * handling code implies lock dependency:
1370 * -> sb_start_pagefault
1372 static inline void sb_start_pagefault(struct super_block *sb)
1374 __sb_start_write(sb, SB_FREEZE_PAGEFAULT, true);
1378 * sb_start_intwrite - get write access to a superblock for internal fs purposes
1379 * @sb: the super we write to
1381 * This is the third level of protection against filesystem freezing. It is
1382 * free for use by a filesystem. The only requirement is that it must rank
1383 * below sb_start_pagefault.
1385 * For example filesystem can call sb_start_intwrite() when starting a
1386 * transaction which somewhat eases handling of freezing for internal sources
1387 * of filesystem changes (internal fs threads, discarding preallocation on file
1390 static inline void sb_start_intwrite(struct super_block *sb)
1392 __sb_start_write(sb, SB_FREEZE_FS, true);
1395 static inline int sb_start_intwrite_trylock(struct super_block *sb)
1397 return __sb_start_write(sb, SB_FREEZE_FS, false);
1401 extern bool inode_owner_or_capable(const struct inode *inode);
1404 * VFS helper functions..
1406 extern int vfs_create(struct inode *, struct dentry *, umode_t, bool);
1407 extern int vfs_mkdir(struct inode *, struct dentry *, umode_t);
1408 extern int vfs_mknod(struct inode *, struct dentry *, umode_t, dev_t);
1409 extern int vfs_symlink(struct inode *, struct dentry *, const char *);
1410 extern int vfs_link(struct dentry *, struct inode *, struct dentry *, struct inode **);
1411 extern int vfs_rmdir(struct inode *, struct dentry *);
1412 extern int vfs_unlink(struct inode *, struct dentry *, struct inode **);
1413 extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *, struct inode **, unsigned int);
1416 * VFS dentry helper functions.
1418 extern void dentry_unhash(struct dentry *dentry);
1421 * VFS file helper functions.
1423 extern void inode_init_owner(struct inode *inode, const struct inode *dir,
1426 * VFS FS_IOC_FIEMAP helper definitions.
1428 struct fiemap_extent_info {
1429 unsigned int fi_flags; /* Flags as passed from user */
1430 unsigned int fi_extents_mapped; /* Number of mapped extents */
1431 unsigned int fi_extents_max; /* Size of fiemap_extent array */
1432 struct fiemap_extent __user *fi_extents_start; /* Start of
1433 fiemap_extent array */
1435 int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
1436 u64 phys, u64 len, u32 flags);
1437 int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
1442 * NOTE! These match bits 12..15 of stat.st_mode
1443 * (ie "(i_mode >> 12) & 15").
1445 #define DT_UNKNOWN 0
1456 * This is the "filldir" function type, used by readdir() to let
1457 * the kernel specify what kind of dirent layout it wants to have.
1458 * This allows the kernel to read directories into kernel space or
1459 * to have different dirent layouts depending on the binary type.
1461 typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned);
1462 struct dir_context {
1463 const filldir_t actor;
1467 struct block_device_operations;
1469 /* These macros are for out of kernel modules to test that
1470 * the kernel supports the unlocked_ioctl and compat_ioctl
1471 * fields in struct file_operations. */
1472 #define HAVE_COMPAT_IOCTL 1
1473 #define HAVE_UNLOCKED_IOCTL 1
1477 struct file_operations {
1478 struct module *owner;
1479 loff_t (*llseek) (struct file *, loff_t, int);
1480 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1481 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
1482 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1483 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1484 ssize_t (*read_iter) (struct kiocb *, struct iov_iter *);
1485 ssize_t (*write_iter) (struct kiocb *, struct iov_iter *);
1486 int (*iterate) (struct file *, struct dir_context *);
1487 unsigned int (*poll) (struct file *, struct poll_table_struct *);
1488 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1489 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1490 int (*mmap) (struct file *, struct vm_area_struct *);
1491 int (*open) (struct inode *, struct file *);
1492 int (*flush) (struct file *, fl_owner_t id);
1493 int (*release) (struct inode *, struct file *);
1494 int (*fsync) (struct file *, loff_t, loff_t, int datasync);
1495 int (*aio_fsync) (struct kiocb *, int datasync);
1496 int (*fasync) (int, struct file *, int);
1497 int (*lock) (struct file *, int, struct file_lock *);
1498 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1499 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1500 int (*check_flags)(int);
1501 int (*flock) (struct file *, int, struct file_lock *);
1502 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1503 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
1504 int (*setlease)(struct file *, long, struct file_lock **);
1505 long (*fallocate)(struct file *file, int mode, loff_t offset,
1507 int (*show_fdinfo)(struct seq_file *m, struct file *f);
1510 struct inode_operations {
1511 struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int);
1512 void * (*follow_link) (struct dentry *, struct nameidata *);
1513 int (*permission) (struct inode *, int);
1514 struct posix_acl * (*get_acl)(struct inode *, int);
1516 int (*readlink) (struct dentry *, char __user *,int);
1517 void (*put_link) (struct dentry *, struct nameidata *, void *);
1519 int (*create) (struct inode *,struct dentry *, umode_t, bool);
1520 int (*link) (struct dentry *,struct inode *,struct dentry *);
1521 int (*unlink) (struct inode *,struct dentry *);
1522 int (*symlink) (struct inode *,struct dentry *,const char *);
1523 int (*mkdir) (struct inode *,struct dentry *,umode_t);
1524 int (*rmdir) (struct inode *,struct dentry *);
1525 int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t);
1526 int (*rename) (struct inode *, struct dentry *,
1527 struct inode *, struct dentry *);
1528 int (*rename2) (struct inode *, struct dentry *,
1529 struct inode *, struct dentry *, unsigned int);
1530 int (*setattr) (struct dentry *, struct iattr *);
1531 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1532 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1533 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1534 ssize_t (*listxattr) (struct dentry *, char *, size_t);
1535 int (*removexattr) (struct dentry *, const char *);
1536 int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
1538 int (*update_time)(struct inode *, struct timespec *, int);
1539 int (*atomic_open)(struct inode *, struct dentry *,
1540 struct file *, unsigned open_flag,
1541 umode_t create_mode, int *opened);
1542 int (*tmpfile) (struct inode *, struct dentry *, umode_t);
1543 int (*set_acl)(struct inode *, struct posix_acl *, int);
1544 } ____cacheline_aligned;
1546 ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
1547 unsigned long nr_segs, unsigned long fast_segs,
1548 struct iovec *fast_pointer,
1549 struct iovec **ret_pointer);
1551 extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1552 extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1553 extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1554 unsigned long, loff_t *);
1555 extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1556 unsigned long, loff_t *);
1558 struct super_operations {
1559 struct inode *(*alloc_inode)(struct super_block *sb);
1560 void (*destroy_inode)(struct inode *);
1562 void (*dirty_inode) (struct inode *, int flags);
