5 #include <linux/linkage.h>
6 #include <linux/wait.h>
7 #include <linux/kdev_t.h>
8 #include <linux/dcache.h>
9 #include <linux/path.h>
10 #include <linux/stat.h>
11 #include <linux/cache.h>
12 #include <linux/list.h>
13 #include <linux/list_lru.h>
14 #include <linux/llist.h>
15 #include <linux/radix-tree.h>
16 #include <linux/rbtree.h>
17 #include <linux/init.h>
18 #include <linux/pid.h>
19 #include <linux/bug.h>
20 #include <linux/mutex.h>
21 #include <linux/capability.h>
22 #include <linux/semaphore.h>
23 #include <linux/fiemap.h>
24 #include <linux/rculist_bl.h>
25 #include <linux/atomic.h>
26 #include <linux/shrinker.h>
27 #include <linux/migrate_mode.h>
28 #include <linux/uidgid.h>
29 #include <linux/lockdep.h>
30 #include <linux/percpu-rwsem.h>
31 #include <linux/blk_types.h>
33 #include <asm/byteorder.h>
34 #include <uapi/linux/fs.h>
36 struct export_operations;
42 struct pipe_inode_info;
43 struct poll_table_struct;
45 struct vm_area_struct;
48 struct swap_info_struct;
50 struct workqueue_struct;
52 extern void __init inode_init(void);
53 extern void __init inode_init_early(void);
54 extern void __init files_init(unsigned long);
56 extern struct files_stat_struct files_stat;
57 extern unsigned long get_max_files(void);
58 extern int sysctl_nr_open;
59 extern struct inodes_stat_t inodes_stat;
60 extern int leases_enable, lease_break_time;
61 extern int sysctl_protected_symlinks;
62 extern int sysctl_protected_hardlinks;
65 typedef int (get_block_t)(struct inode *inode, sector_t iblock,
66 struct buffer_head *bh_result, int create);
67 typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
68 ssize_t bytes, void *private);
70 #define MAY_EXEC 0x00000001
71 #define MAY_WRITE 0x00000002
72 #define MAY_READ 0x00000004
73 #define MAY_APPEND 0x00000008
74 #define MAY_ACCESS 0x00000010
75 #define MAY_OPEN 0x00000020
76 #define MAY_CHDIR 0x00000040
77 /* called from RCU mode, don't block */
78 #define MAY_NOT_BLOCK 0x00000080
81 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
82 * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
85 /* file is open for reading */
86 #define FMODE_READ ((__force fmode_t)0x1)
87 /* file is open for writing */
88 #define FMODE_WRITE ((__force fmode_t)0x2)
89 /* file is seekable */
90 #define FMODE_LSEEK ((__force fmode_t)0x4)
91 /* file can be accessed using pread */
92 #define FMODE_PREAD ((__force fmode_t)0x8)
93 /* file can be accessed using pwrite */
94 #define FMODE_PWRITE ((__force fmode_t)0x10)
95 /* File is opened for execution with sys_execve / sys_uselib */
96 #define FMODE_EXEC ((__force fmode_t)0x20)
97 /* File is opened with O_NDELAY (only set for block devices) */
98 #define FMODE_NDELAY ((__force fmode_t)0x40)
99 /* File is opened with O_EXCL (only set for block devices) */
100 #define FMODE_EXCL ((__force fmode_t)0x80)
101 /* File is opened using open(.., 3, ..) and is writeable only for ioctls
102 (specialy hack for floppy.c) */
103 #define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
104 /* 32bit hashes as llseek() offset (for directories) */
105 #define FMODE_32BITHASH ((__force fmode_t)0x200)
106 /* 64bit hashes as llseek() offset (for directories) */
107 #define FMODE_64BITHASH ((__force fmode_t)0x400)
110 * Don't update ctime and mtime.
112 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
113 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
115 #define FMODE_NOCMTIME ((__force fmode_t)0x800)
117 /* Expect random access pattern */
118 #define FMODE_RANDOM ((__force fmode_t)0x1000)
120 /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
121 #define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
123 /* File is opened with O_PATH; almost nothing can be done with it */
124 #define FMODE_PATH ((__force fmode_t)0x4000)
126 /* File needs atomic accesses to f_pos */
127 #define FMODE_ATOMIC_POS ((__force fmode_t)0x8000)
128 /* Write access to underlying fs */
129 #define FMODE_WRITER ((__force fmode_t)0x10000)
131 /* File was opened by fanotify and shouldn't generate fanotify events */
132 #define FMODE_NONOTIFY ((__force fmode_t)0x1000000)
135 * Flag for rw_copy_check_uvector and compat_rw_copy_check_uvector
136 * that indicates that they should check the contents of the iovec are
137 * valid, but not check the memory that the iovec elements
140 #define CHECK_IOVEC_ONLY -1
143 * The below are the various read and write types that we support. Some of
144 * them include behavioral modifiers that send information down to the
145 * block layer and IO scheduler. Terminology:
147 * The block layer uses device plugging to defer IO a little bit, in
148 * the hope that we will see more IO very shortly. This increases
149 * coalescing of adjacent IO and thus reduces the number of IOs we
150 * have to send to the device. It also allows for better queuing,
151 * if the IO isn't mergeable. If the caller is going to be waiting
152 * for the IO, then he must ensure that the device is unplugged so
153 * that the IO is dispatched to the driver.
155 * All IO is handled async in Linux. This is fine for background
156 * writes, but for reads or writes that someone waits for completion
157 * on, we want to notify the block layer and IO scheduler so that they
158 * know about it. That allows them to make better scheduling
159 * decisions. So when the below references 'sync' and 'async', it
160 * is referencing this priority hint.
162 * With that in mind, the available types are:
164 * READ A normal read operation. Device will be plugged.
165 * READ_SYNC A synchronous read. Device is not plugged, caller can
166 * immediately wait on this read without caring about
168 * READA Used for read-ahead operations. Lower priority, and the
169 * block layer could (in theory) choose to ignore this
170 * request if it runs into resource problems.
171 * WRITE A normal async write. Device will be plugged.
172 * WRITE_SYNC Synchronous write. Identical to WRITE, but passes down
173 * the hint that someone will be waiting on this IO
174 * shortly. The write equivalent of READ_SYNC.
175 * WRITE_ODIRECT Special case write for O_DIRECT only.
176 * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
177 * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
178 * non-volatile media on completion.
179 * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
180 * by a cache flush and data is guaranteed to be on
181 * non-volatile media on completion.
184 #define RW_MASK REQ_WRITE
185 #define RWA_MASK REQ_RAHEAD
188 #define WRITE RW_MASK
189 #define READA RWA_MASK
190 #define KERNEL_READ (READ|REQ_KERNEL)
191 #define KERNEL_WRITE (WRITE|REQ_KERNEL)
193 #define READ_SYNC (READ | REQ_SYNC)
194 #define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE)
195 #define WRITE_ODIRECT (WRITE | REQ_SYNC)
196 #define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
197 #define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
198 #define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
201 * Attribute flags. These should be or-ed together to figure out what
204 #define ATTR_MODE (1 << 0)
205 #define ATTR_UID (1 << 1)
206 #define ATTR_GID (1 << 2)
207 #define ATTR_SIZE (1 << 3)
208 #define ATTR_ATIME (1 << 4)
209 #define ATTR_MTIME (1 << 5)
210 #define ATTR_CTIME (1 << 6)
211 #define ATTR_ATIME_SET (1 << 7)
212 #define ATTR_MTIME_SET (1 << 8)
213 #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
214 #define ATTR_ATTR_FLAG (1 << 10)
215 #define ATTR_KILL_SUID (1 << 11)
216 #define ATTR_KILL_SGID (1 << 12)
217 #define ATTR_FILE (1 << 13)
218 #define ATTR_KILL_PRIV (1 << 14)
219 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
220 #define ATTR_TIMES_SET (1 << 16)
223 * This is the Inode Attributes structure, used for notify_change(). It
224 * uses the above definitions as flags, to know which values have changed.
225 * Also, in this manner, a Filesystem can look at only the values it cares
226 * about. Basically, these are the attributes that the VFS layer can
227 * request to change from the FS layer.
229 * Derek Atkins <warlord@MIT.EDU> 94-10-20
232 unsigned int ia_valid;
237 struct timespec ia_atime;
238 struct timespec ia_mtime;
239 struct timespec ia_ctime;
242 * Not an attribute, but an auxiliary info for filesystems wanting to
243 * implement an ftruncate() like method. NOTE: filesystem should
244 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
246 struct file *ia_file;
250 * Includes for diskquotas.
252 #include <linux/quota.h>
255 * enum positive_aop_returns - aop return codes with specific semantics
257 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
258 * completed, that the page is still locked, and
259 * should be considered active. The VM uses this hint
260 * to return the page to the active list -- it won't
261 * be a candidate for writeback again in the near
262 * future. Other callers must be careful to unlock
263 * the page if they get this return. Returned by
266 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
267 * unlocked it and the page might have been truncated.
268 * The caller should back up to acquiring a new page and
269 * trying again. The aop will be taking reasonable
270 * precautions not to livelock. If the caller held a page
271 * reference, it should drop it before retrying. Returned
274 * address_space_operation functions return these large constants to indicate
275 * special semantics to the caller. These are much larger than the bytes in a
276 * page to allow for functions that return the number of bytes operated on in a
280 enum positive_aop_returns {
281 AOP_WRITEPAGE_ACTIVATE = 0x80000,
282 AOP_TRUNCATED_PAGE = 0x80001,
285 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
286 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
287 #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
288 * helper code (eg buffer layer)
289 * to clear GFP_FS from alloc */
292 * oh the beauties of C type declarations.
295 struct address_space;
296 struct writeback_control;
299 * "descriptor" for what we're up to with a read.
300 * This allows us to use the same read code yet
301 * have multiple different users of the data that
302 * we read from a file.
304 * The simplest case just copies the data to user
317 typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
318 unsigned long, unsigned long);
320 struct address_space_operations {
321 int (*writepage)(struct page *page, struct writeback_control *wbc);
322 int (*readpage)(struct file *, struct page *);
324 /* Write back some dirty pages from this mapping. */
325 int (*writepages)(struct address_space *, struct writeback_control *);
327 /* Set a page dirty. Return true if this dirtied it */
328 int (*set_page_dirty)(struct page *page);
330 int (*readpages)(struct file *filp, struct address_space *mapping,
331 struct list_head *pages, unsigned nr_pages);
333 int (*write_begin)(struct file *, struct address_space *mapping,
334 loff_t pos, unsigned len, unsigned flags,
335 struct page **pagep, void **fsdata);
336 int (*write_end)(struct file *, struct address_space *mapping,
337 loff_t pos, unsigned len, unsigned copied,
338 struct page *page, void *fsdata);
340 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
341 sector_t (*bmap)(struct address_space *, sector_t);
342 void (*invalidatepage) (struct page *, unsigned int, unsigned int);
343 int (*releasepage) (struct page *, gfp_t);
344 void (*freepage)(struct page *);
345 ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
346 loff_t offset, unsigned long nr_segs);
347 int (*get_xip_mem)(struct address_space *, pgoff_t, int,
348 void **, unsigned long *);
350 * migrate the contents of a page to the specified target. If
351 * migrate_mode is MIGRATE_ASYNC, it must not block.
353 int (*migratepage) (struct address_space *,
354 struct page *, struct page *, enum migrate_mode);
355 int (*launder_page) (struct page *);
356 int (*is_partially_uptodate) (struct page *, unsigned long,
358 void (*is_dirty_writeback) (struct page *, bool *, bool *);
359 int (*error_remove_page)(struct address_space *, struct page *);
361 /* swapfile support */
362 int (*swap_activate)(struct swap_info_struct *sis, struct file *file,
364 void (*swap_deactivate)(struct file *file);
367 extern const struct address_space_operations empty_aops;
370 * pagecache_write_begin/pagecache_write_end must be used by general code
371 * to write into the pagecache.
373 int pagecache_write_begin(struct file *, struct address_space *mapping,
374 loff_t pos, unsigned len, unsigned flags,
375 struct page **pagep, void **fsdata);
377 int pagecache_write_end(struct file *, struct address_space *mapping,
378 loff_t pos, unsigned len, unsigned copied,
379 struct page *page, void *fsdata);
381 struct backing_dev_info;
382 struct address_space {
383 struct inode *host; /* owner: inode, block_device */
384 struct radix_tree_root page_tree; /* radix tree of all pages */
385 spinlock_t tree_lock; /* and lock protecting it */
386 unsigned int i_mmap_writable;/* count VM_SHARED mappings */
387 struct rb_root i_mmap; /* tree of private and shared mappings */
388 struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
389 struct mutex i_mmap_mutex; /* protect tree, count, list */
390 /* Protected by tree_lock together with the radix tree */
391 unsigned long nrpages; /* number of total pages */
392 pgoff_t writeback_index;/* writeback starts here */
393 const struct address_space_operations *a_ops; /* methods */
394 unsigned long flags; /* error bits/gfp mask */
395 struct backing_dev_info *backing_dev_info; /* device readahead, etc */
396 spinlock_t private_lock; /* for use by the address_space */
397 struct list_head private_list; /* ditto */
398 void *private_data; /* ditto */
399 } __attribute__((aligned(sizeof(long))));
401 * On most architectures that alignment is already the case; but
402 * must be enforced here for CRIS, to let the least significant bit
403 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
405 struct request_queue;
407 struct block_device {
408 dev_t bd_dev; /* not a kdev_t - it's a search key */
410 struct inode * bd_inode; /* will die */
411 struct super_block * bd_super;
412 struct mutex bd_mutex; /* open/close mutex */
413 struct list_head bd_inodes;
417 bool bd_write_holder;
419 struct list_head bd_holder_disks;
421 struct block_device * bd_contains;
422 unsigned bd_block_size;
423 struct hd_struct * bd_part;
424 /* number of times partitions within this device have been opened. */
425 unsigned bd_part_count;
427 struct gendisk * bd_disk;
428 struct request_queue * bd_queue;
429 struct list_head bd_list;
431 * Private data. You must have bd_claim'ed the block_device
432 * to use this. NOTE: bd_claim allows an owner to claim
433 * the same device multiple times, the owner must take special
434 * care to not mess up bd_private for that case.
436 unsigned long bd_private;
438 /* The counter of freeze processes */
439 int bd_fsfreeze_count;
440 /* Mutex for freeze */
441 struct mutex bd_fsfreeze_mutex;
445 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
448 #define PAGECACHE_TAG_DIRTY 0
449 #define PAGECACHE_TAG_WRITEBACK 1
450 #define PAGECACHE_TAG_TOWRITE 2
452 int mapping_tagged(struct address_space *mapping, int tag);
455 * Might pages of this file be mapped into userspace?
457 static inline int mapping_mapped(struct address_space *mapping)
459 return !RB_EMPTY_ROOT(&mapping->i_mmap) ||
460 !list_empty(&mapping->i_mmap_nonlinear);
464 * Might pages of this file have been modified in userspace?
465 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
466 * marks vma as VM_SHARED if it is shared, and the file was opened for
467 * writing i.e. vma may be mprotected writable even if now readonly.
469 static inline int mapping_writably_mapped(struct address_space *mapping)
471 return mapping->i_mmap_writable != 0;
475 * Use sequence counter to get consistent i_size on 32-bit processors.