1563 int (*write_inode) (struct inode *, struct writeback_control *wbc);
1564 int (*drop_inode) (struct inode *);
1565 void (*evict_inode) (struct inode *);
1566 void (*put_super) (struct super_block *);
1567 int (*sync_fs)(struct super_block *sb, int wait);
1568 int (*freeze_fs) (struct super_block *);
1569 int (*unfreeze_fs) (struct super_block *);
1570 int (*statfs) (struct dentry *, struct kstatfs *);
1571 int (*remount_fs) (struct super_block *, int *, char *);
1572 void (*umount_begin) (struct super_block *);
1574 int (*show_options)(struct seq_file *, struct dentry *);
1575 int (*show_devname)(struct seq_file *, struct dentry *);
1576 int (*show_path)(struct seq_file *, struct dentry *);
1577 int (*show_stats)(struct seq_file *, struct dentry *);
1579 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1580 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
1582 int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
1583 long (*nr_cached_objects)(struct super_block *, int);
1584 long (*free_cached_objects)(struct super_block *, long, int);
1588 * Inode flags - they have no relation to superblock flags now
1590 #define S_SYNC 1 /* Writes are synced at once */
1591 #define S_NOATIME 2 /* Do not update access times */
1592 #define S_APPEND 4 /* Append-only file */
1593 #define S_IMMUTABLE 8 /* Immutable file */
1594 #define S_DEAD 16 /* removed, but still open directory */
1595 #define S_NOQUOTA 32 /* Inode is not counted to quota */
1596 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
1597 #define S_NOCMTIME 128 /* Do not update file c/mtime */
1598 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
1599 #define S_PRIVATE 512 /* Inode is fs-internal */
1600 #define S_IMA 1024 /* Inode has an associated IMA struct */
1601 #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
1602 #define S_NOSEC 4096 /* no suid or xattr security attributes */
1605 * Note that nosuid etc flags are inode-specific: setting some file-system
1606 * flags just means all the inodes inherit those flags by default. It might be
1607 * possible to override it selectively if you really wanted to with some
1608 * ioctl() that is not currently implemented.
1610 * Exception: MS_RDONLY is always applied to the entire file system.
1612 * Unfortunately, it is possible to change a filesystems flags with it mounted
1613 * with files in use. This means that all of the inodes will not have their
1614 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
1615 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
1617 #define __IS_FLG(inode, flg) ((inode)->i_sb->s_flags & (flg))
1619 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
1620 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
1621 ((inode)->i_flags & S_SYNC))
1622 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
1623 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
1624 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
1625 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
1626 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
1628 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
1629 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
1630 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
1631 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
1633 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
1634 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
1635 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
1636 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
1637 #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
1638 #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
1639 #define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
1642 * Inode state bits. Protected by inode->i_lock
1644 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1645 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1647 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1648 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1649 * various stages of removing an inode.
1651 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1653 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1654 * fdatasync(). i_atime is the usual cause.
1655 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1656 * these changes separately from I_DIRTY_SYNC so that we
1657 * don't have to write inode on fdatasync() when only
1658 * mtime has changed in it.
1659 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1660 * I_NEW Serves as both a mutex and completion notification.
1661 * New inodes set I_NEW. If two processes both create
1662 * the same inode, one of them will release its inode and
1663 * wait for I_NEW to be released before returning.
1664 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1665 * also cause waiting on I_NEW, without I_NEW actually
1666 * being set. find_inode() uses this to prevent returning
1667 * nearly-dead inodes.
1668 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1669 * is zero. I_FREEING must be set when I_WILL_FREE is
1671 * I_FREEING Set when inode is about to be freed but still has dirty
1672 * pages or buffers attached or the inode itself is still
1674 * I_CLEAR Added by clear_inode(). In this state the inode is
1675 * clean and can be destroyed. Inode keeps I_FREEING.
1677 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1678 * prohibited for many purposes. iget() must wait for
1679 * the inode to be completely released, then create it
1680 * anew. Other functions will just ignore such inodes,
1681 * if appropriate. I_NEW is used for waiting.
1683 * I_SYNC Writeback of inode is running. The bit is set during
1684 * data writeback, and cleared with a wakeup on the bit
1685 * address once it is done. The bit is also used to pin
1686 * the inode in memory for flusher thread.
1688 * I_REFERENCED Marks the inode as recently references on the LRU list.
1690 * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit().
1692 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1694 #define I_DIRTY_SYNC (1 << 0)
1695 #define I_DIRTY_DATASYNC (1 << 1)
1696 #define I_DIRTY_PAGES (1 << 2)
1698 #define I_NEW (1 << __I_NEW)
1699 #define I_WILL_FREE (1 << 4)
1700 #define I_FREEING (1 << 5)
1701 #define I_CLEAR (1 << 6)
1703 #define I_SYNC (1 << __I_SYNC)
1704 #define I_REFERENCED (1 << 8)
1705 #define __I_DIO_WAKEUP 9
1706 #define I_DIO_WAKEUP (1 << I_DIO_WAKEUP)
1707 #define I_LINKABLE (1 << 10)
1709 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1711 extern void __mark_inode_dirty(struct inode *, int);
1712 static inline void mark_inode_dirty(struct inode *inode)
1714 __mark_inode_dirty(inode, I_DIRTY);
1717 static inline void mark_inode_dirty_sync(struct inode *inode)
1719 __mark_inode_dirty(inode, I_DIRTY_SYNC);
1722 extern void inc_nlink(struct inode *inode);
1723 extern void drop_nlink(struct inode *inode);
1724 extern void clear_nlink(struct inode *inode);
1725 extern void set_nlink(struct inode *inode, unsigned int nlink);
1727 static inline void inode_inc_link_count(struct inode *inode)
1730 mark_inode_dirty(inode);
1733 static inline void inode_dec_link_count(struct inode *inode)
1736 mark_inode_dirty(inode);
1740 * inode_inc_iversion - increments i_version
1741 * @inode: inode that need to be updated
1743 * Every time the inode is modified, the i_version field will be incremented.