477 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
478 #include <linux/seqlock.h>
479 #define __NEED_I_SIZE_ORDERED
480 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
482 #define i_size_ordered_init(inode) do { } while (0)
486 #define ACL_NOT_CACHED ((void *)(-1))
488 #define IOP_FASTPERM 0x0001
489 #define IOP_LOOKUP 0x0002
490 #define IOP_NOFOLLOW 0x0004
493 * Keep mostly read-only and often accessed (especially for
494 * the RCU path lookup and 'stat' data) fields at the beginning
495 * of the 'struct inode'
499 unsigned short i_opflags;
502 unsigned int i_flags;
504 #ifdef CONFIG_FS_POSIX_ACL
505 struct posix_acl *i_acl;
506 struct posix_acl *i_default_acl;
509 const struct inode_operations *i_op;
510 struct super_block *i_sb;
511 struct address_space *i_mapping;
513 #ifdef CONFIG_SECURITY
517 /* Stat data, not accessed from path walking */
520 * Filesystems may only read i_nlink directly. They shall use the
521 * following functions for modification:
523 * (set|clear|inc|drop)_nlink
524 * inode_(inc|dec)_link_count
527 const unsigned int i_nlink;
528 unsigned int __i_nlink;
532 struct timespec i_atime;
533 struct timespec i_mtime;
534 struct timespec i_ctime;
535 spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
536 unsigned short i_bytes;
537 unsigned int i_blkbits;
540 #ifdef __NEED_I_SIZE_ORDERED
541 seqcount_t i_size_seqcount;
545 unsigned long i_state;
546 struct mutex i_mutex;
548 unsigned long dirtied_when; /* jiffies of first dirtying */
550 struct hlist_node i_hash;
551 struct list_head i_wb_list; /* backing dev IO list */
552 struct list_head i_lru; /* inode LRU list */
553 struct list_head i_sb_list;
555 struct hlist_head i_dentry;
556 struct rcu_head i_rcu;
560 atomic_t i_dio_count;
561 atomic_t i_writecount;
562 const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
563 struct file_lock *i_flock;
564 struct address_space i_data;
566 struct dquot *i_dquot[MAXQUOTAS];
568 struct list_head i_devices;
570 struct pipe_inode_info *i_pipe;
571 struct block_device *i_bdev;
577 #ifdef CONFIG_FSNOTIFY
578 __u32 i_fsnotify_mask; /* all events this inode cares about */
579 struct hlist_head i_fsnotify_marks;
583 atomic_t i_readcount; /* struct files open RO */
585 void *i_private; /* fs or device private pointer */
588 static inline int inode_unhashed(struct inode *inode)
590 return hlist_unhashed(&inode->i_hash);
594 * inode->i_mutex nesting subclasses for the lock validator:
596 * 0: the object of the current VFS operation
600 * 4: second non-directory
601 * The last is for certain operations (such as rename) which lock two
602 * non-directories at once.
604 * The locking order between these classes is
605 * parent -> child -> normal -> xattr -> second non-directory
607 enum inode_i_mutex_lock_class
616 void lock_two_nondirectories(struct inode *, struct inode*);
617 void unlock_two_nondirectories(struct inode *, struct inode*);
620 * NOTE: in a 32bit arch with a preemptable kernel and
621 * an UP compile the i_size_read/write must be atomic
622 * with respect to the local cpu (unlike with preempt disabled),
623 * but they don't need to be atomic with respect to other cpus like in
624 * true SMP (so they need either to either locally disable irq around
625 * the read or for example on x86 they can be still implemented as a
626 * cmpxchg8b without the need of the lock prefix). For SMP compiles
627 * and 64bit archs it makes no difference if preempt is enabled or not.
629 static inline loff_t i_size_read(const struct inode *inode)
631 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
636 seq = read_seqcount_begin(&inode->i_size_seqcount);
637 i_size = inode->i_size;
638 } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
640 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
644 i_size = inode->i_size;
648 return inode->i_size;
653 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
654 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
655 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
657 static inline void i_size_write(struct inode *inode, loff_t i_size)
659 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
661 write_seqcount_begin(&inode->i_size_seqcount);
662 inode->i_size = i_size;
663 write_seqcount_end(&inode->i_size_seqcount);
665 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
667 inode->i_size = i_size;
670 inode->i_size = i_size;
674 /* Helper functions so that in most cases filesystems will
675 * not need to deal directly with kuid_t and kgid_t and can
676 * instead deal with the raw numeric values that are stored
679 static inline uid_t i_uid_read(const struct inode *inode)
681 return from_kuid(&init_user_ns, inode->i_uid);
684 static inline gid_t i_gid_read(const struct inode *inode)
686 return from_kgid(&init_user_ns, inode->i_gid);
689 static inline void i_uid_write(struct inode *inode, uid_t uid)
691 inode->i_uid = make_kuid(&init_user_ns, uid);
694 static inline void i_gid_write(struct inode *inode, gid_t gid)
696 inode->i_gid = make_kgid(&init_user_ns, gid);
699 static inline unsigned iminor(const struct inode *inode)
701 return MINOR(inode->i_rdev);
704 static inline unsigned imajor(const struct inode *inode)
706 return MAJOR(inode->i_rdev);
709 extern struct block_device *I_BDEV(struct inode *inode);
712 rwlock_t lock; /* protects pid, uid, euid fields */
713 struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
714 enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
715 kuid_t uid, euid; /* uid/euid of process setting the owner */
716 int signum; /* posix.1b rt signal to be delivered on IO */
720 * Track a single file's readahead state
722 struct file_ra_state {
723 pgoff_t start; /* where readahead started */
724 unsigned int size; /* # of readahead pages */
725 unsigned int async_size; /* do asynchronous readahead when
726 there are only # of pages ahead */
728 unsigned int ra_pages; /* Maximum readahead window */
729 unsigned int mmap_miss; /* Cache miss stat for mmap accesses */
730 loff_t prev_pos; /* Cache last read() position */
734 * Check if @index falls in the readahead windows.
736 static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
738 return (index >= ra->start &&
739 index < ra->start + ra->size);
744 struct llist_node fu_llist;
745 struct rcu_head fu_rcuhead;
748 #define f_dentry f_path.dentry
749 struct inode *f_inode; /* cached value */
750 const struct file_operations *f_op;
753 * Protects f_ep_links, f_flags.
754 * Must not be taken from IRQ context.
757 atomic_long_t f_count;
758 unsigned int f_flags;
760 struct mutex f_pos_lock;
762 struct fown_struct f_owner;
763 const struct cred *f_cred;
764 struct file_ra_state f_ra;
767 #ifdef CONFIG_SECURITY
770 /* needed for tty driver, and maybe others */
774 /* Used by fs/eventpoll.c to link all the hooks to this file */
775 struct list_head f_ep_links;
776 struct list_head f_tfile_llink;
777 #endif /* #ifdef CONFIG_EPOLL */
778 struct address_space *f_mapping;
779 } __attribute__((aligned(4))); /* lest something weird decides that 2 is OK */
784 /* file identifier */
785 unsigned char f_handle[0];
788 static inline struct file *get_file(struct file *f)
790 atomic_long_inc(&f->f_count);
793 #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
794 #define file_count(x) atomic_long_read(&(x)->f_count)
796 #define MAX_NON_LFS ((1UL<<31) - 1)
798 /* Page cache limit. The filesystems should put that into their s_maxbytes
799 limits, otherwise bad things can happen in VM. */
800 #if BITS_PER_LONG==32
801 #define MAX_LFS_FILESIZE (((loff_t)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
802 #elif BITS_PER_LONG==64
803 #define MAX_LFS_FILESIZE ((loff_t)0x7fffffffffffffffLL)
808 #define FL_DELEG 4 /* NFSv4 delegation */
809 #define FL_ACCESS 8 /* not trying to lock, just looking */
810 #define FL_EXISTS 16 /* when unlocking, test for existence */
811 #define FL_LEASE 32 /* lease held on this file */
812 #define FL_CLOSE 64 /* unlock on close */
813 #define FL_SLEEP 128 /* A blocking lock */
814 #define FL_DOWNGRADE_PENDING 256 /* Lease is being downgraded */
815 #define FL_UNLOCK_PENDING 512 /* Lease is being broken */
818 * Special return value from posix_lock_file() and vfs_lock_file() for
819 * asynchronous locking.
821 #define FILE_LOCK_DEFERRED 1
824 * The POSIX file lock owner is determined by
825 * the "struct files_struct" in the thread group
826 * (or NULL for no owner - BSD locks).