1744 * The filesystem has to be mounted with i_version flag
1747 static inline void inode_inc_iversion(struct inode *inode)
1749 spin_lock(&inode->i_lock);
1751 spin_unlock(&inode->i_lock);
1754 enum file_time_flags {
1761 extern void touch_atime(const struct path *);
1762 static inline void file_accessed(struct file *file)
1764 if (!(file->f_flags & O_NOATIME))
1765 touch_atime(&file->f_path);
1768 int sync_inode(struct inode *inode, struct writeback_control *wbc);
1769 int sync_inode_metadata(struct inode *inode, int wait);
1771 struct file_system_type {
1774 #define FS_REQUIRES_DEV 1
1775 #define FS_BINARY_MOUNTDATA 2
1776 #define FS_HAS_SUBTYPE 4
1777 #define FS_USERNS_MOUNT 8 /* Can be mounted by userns root */
1778 #define FS_USERNS_DEV_MOUNT 16 /* A userns mount does not imply MNT_NODEV */
1779 #define FS_USERNS_VISIBLE 32 /* FS must already be visible */
1780 #define FS_NOEXEC 64 /* Ignore executables on this fs */
1781 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() during rename() internally. */
1782 struct dentry *(*mount) (struct file_system_type *, int,
1783 const char *, void *);
1784 void (*kill_sb) (struct super_block *);
1785 struct module *owner;
1786 struct file_system_type * next;
1787 struct hlist_head fs_supers;
1789 struct lock_class_key s_lock_key;
1790 struct lock_class_key s_umount_key;
1791 struct lock_class_key s_vfs_rename_key;
1792 struct lock_class_key s_writers_key[SB_FREEZE_LEVELS];
1794 struct lock_class_key i_lock_key;
1795 struct lock_class_key i_mutex_key;
1796 struct lock_class_key i_mutex_dir_key;
1799 #define MODULE_ALIAS_FS(NAME) MODULE_ALIAS("fs-" NAME)
1801 extern struct dentry *mount_ns(struct file_system_type *fs_type, int flags,
1802 void *data, int (*fill_super)(struct super_block *, void *, int));
1803 extern struct dentry *mount_bdev(struct file_system_type *fs_type,
1804 int flags, const char *dev_name, void *data,
1805 int (*fill_super)(struct super_block *, void *, int));
1806 extern struct dentry *mount_single(struct file_system_type *fs_type,
1807 int flags, void *data,
1808 int (*fill_super)(struct super_block *, void *, int));
1809 extern struct dentry *mount_nodev(struct file_system_type *fs_type,
1810 int flags, void *data,
1811 int (*fill_super)(struct super_block *, void *, int));
1812 extern struct dentry *mount_subtree(struct vfsmount *mnt, const char *path);
1813 void generic_shutdown_super(struct super_block *sb);
1814 void kill_block_super(struct super_block *sb);
1815 void kill_anon_super(struct super_block *sb);
1816 void kill_litter_super(struct super_block *sb);
1817 void deactivate_super(struct super_block *sb);
1818 void deactivate_locked_super(struct super_block *sb);
1819 int set_anon_super(struct super_block *s, void *data);
1820 int get_anon_bdev(dev_t *);
1821 void free_anon_bdev(dev_t);
1822 struct super_block *sget(struct file_system_type *type,
1823 int (*test)(struct super_block *,void *),
1824 int (*set)(struct super_block *,void *),
1825 int flags, void *data);
1826 extern struct dentry *mount_pseudo(struct file_system_type *, char *,
1827 const struct super_operations *ops,
1828 const struct dentry_operations *dops,
1831 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1832 #define fops_get(fops) \
1833 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1834 #define fops_put(fops) \
1835 do { if (fops) module_put((fops)->owner); } while(0)
1837 * This one is to be used *ONLY* from ->open() instances.
1838 * fops must be non-NULL, pinned down *and* module dependencies
1839 * should be sufficient to pin the caller down as well.
1841 #define replace_fops(f, fops) \
1843 struct file *__file = (f); \
1844 fops_put(__file->f_op); \
1845 BUG_ON(!(__file->f_op = (fops))); \
1848 extern int register_filesystem(struct file_system_type *);
1849 extern int unregister_filesystem(struct file_system_type *);
1850 extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
1851 #define kern_mount(type) kern_mount_data(type, NULL)
1852 extern void kern_unmount(struct vfsmount *mnt);
1853 extern int may_umount_tree(struct vfsmount *);
1854 extern int may_umount(struct vfsmount *);
1855 extern long do_mount(const char *, const char *, const char *, unsigned long, void *);
1856 extern struct vfsmount *collect_mounts(struct path *);
1857 extern void drop_collected_mounts(struct vfsmount *);
1858 extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
1860 extern int vfs_statfs(struct path *, struct kstatfs *);
1861 extern int user_statfs(const char __user *, struct kstatfs *);
1862 extern int fd_statfs(int, struct kstatfs *);
1863 extern int vfs_ustat(dev_t, struct kstatfs *);
1864 extern int freeze_super(struct super_block *super);
1865 extern int thaw_super(struct super_block *super);
1866 extern bool our_mnt(struct vfsmount *mnt);
1868 extern int current_umask(void);
1870 extern void ihold(struct inode * inode);
1871 extern void iput(struct inode *);
1873 static inline struct inode *file_inode(struct file *f)
1879 extern struct kobject *fs_kobj;
1881 #define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
1883 #define FLOCK_VERIFY_READ 1
1884 #define FLOCK_VERIFY_WRITE 2
1886 #ifdef CONFIG_FILE_LOCKING
1887 extern int locks_mandatory_locked(struct file *);
1888 extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
1891 * Candidates for mandatory locking have the setgid bit set
1892 * but no group execute bit - an otherwise meaningless combination.