828 * Lockd stuffs a "host" pointer into this.
830 typedef struct files_struct *fl_owner_t;
832 struct file_lock_operations {
833 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
834 void (*fl_release_private)(struct file_lock *);
837 struct lock_manager_operations {
838 int (*lm_compare_owner)(struct file_lock *, struct file_lock *);
839 unsigned long (*lm_owner_key)(struct file_lock *);
840 void (*lm_notify)(struct file_lock *); /* unblock callback */
841 int (*lm_grant)(struct file_lock *, struct file_lock *, int);
842 void (*lm_break)(struct file_lock *);
843 int (*lm_change)(struct file_lock **, int);
846 struct lock_manager {
847 struct list_head list;
851 void locks_start_grace(struct net *, struct lock_manager *);
852 void locks_end_grace(struct lock_manager *);
853 int locks_in_grace(struct net *);
855 /* that will die - we need it for nfs_lock_info */
856 #include <linux/nfs_fs_i.h>
859 * struct file_lock represents a generic "file lock". It's used to represent
860 * POSIX byte range locks, BSD (flock) locks, and leases. It's important to
861 * note that the same struct is used to represent both a request for a lock and
862 * the lock itself, but the same object is never used for both.
864 * FIXME: should we create a separate "struct lock_request" to help distinguish
867 * The i_flock list is ordered by:
869 * 1) lock type -- FL_LEASEs first, then FL_FLOCK, and finally FL_POSIX
871 * 3) lock range start
874 * Obviously, the last two criteria only matter for POSIX locks.
877 struct file_lock *fl_next; /* singly linked list for this inode */
878 struct hlist_node fl_link; /* node in global lists */
879 struct list_head fl_block; /* circular list of blocked processes */
881 unsigned int fl_flags;
882 unsigned char fl_type;
884 int fl_link_cpu; /* what cpu's list is this on? */
885 struct pid *fl_nspid;
886 wait_queue_head_t fl_wait;
887 struct file *fl_file;
891 struct fasync_struct * fl_fasync; /* for lease break notifications */
892 /* for lease breaks: */
893 unsigned long fl_break_time;
894 unsigned long fl_downgrade_time;
896 const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
897 const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
899 struct nfs_lock_info nfs_fl;
900 struct nfs4_lock_info nfs4_fl;
902 struct list_head link; /* link in AFS vnode's pending_locks list */
903 int state; /* state of grant or error if -ve */
908 /* The following constant reflects the upper bound of the file/locking space */
910 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
911 #define OFFSET_MAX INT_LIMIT(loff_t)
912 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
915 #include <linux/fcntl.h>
917 extern void send_sigio(struct fown_struct *fown, int fd, int band);
919 #ifdef CONFIG_FILE_LOCKING
920 extern int fcntl_getlk(struct file *, struct flock __user *);
921 extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
922 struct flock __user *);
924 #if BITS_PER_LONG == 32
925 extern int fcntl_getlk64(struct file *, struct flock64 __user *);
926 extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
927 struct flock64 __user *);
930 extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
931 extern int fcntl_getlease(struct file *filp);
934 void locks_free_lock(struct file_lock *fl);
935 extern void locks_init_lock(struct file_lock *);
936 extern struct file_lock * locks_alloc_lock(void);
937 extern void locks_copy_lock(struct file_lock *, struct file_lock *);
938 extern void __locks_copy_lock(struct file_lock *, const struct file_lock *);
939 extern void locks_remove_posix(struct file *, fl_owner_t);
940 extern void locks_remove_flock(struct file *);
941 extern void locks_release_private(struct file_lock *);
942 extern void posix_test_lock(struct file *, struct file_lock *);
943 extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
944 extern int posix_lock_file_wait(struct file *, struct file_lock *);
945 extern int posix_unblock_lock(struct file_lock *);
946 extern int vfs_test_lock(struct file *, struct file_lock *);
947 extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
948 extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
949 extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
950 extern int __break_lease(struct inode *inode, unsigned int flags, unsigned int type);
951 extern void lease_get_mtime(struct inode *, struct timespec *time);
952 extern int generic_setlease(struct file *, long, struct file_lock **);
953 extern int vfs_setlease(struct file *, long, struct file_lock **);
954 extern int lease_modify(struct file_lock **, int);
955 extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
956 extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
957 #else /* !CONFIG_FILE_LOCKING */
958 static inline int fcntl_getlk(struct file *file, struct flock __user *user)
963 static inline int fcntl_setlk(unsigned int fd, struct file *file,
964 unsigned int cmd, struct flock __user *user)
969 #if BITS_PER_LONG == 32
970 static inline int fcntl_getlk64(struct file *file, struct flock64 __user *user)
975 static inline int fcntl_setlk64(unsigned int fd, struct file *file,
976 unsigned int cmd, struct flock64 __user *user)
981 static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
986 static inline int fcntl_getlease(struct file *filp)
991 static inline void locks_init_lock(struct file_lock *fl)
996 static inline void __locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1001 static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1006 static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
1011 static inline void locks_remove_flock(struct file *filp)
1016 static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
1021 static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
1022 struct file_lock *conflock)
1027 static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
1032 static inline int posix_unblock_lock(struct file_lock *waiter)
1037 static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
1042 static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
1043 struct file_lock *fl, struct file_lock *conf)
1048 static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
1053 static inline int flock_lock_file_wait(struct file *filp,
1054 struct file_lock *request)
1059 static inline int __break_lease(struct inode *inode, unsigned int mode, unsigned int type)
1064 static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
1069 static inline int generic_setlease(struct file *filp, long arg,
1070 struct file_lock **flp)
1075 static inline int vfs_setlease(struct file *filp, long arg,
1076 struct file_lock **lease)
1081 static inline int lease_modify(struct file_lock **before, int arg)
1086 static inline int lock_may_read(struct inode *inode, loff_t start,
1092 static inline int lock_may_write(struct inode *inode, loff_t start,
1097 #endif /* !CONFIG_FILE_LOCKING */
1100 struct fasync_struct {
1104 struct fasync_struct *fa_next; /* singly linked list */
1105 struct file *fa_file;
1106 struct rcu_head fa_rcu;
1109 #define FASYNC_MAGIC 0x4601
1111 /* SMP safe fasync helpers: */
1112 extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
1113 extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
1114 extern int fasync_remove_entry(struct file *, struct fasync_struct **);
1115 extern struct fasync_struct *fasync_alloc(void);
1116 extern void fasync_free(struct fasync_struct *);
1118 /* can be called from interrupts */
1119 extern void kill_fasync(struct fasync_struct **, int, int);
1121 extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
1122 extern int f_setown(struct file *filp, unsigned long arg, int force);
1123 extern void f_delown(struct file *filp);
1124 extern pid_t f_getown(struct file *filp);
1125 extern int send_sigurg(struct fown_struct *fown);
1133 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1134 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1135 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1136 #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1137 #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
1139 extern struct list_head super_blocks;
1140 extern spinlock_t sb_lock;
1142 /* Possible states of 'frozen' field */
1144 SB_UNFROZEN = 0, /* FS is unfrozen */
1145 SB_FREEZE_WRITE = 1, /* Writes, dir ops, ioctls frozen */
1146 SB_FREEZE_PAGEFAULT = 2, /* Page faults stopped as well */
1147 SB_FREEZE_FS = 3, /* For internal FS use (e.g. to stop
1148 * internal threads if needed) */
1149 SB_FREEZE_COMPLETE = 4, /* ->freeze_fs finished successfully */
1152 #define SB_FREEZE_LEVELS (SB_FREEZE_COMPLETE - 1)
1155 /* Counters for counting writers at each level */
1156 struct percpu_counter counter[SB_FREEZE_LEVELS];
1157 wait_queue_head_t wait; /* queue for waiting for
1158 writers / faults to finish */
1159 int frozen; /* Is sb frozen? */
1160 wait_queue_head_t wait_unfrozen; /* queue for waiting for
1162 #ifdef CONFIG_DEBUG_LOCK_ALLOC
1163 struct lockdep_map lock_map[SB_FREEZE_LEVELS];
1167 struct super_block {
1168 struct list_head s_list; /* Keep this first */
1169 dev_t s_dev; /* search index; _not_ kdev_t */
1170 unsigned char s_blocksize_bits;
1171 unsigned long s_blocksize;
1172 loff_t s_maxbytes; /* Max file size */
1173 struct file_system_type *s_type;
1174 const struct super_operations *s_op;
1175 const struct dquot_operations *dq_op;
1176 const struct quotactl_ops *s_qcop;
1177 const struct export_operations *s_export_op;
1178 unsigned long s_flags;
1179 unsigned long s_magic;
1180 struct dentry *s_root;
1181 struct rw_semaphore s_umount;
1184 #ifdef CONFIG_SECURITY
1187 const struct xattr_handler **s_xattr;
1189 struct list_head s_inodes; /* all inodes */
1190 struct hlist_bl_head s_anon; /* anonymous dentries for (nfs) exporting */
1191 struct list_head s_mounts; /* list of mounts; _not_ for fs use */
1192 struct block_device *s_bdev;
1193 struct backing_dev_info *s_bdi;
1194 struct mtd_info *s_mtd;
1195 struct hlist_node s_instances;
1196 struct quota_info s_dquot; /* Diskquota specific options */
1198 struct sb_writers s_writers;
1200 char s_id[32]; /* Informational name */
1201 u8 s_uuid[16]; /* UUID */
1203 void *s_fs_info; /* Filesystem private info */
1204 unsigned int s_max_links;
1207 /* Granularity of c/m/atime in ns.
1208 Cannot be worse than a second */
1212 * The next field is for VFS *only*. No filesystems have any business
1213 * even looking at it. You had been warned.
1215 struct mutex s_vfs_rename_mutex; /* Kludge */
1218 * Filesystem subtype. If non-empty the filesystem type field
1219 * in /proc/mounts will be "type.subtype"
1224 * Saved mount options for lazy filesystems using
1225 * generic_show_options()
1227 char __rcu *s_options;
1228 const struct dentry_operations *s_d_op; /* default d_op for dentries */
1231 * Saved pool identifier for cleancache (-1 means none)
1233 int cleancache_poolid;
1235 struct shrinker s_shrink; /* per-sb shrinker handle */
1237 /* Number of inodes with nlink == 0 but still referenced */
1238 atomic_long_t s_remove_count;
1240 /* Being remounted read-only */
1241 int s_readonly_remount;
1243 /* AIO completions deferred from interrupt context */
1244 struct workqueue_struct *s_dio_done_wq;
1247 * Keep the lru lists last in the structure so they always sit on their
1248 * own individual cachelines.
1250 struct list_lru s_dentry_lru ____cacheline_aligned_in_smp;
1251 struct list_lru s_inode_lru ____cacheline_aligned_in_smp;
1252 struct rcu_head rcu;
1255 extern struct timespec current_fs_time(struct super_block *sb);
1258 * Snapshotting support.
1261 void __sb_end_write(struct super_block *sb, int level);
1262 int __sb_start_write(struct super_block *sb, int level, bool wait);
1265 * sb_end_write - drop write access to a superblock
1266 * @sb: the super we wrote to
1268 * Decrement number of writers to the filesystem. Wake up possible waiters
1269 * wanting to freeze the filesystem.
1271 static inline void sb_end_write(struct super_block *sb)
1273 __sb_end_write(sb, SB_FREEZE_WRITE);
1277 * sb_end_pagefault - drop write access to a superblock from a page fault
1278 * @sb: the super we wrote to
1280 * Decrement number of processes handling write page fault to the filesystem.
1281 * Wake up possible waiters wanting to freeze the filesystem.
1283 static inline void sb_end_pagefault(struct super_block *sb)
1285 __sb_end_write(sb, SB_FREEZE_PAGEFAULT);
1289 * sb_end_intwrite - drop write access to a superblock for internal fs purposes
1290 * @sb: the super we wrote to
1292 * Decrement fs-internal number of writers to the filesystem. Wake up possible
1293 * waiters wanting to freeze the filesystem.
1295 static inline void sb_end_intwrite(struct super_block *sb)
1297 __sb_end_write(sb, SB_FREEZE_FS);
1301 * sb_start_write - get write access to a superblock
1302 * @sb: the super we write to
1304 * When a process wants to write data or metadata to a file system (i.e. dirty
1305 * a page or an inode), it should embed the operation in a sb_start_write() -
1306 * sb_end_write() pair to get exclusion against file system freezing. This
1307 * function increments number of writers preventing freezing. If the file
1308 * system is already frozen, the function waits until the file system is
1311 * Since freeze protection behaves as a lock, users have to preserve
1312 * ordering of freeze protection and other filesystem locks. Generally,
1313 * freeze protection should be the outermost lock. In particular, we have:
1316 * -> i_mutex (write path, truncate, directory ops, ...)
1317 * -> s_umount (freeze_super, thaw_super)
1319 static inline void sb_start_write(struct super_block *sb)
1321 __sb_start_write(sb, SB_FREEZE_WRITE, true);
1324 static inline int sb_start_write_trylock(struct super_block *sb)
1326 return __sb_start_write(sb, SB_FREEZE_WRITE, false);
1330 * sb_start_pagefault - get write access to a superblock from a page fault
1331 * @sb: the super we write to
1333 * When a process starts handling write page fault, it should embed the
1334 * operation into sb_start_pagefault() - sb_end_pagefault() pair to get
1335 * exclusion against file system freezing. This is needed since the page fault
1336 * is going to dirty a page. This function increments number of running page
1337 * faults preventing freezing. If the file system is already frozen, the
1338 * function waits until the file system is thawed.
1340 * Since page fault freeze protection behaves as a lock, users have to preserve
1341 * ordering of freeze protection and other filesystem locks. It is advised to
1342 * put sb_start_pagefault() close to mmap_sem in lock ordering. Page fault
1343 * handling code implies lock dependency:
1346 * -> sb_start_pagefault
1348 static inline void sb_start_pagefault(struct super_block *sb)
1350 __sb_start_write(sb, SB_FREEZE_PAGEFAULT, true);
1354 * sb_start_intwrite - get write access to a superblock for internal fs purposes
1355 * @sb: the super we write to
1357 * This is the third level of protection against filesystem freezing. It is
1358 * free for use by a filesystem. The only requirement is that it must rank
1359 * below sb_start_pagefault.
1361 * For example filesystem can call sb_start_intwrite() when starting a
1362 * transaction which somewhat eases handling of freezing for internal sources
1363 * of filesystem changes (internal fs threads, discarding preallocation on file
1366 static inline void sb_start_intwrite(struct super_block *sb)
1368 __sb_start_write(sb, SB_FREEZE_FS, true);
1372 extern bool inode_owner_or_capable(const struct inode *inode);
1375 * VFS helper functions..
1377 extern int vfs_create(struct inode *, struct dentry *, umode_t, bool);
1378 extern int vfs_mkdir(struct inode *, struct dentry *, umode_t);
1379 extern int vfs_mknod(struct inode *, struct dentry *, umode_t, dev_t);
1380 extern int vfs_symlink(struct inode *, struct dentry *, const char *);
1381 extern int vfs_link(struct dentry *, struct inode *, struct dentry *, struct inode **);
1382 extern int vfs_rmdir(struct inode *, struct dentry *);
1383 extern int vfs_unlink(struct inode *, struct dentry *, struct inode **);
1384 extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *, struct inode **);
1387 * VFS dentry helper functions.
1389 extern void dentry_unhash(struct dentry *dentry);
1392 * VFS file helper functions.
1394 extern void inode_init_owner(struct inode *inode, const struct inode *dir,
1397 * VFS FS_IOC_FIEMAP helper definitions.