1895 static inline int __mandatory_lock(struct inode *ino)
1897 return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
1901 * ... and these candidates should be on MS_MANDLOCK mounted fs,
1902 * otherwise these will be advisory locks
1905 static inline int mandatory_lock(struct inode *ino)
1907 return IS_MANDLOCK(ino) && __mandatory_lock(ino);
1910 static inline int locks_verify_locked(struct file *file)
1912 if (mandatory_lock(file_inode(file)))
1913 return locks_mandatory_locked(file);
1917 static inline int locks_verify_truncate(struct inode *inode,
1921 if (inode->i_flock && mandatory_lock(inode))
1922 return locks_mandatory_area(
1923 FLOCK_VERIFY_WRITE, inode, filp,
1924 size < inode->i_size ? size : inode->i_size,
1925 (size < inode->i_size ? inode->i_size - size
1926 : size - inode->i_size)
1931 static inline int break_lease(struct inode *inode, unsigned int mode)
1934 * Since this check is lockless, we must ensure that any refcounts
1935 * taken are done before checking inode->i_flock. Otherwise, we could
1936 * end up racing with tasks trying to set a new lease on this file.
1940 return __break_lease(inode, mode, FL_LEASE);
1944 static inline int break_deleg(struct inode *inode, unsigned int mode)
1947 * Since this check is lockless, we must ensure that any refcounts
1948 * taken are done before checking inode->i_flock. Otherwise, we could
1949 * end up racing with tasks trying to set a new lease on this file.
1953 return __break_lease(inode, mode, FL_DELEG);
1957 static inline int try_break_deleg(struct inode *inode, struct inode **delegated_inode)
1961 ret = break_deleg(inode, O_WRONLY|O_NONBLOCK);
1962 if (ret == -EWOULDBLOCK && delegated_inode) {
1963 *delegated_inode = inode;
1969 static inline int break_deleg_wait(struct inode **delegated_inode)
1973 ret = break_deleg(*delegated_inode, O_WRONLY);
1974 iput(*delegated_inode);
1975 *delegated_inode = NULL;
1979 #else /* !CONFIG_FILE_LOCKING */
1980 static inline int locks_mandatory_locked(struct file *file)
1985 static inline int locks_mandatory_area(int rw, struct inode *inode,
1986 struct file *filp, loff_t offset,
1992 static inline int __mandatory_lock(struct inode *inode)
1997 static inline int mandatory_lock(struct inode *inode)
2002 static inline int locks_verify_locked(struct file *file)
2007 static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
2013 static inline int break_lease(struct inode *inode, unsigned int mode)
2018 static inline int break_deleg(struct inode *inode, unsigned int mode)
2023 static inline int try_break_deleg(struct inode *inode, struct inode **delegated_inode)
2028 static inline int break_deleg_wait(struct inode **delegated_inode)
2034 #endif /* CONFIG_FILE_LOCKING */
2039 const char *name; /* pointer to actual string */
2040 const __user char *uptr; /* original userland pointer */
2041 struct audit_names *aname;
2042 bool separate; /* should "name" be freed? */
2045 extern long vfs_truncate(struct path *, loff_t);
2046 extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
2048 extern int do_fallocate(struct file *file, int mode, loff_t offset,
2050 extern long do_sys_open(int dfd, const char __user *filename, int flags,
2052 extern struct file *file_open_name(struct filename *, int, umode_t);
2053 extern struct file *filp_open(const char *, int, umode_t);
2054 extern struct file *file_open_root(struct dentry *, struct vfsmount *,
2055 const char *, int, umode_t);
2056 extern struct file * dentry_open(const struct path *, int, const struct cred *);
2057 extern int filp_close(struct file *, fl_owner_t id);
2059 extern struct filename *getname(const char __user *);
2060 extern struct filename *getname_kernel(const char *);
2066 extern int finish_open(struct file *file, struct dentry *dentry,
2067 int (*open)(struct inode *, struct file *),
2069 extern int finish_no_open(struct file *file, struct dentry *dentry);
2073 extern int ioctl_preallocate(struct file *filp, void __user *argp);
2076 extern void __init vfs_caches_init_early(void);
2077 extern void __init vfs_caches_init(unsigned long);
2079 extern struct kmem_cache *names_cachep;
2081 extern void final_putname(struct filename *name);
2083 #define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL)
2084 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
2085 #ifndef CONFIG_AUDITSYSCALL
2086 #define putname(name) final_putname(name)
2088 extern void putname(struct filename *name);
2092 extern int register_blkdev(unsigned int, const char *);
2093 extern void unregister_blkdev(unsigned int, const char *);
2094 extern struct block_device *bdget(dev_t);
2095 extern struct block_device *bdgrab(struct block_device *bdev);
2096 extern void bd_set_size(struct block_device *, loff_t size);
2097 extern void bd_forget(struct inode *inode);
2098 extern void bdput(struct block_device *);
2099 extern void invalidate_bdev(struct block_device *);
2100 extern void iterate_bdevs(void (*)(struct block_device *, void *), void *);
2101 extern int sync_blockdev(struct block_device *bdev);
2102 extern void kill_bdev(struct block_device *);
2103 extern struct super_block *freeze_bdev(struct block_device *);
2104 extern void emergency_thaw_all(void);
2105 extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
2106 extern int fsync_bdev(struct block_device *);
2107 extern int sb_is_blkdev_sb(struct super_block *sb);
2109 static inline void bd_forget(struct inode *inode) {}
2110 static inline int sync_blockdev(struct block_device *bdev) { return 0; }
2111 static inline void kill_bdev(struct block_device *bdev) {}
2112 static inline void invalidate_bdev(struct block_device *bdev) {}
2114 static inline struct super_block *freeze_bdev(struct block_device *sb)
2119 static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
2124 static inline void iterate_bdevs(void (*f)(struct block_device *, void *), void *arg)
2128 static inline int sb_is_blkdev_sb(struct super_block *sb)
2133 extern int sync_filesystem(struct super_block *);
2134 extern const struct file_operations def_blk_fops;
2135 extern const struct file_operations def_chr_fops;
2136 extern const struct file_operations bad_sock_fops;
2138 extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
2139 extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
2140 extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
2141 extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
2142 extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
2144 extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
2146 extern void blkdev_put(struct block_device *bdev, fmode_t mode);
2148 extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
2149 extern void bd_unlink_disk_holder(struct block_device *bdev,
2150 struct gendisk *disk);
2152 static inline int bd_link_disk_holder(struct block_device *bdev,
2153 struct gendisk *disk)
2157 static inline void bd_unlink_disk_holder(struct block_device *bdev,
2158 struct gendisk *disk)
2165 #define CHRDEV_MAJOR_HASH_SIZE 255
2166 extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
2167 extern int register_chrdev_region(dev_t, unsigned, const char *);
2168 extern int __register_chrdev(unsigned int major, unsigned int baseminor,
2169 unsigned int count, const char *name,
2170 const struct file_operations *fops);
2171 extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
2172 unsigned int count, const char *name);
2173 extern void unregister_chrdev_region(dev_t, unsigned);
2174 extern void chrdev_show(struct seq_file *,off_t);
2176 static inline int register_chrdev(unsigned int major, const char *name,
2177 const struct file_operations *fops)
2179 return __register_chrdev(major, 0, 256, name, fops);
2182 static inline void unregister_chrdev(unsigned int major, const char *name)
2184 __unregister_chrdev(major, 0, 256, name);
2187 /* fs/block_dev.c */
2188 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
2189 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
2192 #define BLKDEV_MAJOR_HASH_SIZE 255
2193 extern const char *__bdevname(dev_t, char *buffer);
2194 extern const char *bdevname(struct block_device *bdev, char *buffer);
2195 extern struct block_device *lookup_bdev(const char *);
2196 extern void blkdev_show(struct seq_file *,off_t);
2199 #define BLKDEV_MAJOR_HASH_SIZE 0
2202 extern void init_special_inode(struct inode *, umode_t, dev_t);
2204 /* Invalid inode operations -- fs/bad_inode.c */
2205 extern void make_bad_inode(struct inode *);
2206 extern int is_bad_inode(struct inode *);
2210 * return READ, READA, or WRITE
2212 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2215 * return data direction, READ or WRITE
2217 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
2219 extern void check_disk_size_change(struct gendisk *disk,
2220 struct block_device *bdev);
2221 extern int revalidate_disk(struct gendisk *);
2222 extern int check_disk_change(struct block_device *);
2223 extern int __invalidate_device(struct block_device *, bool);
2224 extern int invalidate_partition(struct gendisk *, int);
2226 unsigned long invalidate_mapping_pages(struct address_space *mapping,
2227 pgoff_t start, pgoff_t end);
2229 static inline void invalidate_remote_inode(struct inode *inode)
2231 if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
2232 S_ISLNK(inode->i_mode))
2233 invalidate_mapping_pages(inode->i_mapping, 0, -1);
2235 extern int invalidate_inode_pages2(struct address_space *mapping);
2236 extern int invalidate_inode_pages2_range(struct address_space *mapping,
2237 pgoff_t start, pgoff_t end);
2238 extern int write_inode_now(struct inode *, int);
2239 extern int filemap_fdatawrite(struct address_space *);
2240 extern int filemap_flush(struct address_space *);
2241 extern int filemap_fdatawait(struct address_space *);
2242 extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
2244 extern int filemap_write_and_wait(struct address_space *mapping);
2245 extern int filemap_write_and_wait_range(struct address_space *mapping,
2246 loff_t lstart, loff_t lend);
2247 extern int __filemap_fdatawrite_range(struct address_space *mapping,
2248 loff_t start, loff_t end, int sync_mode);
2249 extern int filemap_fdatawrite_range(struct address_space *mapping,
2250 loff_t start, loff_t end);
2252 extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
2254 extern int vfs_fsync(struct file *file, int datasync);
2255 static inline int generic_write_sync(struct file *file, loff_t pos, loff_t count)
2257 if (!(file->f_flags & O_DSYNC) && !IS_SYNC(file->f_mapping->host))
2259 return vfs_fsync_range(file, pos, pos + count - 1,
2260 (file->f_flags & __O_SYNC) ? 0 : 1);
2262 extern void emergency_sync(void);
2263 extern void emergency_remount(void);
2265 extern sector_t bmap(struct inode *, sector_t);
2267 extern int notify_change(struct dentry *, struct iattr *, struct inode **);
2268 extern int inode_permission(struct inode *, int);
2269 extern int generic_permission(struct inode *, int);
2271 static inline bool execute_ok(struct inode *inode)
2273 return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
2276 static inline void file_start_write(struct file *file)
2278 if (!S_ISREG(file_inode(file)->i_mode))
2280 __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, true);
2283 static inline bool file_start_write_trylock(struct file *file)
2285 if (!S_ISREG(file_inode(file)->i_mode))
2287 return __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, false);
2290 static inline void file_end_write(struct file *file)
2292 if (!S_ISREG(file_inode(file)->i_mode))
2294 __sb_end_write(file_inode(file)->i_sb, SB_FREEZE_WRITE);
2298 * get_write_access() gets write permission for a file.
2299 * put_write_access() releases this write permission.
2300 * This is used for regular files.
2301 * We cannot support write (and maybe mmap read-write shared) accesses and
2302 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2303 * can have the following values:
2304 * 0: no writers, no VM_DENYWRITE mappings
2305 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2306 * > 0: (i_writecount) users are writing to the file.
2308 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2309 * except for the cases where we don't hold i_writecount yet. Then we need to
2310 * use {get,deny}_write_access() - these functions check the sign and refuse
2311 * to do the change if sign is wrong.