1399 struct fiemap_extent_info {
1400 unsigned int fi_flags; /* Flags as passed from user */
1401 unsigned int fi_extents_mapped; /* Number of mapped extents */
1402 unsigned int fi_extents_max; /* Size of fiemap_extent array */
1403 struct fiemap_extent __user *fi_extents_start; /* Start of
1404 fiemap_extent array */
1406 int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
1407 u64 phys, u64 len, u32 flags);
1408 int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
1413 * NOTE! These match bits 12..15 of stat.st_mode
1414 * (ie "(i_mode >> 12) & 15").
1416 #define DT_UNKNOWN 0
1427 * This is the "filldir" function type, used by readdir() to let
1428 * the kernel specify what kind of dirent layout it wants to have.
1429 * This allows the kernel to read directories into kernel space or
1430 * to have different dirent layouts depending on the binary type.
1432 typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned);
1433 struct dir_context {
1434 const filldir_t actor;
1438 struct block_device_operations;
1440 /* These macros are for out of kernel modules to test that
1441 * the kernel supports the unlocked_ioctl and compat_ioctl
1442 * fields in struct file_operations. */
1443 #define HAVE_COMPAT_IOCTL 1
1444 #define HAVE_UNLOCKED_IOCTL 1
1446 struct file_operations {
1447 struct module *owner;
1448 loff_t (*llseek) (struct file *, loff_t, int);
1449 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1450 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
1451 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1452 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1453 int (*iterate) (struct file *, struct dir_context *);
1454 unsigned int (*poll) (struct file *, struct poll_table_struct *);
1455 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1456 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1457 int (*mmap) (struct file *, struct vm_area_struct *);
1458 int (*open) (struct inode *, struct file *);
1459 int (*flush) (struct file *, fl_owner_t id);
1460 int (*release) (struct inode *, struct file *);
1461 int (*fsync) (struct file *, loff_t, loff_t, int datasync);
1462 int (*aio_fsync) (struct kiocb *, int datasync);
1463 int (*fasync) (int, struct file *, int);
1464 int (*lock) (struct file *, int, struct file_lock *);
1465 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1466 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1467 int (*check_flags)(int);
1468 int (*flock) (struct file *, int, struct file_lock *);
1469 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1470 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
1471 int (*setlease)(struct file *, long, struct file_lock **);
1472 long (*fallocate)(struct file *file, int mode, loff_t offset,
1474 int (*show_fdinfo)(struct seq_file *m, struct file *f);
1477 struct inode_operations {
1478 struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int);
1479 void * (*follow_link) (struct dentry *, struct nameidata *);
1480 int (*permission) (struct inode *, int);
1481 struct posix_acl * (*get_acl)(struct inode *, int);
1483 int (*readlink) (struct dentry *, char __user *,int);
1484 void (*put_link) (struct dentry *, struct nameidata *, void *);
1486 int (*create) (struct inode *,struct dentry *, umode_t, bool);
1487 int (*link) (struct dentry *,struct inode *,struct dentry *);
1488 int (*unlink) (struct inode *,struct dentry *);
1489 int (*symlink) (struct inode *,struct dentry *,const char *);
1490 int (*mkdir) (struct inode *,struct dentry *,umode_t);
1491 int (*rmdir) (struct inode *,struct dentry *);
1492 int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t);
1493 int (*rename) (struct inode *, struct dentry *,
1494 struct inode *, struct dentry *);
1495 int (*setattr) (struct dentry *, struct iattr *);
1496 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1497 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1498 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1499 ssize_t (*listxattr) (struct dentry *, char *, size_t);
1500 int (*removexattr) (struct dentry *, const char *);
1501 int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
1503 int (*update_time)(struct inode *, struct timespec *, int);
1504 int (*atomic_open)(struct inode *, struct dentry *,
1505 struct file *, unsigned open_flag,
1506 umode_t create_mode, int *opened);
1507 int (*tmpfile) (struct inode *, struct dentry *, umode_t);
1508 int (*set_acl)(struct inode *, struct posix_acl *, int);
1509 } ____cacheline_aligned;
1511 ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
1512 unsigned long nr_segs, unsigned long fast_segs,
1513 struct iovec *fast_pointer,
1514 struct iovec **ret_pointer);
1516 extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1517 extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1518 extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1519 unsigned long, loff_t *);
1520 extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1521 unsigned long, loff_t *);
1523 struct super_operations {
1524 struct inode *(*alloc_inode)(struct super_block *sb);
1525 void (*destroy_inode)(struct inode *);
1527 void (*dirty_inode) (struct inode *, int flags);
1528 int (*write_inode) (struct inode *, struct writeback_control *wbc);
1529 int (*drop_inode) (struct inode *);
1530 void (*evict_inode) (struct inode *);
1531 void (*put_super) (struct super_block *);
1532 int (*sync_fs)(struct super_block *sb, int wait);
1533 int (*freeze_fs) (struct super_block *);
1534 int (*unfreeze_fs) (struct super_block *);
1535 int (*statfs) (struct dentry *, struct kstatfs *);
1536 int (*remount_fs) (struct super_block *, int *, char *);
1537 void (*umount_begin) (struct super_block *);
1539 int (*show_options)(struct seq_file *, struct dentry *);
1540 int (*show_devname)(struct seq_file *, struct dentry *);
1541 int (*show_path)(struct seq_file *, struct dentry *);
1542 int (*show_stats)(struct seq_file *, struct dentry *);
1544 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1545 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
1547 int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
1548 long (*nr_cached_objects)(struct super_block *, int);
1549 long (*free_cached_objects)(struct super_block *, long, int);
1553 * Inode flags - they have no relation to superblock flags now
1555 #define S_SYNC 1 /* Writes are synced at once */
1556 #define S_NOATIME 2 /* Do not update access times */
1557 #define S_APPEND 4 /* Append-only file */
1558 #define S_IMMUTABLE 8 /* Immutable file */
1559 #define S_DEAD 16 /* removed, but still open directory */
1560 #define S_NOQUOTA 32 /* Inode is not counted to quota */
1561 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
1562 #define S_NOCMTIME 128 /* Do not update file c/mtime */
1563 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
1564 #define S_PRIVATE 512 /* Inode is fs-internal */
1565 #define S_IMA 1024 /* Inode has an associated IMA struct */
1566 #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
1567 #define S_NOSEC 4096 /* no suid or xattr security attributes */
1570 * Note that nosuid etc flags are inode-specific: setting some file-system
1571 * flags just means all the inodes inherit those flags by default. It might be
1572 * possible to override it selectively if you really wanted to with some
1573 * ioctl() that is not currently implemented.
1575 * Exception: MS_RDONLY is always applied to the entire file system.
1577 * Unfortunately, it is possible to change a filesystems flags with it mounted
1578 * with files in use. This means that all of the inodes will not have their
1579 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
1580 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
1582 #define __IS_FLG(inode, flg) ((inode)->i_sb->s_flags & (flg))
1584 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
1585 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
1586 ((inode)->i_flags & S_SYNC))
1587 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
1588 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
1589 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
1590 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
1591 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
1593 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
1594 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
1595 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
1596 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
1598 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
1599 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
1600 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
1601 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
1602 #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
1603 #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
1604 #define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
1607 * Inode state bits. Protected by inode->i_lock
1609 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1610 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1612 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1613 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1614 * various stages of removing an inode.
1616 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1618 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1619 * fdatasync(). i_atime is the usual cause.
1620 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1621 * these changes separately from I_DIRTY_SYNC so that we
1622 * don't have to write inode on fdatasync() when only
1623 * mtime has changed in it.
1624 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1625 * I_NEW Serves as both a mutex and completion notification.
1626 * New inodes set I_NEW. If two processes both create
1627 * the same inode, one of them will release its inode and
1628 * wait for I_NEW to be released before returning.
1629 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1630 * also cause waiting on I_NEW, without I_NEW actually
1631 * being set. find_inode() uses this to prevent returning
1632 * nearly-dead inodes.
1633 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1634 * is zero. I_FREEING must be set when I_WILL_FREE is
1636 * I_FREEING Set when inode is about to be freed but still has dirty
1637 * pages or buffers attached or the inode itself is still
1639 * I_CLEAR Added by clear_inode(). In this state the inode is
1640 * clean and can be destroyed. Inode keeps I_FREEING.