2313 static inline int get_write_access(struct inode *inode)
2315 return atomic_inc_unless_negative(&inode->i_writecount) ? 0 : -ETXTBSY;
2317 static inline int deny_write_access(struct file *file)
2319 struct inode *inode = file_inode(file);
2320 return atomic_dec_unless_positive(&inode->i_writecount) ? 0 : -ETXTBSY;
2322 static inline void put_write_access(struct inode * inode)
2324 atomic_dec(&inode->i_writecount);
2326 static inline void allow_write_access(struct file *file)
2329 atomic_inc(&file_inode(file)->i_writecount);
2331 static inline bool inode_is_open_for_write(const struct inode *inode)
2333 return atomic_read(&inode->i_writecount) > 0;
2337 static inline void i_readcount_dec(struct inode *inode)
2339 BUG_ON(!atomic_read(&inode->i_readcount));
2340 atomic_dec(&inode->i_readcount);
2342 static inline void i_readcount_inc(struct inode *inode)
2344 atomic_inc(&inode->i_readcount);
2347 static inline void i_readcount_dec(struct inode *inode)
2351 static inline void i_readcount_inc(struct inode *inode)
2356 extern int do_pipe_flags(int *, int);
2358 extern int kernel_read(struct file *, loff_t, char *, unsigned long);
2359 extern ssize_t kernel_write(struct file *, const char *, size_t, loff_t);
2360 extern struct file * open_exec(const char *);
2362 /* fs/dcache.c -- generic fs support functions */
2363 extern int is_subdir(struct dentry *, struct dentry *);
2364 extern int path_is_under(struct path *, struct path *);
2366 #include <linux/err.h>
2368 /* needed for stackable file system support */
2369 extern loff_t default_llseek(struct file *file, loff_t offset, int whence);
2371 extern loff_t vfs_llseek(struct file *file, loff_t offset, int whence);
2373 extern int inode_init_always(struct super_block *, struct inode *);
2374 extern void inode_init_once(struct inode *);
2375 extern void address_space_init_once(struct address_space *mapping);
2376 extern struct inode * igrab(struct inode *);
2377 extern ino_t iunique(struct super_block *, ino_t);
2378 extern int inode_needs_sync(struct inode *inode);
2379 extern int generic_delete_inode(struct inode *inode);
2380 static inline int generic_drop_inode(struct inode *inode)
2382 return !inode->i_nlink || inode_unhashed(inode);
2385 extern struct inode *ilookup5_nowait(struct super_block *sb,
2386 unsigned long hashval, int (*test)(struct inode *, void *),
2388 extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
2389 int (*test)(struct inode *, void *), void *data);
2390 extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
2392 extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
2393 extern struct inode * iget_locked(struct super_block *, unsigned long);
2394 extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
2395 extern int insert_inode_locked(struct inode *);
2396 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2397 extern void lockdep_annotate_inode_mutex_key(struct inode *inode);
2399 static inline void lockdep_annotate_inode_mutex_key(struct inode *inode) { };
2401 extern void unlock_new_inode(struct inode *);
2402 extern unsigned int get_next_ino(void);
2404 extern void __iget(struct inode * inode);
2405 extern void iget_failed(struct inode *);
2406 extern void clear_inode(struct inode *);
2407 extern void __destroy_inode(struct inode *);
2408 extern struct inode *new_inode_pseudo(struct super_block *sb);
2409 extern struct inode *new_inode(struct super_block *sb);
2410 extern void free_inode_nonrcu(struct inode *inode);
2411 extern int should_remove_suid(struct dentry *);
2412 extern int file_remove_suid(struct file *);
2414 extern void __insert_inode_hash(struct inode *, unsigned long hashval);
2415 static inline void insert_inode_hash(struct inode *inode)
2417 __insert_inode_hash(inode, inode->i_ino);
2420 extern void __remove_inode_hash(struct inode *);
2421 static inline void remove_inode_hash(struct inode *inode)
2423 if (!inode_unhashed(inode))
2424 __remove_inode_hash(inode);
2427 extern void inode_sb_list_add(struct inode *inode);
2430 extern void submit_bio(int, struct bio *);
2431 extern int bdev_read_only(struct block_device *);
2433 extern int set_blocksize(struct block_device *, int);
2434 extern int sb_set_blocksize(struct super_block *, int);
2435 extern int sb_min_blocksize(struct super_block *, int);
2437 extern int generic_file_mmap(struct file *, struct vm_area_struct *);
2438 extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
2439 int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
2440 extern ssize_t generic_file_read_iter(struct kiocb *, struct iov_iter *);
2441 extern ssize_t __generic_file_write_iter(struct kiocb *, struct iov_iter *);
2442 extern ssize_t generic_file_write_iter(struct kiocb *, struct iov_iter *);
2443 extern ssize_t generic_file_direct_write(struct kiocb *, struct iov_iter *, loff_t);
2444 extern ssize_t generic_perform_write(struct file *, struct iov_iter *, loff_t);
2445 extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
2446 extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
2447 extern ssize_t new_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
2448 extern ssize_t new_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
2450 /* fs/block_dev.c */
2451 extern ssize_t blkdev_write_iter(struct kiocb *iocb, struct iov_iter *from);
2452 extern int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
2454 extern void block_sync_page(struct page *page);
2457 extern ssize_t generic_file_splice_read(struct file *, loff_t *,
2458 struct pipe_inode_info *, size_t, unsigned int);
2459 extern ssize_t default_file_splice_read(struct file *, loff_t *,
2460 struct pipe_inode_info *, size_t, unsigned int);
2461 extern ssize_t iter_file_splice_write(struct pipe_inode_info *,
2462 struct file *, loff_t *, size_t, unsigned int);
2463 extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
2464 struct file *out, loff_t *, size_t len, unsigned int flags);
2467 file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
2468 extern loff_t noop_llseek(struct file *file, loff_t offset, int whence);
2469 extern loff_t no_llseek(struct file *file, loff_t offset, int whence);
2470 extern loff_t vfs_setpos(struct file *file, loff_t offset, loff_t maxsize);
2471 extern loff_t generic_file_llseek(struct file *file, loff_t offset, int whence);
2472 extern loff_t generic_file_llseek_size(struct file *file, loff_t offset,
2473 int whence, loff_t maxsize, loff_t eof);
2474 extern loff_t fixed_size_llseek(struct file *file, loff_t offset,
2475 int whence, loff_t size);
2476 extern int generic_file_open(struct inode * inode, struct file * filp);
2477 extern int nonseekable_open(struct inode * inode, struct file * filp);
2478 extern int stream_open(struct inode * inode, struct file * filp);
2480 #ifdef CONFIG_FS_XIP
2481 extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
2483 extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
2484 extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
2485 size_t len, loff_t *ppos);
2486 extern int xip_truncate_page(struct address_space *mapping, loff_t from);
2488 static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
2495 typedef void (dio_submit_t)(int rw, struct bio *bio, struct inode *inode,
2496 loff_t file_offset);
2499 /* need locking between buffered and direct access */
2502 /* filesystem does not support filling holes */
2503 DIO_SKIP_HOLES = 0x02,
2505 /* filesystem can handle aio writes beyond i_size */
2506 DIO_ASYNC_EXTEND = 0x04,
2509 void dio_end_io(struct bio *bio, int error);
2511 ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
2512 struct block_device *bdev, struct iov_iter *iter, loff_t offset,
2513 get_block_t get_block, dio_iodone_t end_io,
2514 dio_submit_t submit_io, int flags);
2516 static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
2517 struct inode *inode, struct iov_iter *iter, loff_t offset,
2518 get_block_t get_block)
2520 return __blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iter,
2521 offset, get_block, NULL, NULL,
2522 DIO_LOCKING | DIO_SKIP_HOLES);
2526 void inode_dio_wait(struct inode *inode);
2527 void inode_dio_done(struct inode *inode);
2529 extern void inode_set_flags(struct inode *inode, unsigned int flags,
2532 extern const struct file_operations generic_ro_fops;
2534 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2536 extern int readlink_copy(char __user *, int, const char *);
2537 extern int page_readlink(struct dentry *, char __user *, int);
2538 extern void *page_follow_link_light(struct dentry *, struct nameidata *);
2539 extern void page_put_link(struct dentry *, struct nameidata *, void *);
2540 extern int __page_symlink(struct inode *inode, const char *symname, int len,
2542 extern int page_symlink(struct inode *inode, const char *symname, int len);
2543 extern const struct inode_operations page_symlink_inode_operations;
2544 extern void kfree_put_link(struct dentry *, struct nameidata *, void *);
2545 extern int generic_readlink(struct dentry *, char __user *, int);
2546 extern void generic_fillattr(struct inode *, struct kstat *);
2547 int vfs_getattr_nosec(struct path *path, struct kstat *stat);
2548 extern int vfs_getattr(struct path *, struct kstat *);
2549 void __inode_add_bytes(struct inode *inode, loff_t bytes);
2550 void inode_add_bytes(struct inode *inode, loff_t bytes);
2551 void __inode_sub_bytes(struct inode *inode, loff_t bytes);
2552 void inode_sub_bytes(struct inode *inode, loff_t bytes);
2553 loff_t inode_get_bytes(struct inode *inode);
2554 void inode_set_bytes(struct inode *inode, loff_t bytes);
2556 extern int vfs_readdir(struct file *, filldir_t, void *);
2557 extern int iterate_dir(struct file *, struct dir_context *);
2559 extern int vfs_stat(const char __user *, struct kstat *);
2560 extern int vfs_lstat(const char __user *, struct kstat *);
2561 extern int vfs_fstat(unsigned int, struct kstat *);
2562 extern int vfs_fstatat(int , const char __user *, struct kstat *, int);
2564 extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
2566 extern int __generic_block_fiemap(struct inode *inode,
2567 struct fiemap_extent_info *fieinfo,
2568 loff_t start, loff_t len,
2569 get_block_t *get_block);
2570 extern int generic_block_fiemap(struct inode *inode,
2571 struct fiemap_extent_info *fieinfo, u64 start,
2572 u64 len, get_block_t *get_block);
2574 extern void get_filesystem(struct file_system_type *fs);
2575 extern void put_filesystem(struct file_system_type *fs);
2576 extern struct file_system_type *get_fs_type(const char *name);
2577 extern struct super_block *get_super(struct block_device *);
2578 extern struct super_block *get_super_thawed(struct block_device *);
2579 extern struct super_block *get_active_super(struct block_device *bdev);
2580 extern void drop_super(struct super_block *sb);
2581 extern void iterate_supers(void (*)(struct super_block *, void *), void *);
2582 extern void iterate_supers_type(struct file_system_type *,
2583 void (*)(struct super_block *, void *), void *);
2585 extern int dcache_dir_open(struct inode *, struct file *);
2586 extern int dcache_dir_close(struct inode *, struct file *);
2587 extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
2588 extern int dcache_readdir(struct file *, struct dir_context *);
2589 extern int simple_setattr(struct dentry *, struct iattr *);
2590 extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2591 extern int simple_statfs(struct dentry *, struct kstatfs *);
2592 extern int simple_open(struct inode *inode, struct file *file);
2593 extern int simple_link(struct dentry *, struct inode *, struct dentry *);
2594 extern int simple_unlink(struct inode *, struct dentry *);
2595 extern int simple_rmdir(struct inode *, struct dentry *);
2596 extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
2597 extern int noop_fsync(struct file *, loff_t, loff_t, int);
2598 extern int simple_empty(struct dentry *);
2599 extern int simple_readpage(struct file *file, struct page *page);
2600 extern int simple_write_begin(struct file *file, struct address_space *mapping,
2601 loff_t pos, unsigned len, unsigned flags,
2602 struct page **pagep, void **fsdata);
2603 extern int simple_write_end(struct file *file, struct address_space *mapping,
2604 loff_t pos, unsigned len, unsigned copied,
2605 struct page *page, void *fsdata);
2606 extern int always_delete_dentry(const struct dentry *);
2607 extern struct inode *alloc_anon_inode(struct super_block *);
2608 extern const struct dentry_operations simple_dentry_operations;
2610 extern struct dentry *simple_lookup(struct inode *, struct dentry *, unsigned int flags);
2611 extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
2612 extern const struct file_operations simple_dir_operations;
2613 extern const struct inode_operations simple_dir_inode_operations;
2614 extern void make_empty_dir_inode(struct inode *inode);
2615 extern bool is_empty_dir_inode(struct inode *inode);
2616 struct tree_descr { char *name; const struct file_operations *ops; int mode; };
2617 struct dentry *d_alloc_name(struct dentry *, const char *);
2618 extern int simple_fill_super(struct super_block *, unsigned long, struct tree_descr *);
2619 extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
2620 extern void simple_release_fs(struct vfsmount **mount, int *count);
2622 extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
2623 loff_t *ppos, const void *from, size_t available);
2624 extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
2625 const void __user *from, size_t count);
2627 extern int __generic_file_fsync(struct file *, loff_t, loff_t, int);
2628 extern int generic_file_fsync(struct file *, loff_t, loff_t, int);
2630 extern int generic_check_addressable(unsigned, u64);
2632 #ifdef CONFIG_MIGRATION
2633 extern int buffer_migrate_page(struct address_space *,
2634 struct page *, struct page *,
2637 #define buffer_migrate_page NULL
2640 extern int setattr_prepare(struct dentry *, struct iattr *);
2641 extern int inode_newsize_ok(const struct inode *, loff_t offset);
2642 extern void setattr_copy(struct inode *inode, const struct iattr *attr);
2644 extern int file_update_time(struct file *file);
2646 extern int generic_show_options(struct seq_file *m, struct dentry *root);
2647 extern void save_mount_options(struct super_block *sb, char *options);
2648 extern void replace_mount_options(struct super_block *sb, char *options);
2650 static inline ino_t parent_ino(struct dentry *dentry)
2655 * Don't strictly need d_lock here? If the parent ino could change
2656 * then surely we'd have a deeper race in the caller?