1642 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1643 * prohibited for many purposes. iget() must wait for
1644 * the inode to be completely released, then create it
1645 * anew. Other functions will just ignore such inodes,
1646 * if appropriate. I_NEW is used for waiting.
1648 * I_SYNC Writeback of inode is running. The bit is set during
1649 * data writeback, and cleared with a wakeup on the bit
1650 * address once it is done. The bit is also used to pin
1651 * the inode in memory for flusher thread.
1653 * I_REFERENCED Marks the inode as recently references on the LRU list.
1655 * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit().
1657 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1659 #define I_DIRTY_SYNC (1 << 0)
1660 #define I_DIRTY_DATASYNC (1 << 1)
1661 #define I_DIRTY_PAGES (1 << 2)
1663 #define I_NEW (1 << __I_NEW)
1664 #define I_WILL_FREE (1 << 4)
1665 #define I_FREEING (1 << 5)
1666 #define I_CLEAR (1 << 6)
1668 #define I_SYNC (1 << __I_SYNC)
1669 #define I_REFERENCED (1 << 8)
1670 #define __I_DIO_WAKEUP 9
1671 #define I_DIO_WAKEUP (1 << I_DIO_WAKEUP)
1672 #define I_LINKABLE (1 << 10)
1674 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1676 extern void __mark_inode_dirty(struct inode *, int);
1677 static inline void mark_inode_dirty(struct inode *inode)
1679 __mark_inode_dirty(inode, I_DIRTY);
1682 static inline void mark_inode_dirty_sync(struct inode *inode)
1684 __mark_inode_dirty(inode, I_DIRTY_SYNC);
1687 extern void inc_nlink(struct inode *inode);
1688 extern void drop_nlink(struct inode *inode);
1689 extern void clear_nlink(struct inode *inode);
1690 extern void set_nlink(struct inode *inode, unsigned int nlink);
1692 static inline void inode_inc_link_count(struct inode *inode)
1695 mark_inode_dirty(inode);
1698 static inline void inode_dec_link_count(struct inode *inode)
1701 mark_inode_dirty(inode);
1705 * inode_inc_iversion - increments i_version
1706 * @inode: inode that need to be updated
1708 * Every time the inode is modified, the i_version field will be incremented.
1709 * The filesystem has to be mounted with i_version flag
1712 static inline void inode_inc_iversion(struct inode *inode)
1714 spin_lock(&inode->i_lock);
1716 spin_unlock(&inode->i_lock);
1719 enum file_time_flags {
1726 extern void touch_atime(const struct path *);
1727 static inline void file_accessed(struct file *file)
1729 if (!(file->f_flags & O_NOATIME))
1730 touch_atime(&file->f_path);
1733 int sync_inode(struct inode *inode, struct writeback_control *wbc);
1734 int sync_inode_metadata(struct inode *inode, int wait);
1736 struct file_system_type {
1739 #define FS_REQUIRES_DEV 1
1740 #define FS_BINARY_MOUNTDATA 2
1741 #define FS_HAS_SUBTYPE 4
1742 #define FS_USERNS_MOUNT 8 /* Can be mounted by userns root */
1743 #define FS_USERNS_DEV_MOUNT 16 /* A userns mount does not imply MNT_NODEV */
1744 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() during rename() internally. */
1745 struct dentry *(*mount) (struct file_system_type *, int,
1746 const char *, void *);
1747 void (*kill_sb) (struct super_block *);
1748 struct module *owner;
1749 struct file_system_type * next;
1750 struct hlist_head fs_supers;
1752 struct lock_class_key s_lock_key;
1753 struct lock_class_key s_umount_key;
1754 struct lock_class_key s_vfs_rename_key;
1755 struct lock_class_key s_writers_key[SB_FREEZE_LEVELS];
1757 struct lock_class_key i_lock_key;
1758 struct lock_class_key i_mutex_key;
1759 struct lock_class_key i_mutex_dir_key;
1762 #define MODULE_ALIAS_FS(NAME) MODULE_ALIAS("fs-" NAME)
1764 extern struct dentry *mount_ns(struct file_system_type *fs_type, int flags,
1765 void *data, int (*fill_super)(struct super_block *, void *, int));
1766 extern struct dentry *mount_bdev(struct file_system_type *fs_type,
1767 int flags, const char *dev_name, void *data,
1768 int (*fill_super)(struct super_block *, void *, int));
1769 extern struct dentry *mount_single(struct file_system_type *fs_type,
1770 int flags, void *data,
1771 int (*fill_super)(struct super_block *, void *, int));
1772 extern struct dentry *mount_nodev(struct file_system_type *fs_type,
1773 int flags, void *data,
1774 int (*fill_super)(struct super_block *, void *, int));
1775 extern struct dentry *mount_subtree(struct vfsmount *mnt, const char *path);
1776 void generic_shutdown_super(struct super_block *sb);
1777 void kill_block_super(struct super_block *sb);
1778 void kill_anon_super(struct super_block *sb);
1779 void kill_litter_super(struct super_block *sb);
1780 void deactivate_super(struct super_block *sb);
1781 void deactivate_locked_super(struct super_block *sb);
1782 int set_anon_super(struct super_block *s, void *data);
1783 int get_anon_bdev(dev_t *);
1784 void free_anon_bdev(dev_t);
1785 struct super_block *sget(struct file_system_type *type,
1786 int (*test)(struct super_block *,void *),
1787 int (*set)(struct super_block *,void *),
1788 int flags, void *data);
1789 extern struct dentry *mount_pseudo(struct file_system_type *, char *,
1790 const struct super_operations *ops,
1791 const struct dentry_operations *dops,
1794 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1795 #define fops_get(fops) \
1796 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1797 #define fops_put(fops) \
1798 do { if (fops) module_put((fops)->owner); } while(0)
1800 * This one is to be used *ONLY* from ->open() instances.
1801 * fops must be non-NULL, pinned down *and* module dependencies
1802 * should be sufficient to pin the caller down as well.
1804 #define replace_fops(f, fops) \
1806 struct file *__file = (f); \
1807 fops_put(__file->f_op); \
1808 BUG_ON(!(__file->f_op = (fops))); \
1811 extern int register_filesystem(struct file_system_type *);
1812 extern int unregister_filesystem(struct file_system_type *);
1813 extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
1814 #define kern_mount(type) kern_mount_data(type, NULL)
1815 extern void kern_unmount(struct vfsmount *mnt);
1816 extern int may_umount_tree(struct vfsmount *);
1817 extern int may_umount(struct vfsmount *);
1818 extern long do_mount(const char *, const char *, const char *, unsigned long, void *);
1819 extern struct vfsmount *collect_mounts(struct path *);
1820 extern void drop_collected_mounts(struct vfsmount *);
1821 extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
1823 extern int vfs_statfs(struct path *, struct kstatfs *);
1824 extern int user_statfs(const char __user *, struct kstatfs *);
1825 extern int fd_statfs(int, struct kstatfs *);
1826 extern int vfs_ustat(dev_t, struct kstatfs *);
1827 extern int freeze_super(struct super_block *super);
1828 extern int thaw_super(struct super_block *super);
1829 extern bool our_mnt(struct vfsmount *mnt);
1830 extern bool fs_fully_visible(struct file_system_type *);
1832 extern int current_umask(void);
1834 extern void ihold(struct inode * inode);
1835 extern void iput(struct inode *);
1838 extern struct kobject *fs_kobj;
1840 #define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
1842 #define FLOCK_VERIFY_READ 1
1843 #define FLOCK_VERIFY_WRITE 2
1845 #ifdef CONFIG_FILE_LOCKING
1846 extern int locks_mandatory_locked(struct inode *);
1847 extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
1850 * Candidates for mandatory locking have the setgid bit set
1851 * but no group execute bit - an otherwise meaningless combination.