2658 spin_lock(&dentry->d_lock);
2659 res = dentry->d_parent->d_inode->i_ino;
2660 spin_unlock(&dentry->d_lock);
2664 /* Transaction based IO helpers */
2667 * An argresp is stored in an allocated page and holds the
2668 * size of the argument or response, along with its content
2670 struct simple_transaction_argresp {
2675 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2677 char *simple_transaction_get(struct file *file, const char __user *buf,
2679 ssize_t simple_transaction_read(struct file *file, char __user *buf,
2680 size_t size, loff_t *pos);
2681 int simple_transaction_release(struct inode *inode, struct file *file);
2683 void simple_transaction_set(struct file *file, size_t n);
2686 * simple attribute files
2688 * These attributes behave similar to those in sysfs:
2690 * Writing to an attribute immediately sets a value, an open file can be
2691 * written to multiple times.
2693 * Reading from an attribute creates a buffer from the value that might get
2694 * read with multiple read calls. When the attribute has been read
2695 * completely, no further read calls are possible until the file is opened
2698 * All attributes contain a text representation of a numeric value
2699 * that are accessed with the get() and set() functions.
2701 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2702 static int __fops ## _open(struct inode *inode, struct file *file) \
2704 __simple_attr_check_format(__fmt, 0ull); \
2705 return simple_attr_open(inode, file, __get, __set, __fmt); \
2707 static const struct file_operations __fops = { \
2708 .owner = THIS_MODULE, \
2709 .open = __fops ## _open, \
2710 .release = simple_attr_release, \
2711 .read = simple_attr_read, \
2712 .write = simple_attr_write, \
2713 .llseek = generic_file_llseek, \
2716 static inline __printf(1, 2)
2717 void __simple_attr_check_format(const char *fmt, ...)
2719 /* don't do anything, just let the compiler check the arguments; */
2722 int simple_attr_open(struct inode *inode, struct file *file,
2723 int (*get)(void *, u64 *), int (*set)(void *, u64),
2725 int simple_attr_release(struct inode *inode, struct file *file);
2726 ssize_t simple_attr_read(struct file *file, char __user *buf,
2727 size_t len, loff_t *ppos);
2728 ssize_t simple_attr_write(struct file *file, const char __user *buf,
2729 size_t len, loff_t *ppos);
2732 int proc_nr_files(struct ctl_table *table, int write,
2733 void __user *buffer, size_t *lenp, loff_t *ppos);
2734 int proc_nr_dentry(struct ctl_table *table, int write,
2735 void __user *buffer, size_t *lenp, loff_t *ppos);
2736 int proc_nr_inodes(struct ctl_table *table, int write,
2737 void __user *buffer, size_t *lenp, loff_t *ppos);
2738 int __init get_filesystem_list(char *buf);
2740 #define __FMODE_EXEC ((__force int) FMODE_EXEC)
2741 #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
2743 #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
2744 #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
2745 (flag & __FMODE_NONOTIFY)))
2747 static inline int is_sxid(umode_t mode)
2749 return (mode & S_ISUID) || ((mode & S_ISGID) && (mode & S_IXGRP));
2752 static inline void inode_has_no_xattr(struct inode *inode)
2754 if (!is_sxid(inode->i_mode) && (inode->i_sb->s_flags & MS_NOSEC))
2755 inode->i_flags |= S_NOSEC;
2758 static inline bool dir_emit(struct dir_context *ctx,
2759 const char *name, int namelen,
2760 u64 ino, unsigned type)
2762 return ctx->actor(ctx, name, namelen, ctx->pos, ino, type) == 0;
2764 static inline bool dir_emit_dot(struct file *file, struct dir_context *ctx)
2766 return ctx->actor(ctx, ".", 1, ctx->pos,
2767 file->f_path.dentry->d_inode->i_ino, DT_DIR) == 0;
2769 static inline bool dir_emit_dotdot(struct file *file, struct dir_context *ctx)
2771 return ctx->actor(ctx, "..", 2, ctx->pos,
2772 parent_ino(file->f_path.dentry), DT_DIR) == 0;
2774 static inline bool dir_emit_dots(struct file *file, struct dir_context *ctx)
2776 if (ctx->pos == 0) {
2777 if (!dir_emit_dot(file, ctx))
2781 if (ctx->pos == 1) {
2782 if (!dir_emit_dotdot(file, ctx))
2788 static inline bool dir_relax(struct inode *inode)
2790 mutex_unlock(&inode->i_mutex);
2791 mutex_lock(&inode->i_mutex);
2792 return !IS_DEADDIR(inode);
2795 extern bool path_noexec(const struct path *path);
2797 #endif /* _LINUX_FS_H */