1854 static inline int __mandatory_lock(struct inode *ino)
1856 return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
1860 * ... and these candidates should be on MS_MANDLOCK mounted fs,
1861 * otherwise these will be advisory locks
1864 static inline int mandatory_lock(struct inode *ino)
1866 return IS_MANDLOCK(ino) && __mandatory_lock(ino);
1869 static inline int locks_verify_locked(struct inode *inode)
1871 if (mandatory_lock(inode))
1872 return locks_mandatory_locked(inode);
1876 static inline int locks_verify_truncate(struct inode *inode,
1880 if (inode->i_flock && mandatory_lock(inode))
1881 return locks_mandatory_area(
1882 FLOCK_VERIFY_WRITE, inode, filp,
1883 size < inode->i_size ? size : inode->i_size,
1884 (size < inode->i_size ? inode->i_size - size
1885 : size - inode->i_size)
1890 static inline int break_lease(struct inode *inode, unsigned int mode)
1893 return __break_lease(inode, mode, FL_LEASE);
1897 static inline int break_deleg(struct inode *inode, unsigned int mode)
1900 return __break_lease(inode, mode, FL_DELEG);
1904 static inline int try_break_deleg(struct inode *inode, struct inode **delegated_inode)
1908 ret = break_deleg(inode, O_WRONLY|O_NONBLOCK);
1909 if (ret == -EWOULDBLOCK && delegated_inode) {
1910 *delegated_inode = inode;
1916 static inline int break_deleg_wait(struct inode **delegated_inode)
1920 ret = break_deleg(*delegated_inode, O_WRONLY);
1921 iput(*delegated_inode);
1922 *delegated_inode = NULL;
1926 #else /* !CONFIG_FILE_LOCKING */
1927 static inline int locks_mandatory_locked(struct inode *inode)
1932 static inline int locks_mandatory_area(int rw, struct inode *inode,
1933 struct file *filp, loff_t offset,
1939 static inline int __mandatory_lock(struct inode *inode)
1944 static inline int mandatory_lock(struct inode *inode)
1949 static inline int locks_verify_locked(struct inode *inode)
1954 static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
1960 static inline int break_lease(struct inode *inode, unsigned int mode)
1965 static inline int break_deleg(struct inode *inode, unsigned int mode)
1970 static inline int try_break_deleg(struct inode *inode, struct inode **delegated_inode)
1975 static inline int break_deleg_wait(struct inode **delegated_inode)
1981 #endif /* CONFIG_FILE_LOCKING */
1986 const char *name; /* pointer to actual string */
1987 const __user char *uptr; /* original userland pointer */
1988 struct audit_names *aname;
1989 bool separate; /* should "name" be freed? */
1992 extern long vfs_truncate(struct path *, loff_t);
1993 extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
1995 extern int do_fallocate(struct file *file, int mode, loff_t offset,
1997 extern long do_sys_open(int dfd, const char __user *filename, int flags,
1999 extern struct file *file_open_name(struct filename *, int, umode_t);
2000 extern struct file *filp_open(const char *, int, umode_t);
2001 extern struct file *file_open_root(struct dentry *, struct vfsmount *,
2003 extern struct file * dentry_open(const struct path *, int, const struct cred *);
2004 extern int filp_close(struct file *, fl_owner_t id);
2006 extern struct filename *getname(const char __user *);
2007 extern struct filename *getname_kernel(const char *);
2013 extern int finish_open(struct file *file, struct dentry *dentry,
2014 int (*open)(struct inode *, struct file *),
2016 extern int finish_no_open(struct file *file, struct dentry *dentry);
2020 extern int ioctl_preallocate(struct file *filp, void __user *argp);
2023 extern void __init vfs_caches_init_early(void);
2024 extern void __init vfs_caches_init(unsigned long);
2026 extern struct kmem_cache *names_cachep;
2028 extern void final_putname(struct filename *name);
2030 #define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL)
2031 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
2032 #ifndef CONFIG_AUDITSYSCALL
2033 #define putname(name) final_putname(name)
2035 extern void putname(struct filename *name);
2039 extern int register_blkdev(unsigned int, const char *);
2040 extern void unregister_blkdev(unsigned int, const char *);
2041 extern struct block_device *bdget(dev_t);
2042 extern struct block_device *bdgrab(struct block_device *bdev);
2043 extern void bd_set_size(struct block_device *, loff_t size);
2044 extern void bd_forget(struct inode *inode);
2045 extern void bdput(struct block_device *);
2046 extern void invalidate_bdev(struct block_device *);
2047 extern void iterate_bdevs(void (*)(struct block_device *, void *), void *);
2048 extern int sync_blockdev(struct block_device *bdev);
2049 extern void kill_bdev(struct block_device *);
2050 extern struct super_block *freeze_bdev(struct block_device *);
2051 extern void emergency_thaw_all(void);
2052 extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
2053 extern int fsync_bdev(struct block_device *);
2054 extern int sb_is_blkdev_sb(struct super_block *sb);
2056 static inline void bd_forget(struct inode *inode) {}
2057 static inline int sync_blockdev(struct block_device *bdev) { return 0; }
2058 static inline void kill_bdev(struct block_device *bdev) {}
2059 static inline void invalidate_bdev(struct block_device *bdev) {}
2061 static inline struct super_block *freeze_bdev(struct block_device *sb)
2066 static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
2071 static inline void iterate_bdevs(void (*f)(struct block_device *, void *), void *arg)
2075 static inline int sb_is_blkdev_sb(struct super_block *sb)
2080 extern int sync_filesystem(struct super_block *);
2081 extern const struct file_operations def_blk_fops;
2082 extern const struct file_operations def_chr_fops;
2083 extern const struct file_operations bad_sock_fops;
2085 extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
2086 extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
2087 extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
2088 extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
2089 extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
2091 extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
2093 extern void blkdev_put(struct block_device *bdev, fmode_t mode);
2095 extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
2096 extern void bd_unlink_disk_holder(struct block_device *bdev,
2097 struct gendisk *disk);
2099 static inline int bd_link_disk_holder(struct block_device *bdev,
2100 struct gendisk *disk)
2104 static inline void bd_unlink_disk_holder(struct block_device *bdev,
2105 struct gendisk *disk)
2112 #define CHRDEV_MAJOR_HASH_SIZE 255
2113 extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
2114 extern int register_chrdev_region(dev_t, unsigned, const char *);
2115 extern int __register_chrdev(unsigned int major, unsigned int baseminor,
2116 unsigned int count, const char *name,
2117 const struct file_operations *fops);
2118 extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
2119 unsigned int count, const char *name);
2120 extern void unregister_chrdev_region(dev_t, unsigned);
2121 extern void chrdev_show(struct seq_file *,off_t);
2123 static inline int register_chrdev(unsigned int major, const char *name,
2124 const struct file_operations *fops)
2126 return __register_chrdev(major, 0, 256, name, fops);
2129 static inline void unregister_chrdev(unsigned int major, const char *name)
2131 __unregister_chrdev(major, 0, 256, name);
2134 /* fs/block_dev.c */
2135 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
2136 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
2139 #define BLKDEV_MAJOR_HASH_SIZE 255
2140 extern const char *__bdevname(dev_t, char *buffer);
2141 extern const char *bdevname(struct block_device *bdev, char *buffer);
2142 extern struct block_device *lookup_bdev(const char *);
2143 extern void blkdev_show(struct seq_file *,off_t);
2146 #define BLKDEV_MAJOR_HASH_SIZE 0
2149 extern void init_special_inode(struct inode *, umode_t, dev_t);
2151 /* Invalid inode operations -- fs/bad_inode.c */
2152 extern void make_bad_inode(struct inode *);
2153 extern int is_bad_inode(struct inode *);
2157 * return READ, READA, or WRITE
2159 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2162 * return data direction, READ or WRITE
2164 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
2166 extern void check_disk_size_change(struct gendisk *disk,
2167 struct block_device *bdev);
2168 extern int revalidate_disk(struct gendisk *);
2169 extern int check_disk_change(struct block_device *);
2170 extern int __invalidate_device(struct block_device *, bool);
2171 extern int invalidate_partition(struct gendisk *, int);
2173 unsigned long invalidate_mapping_pages(struct address_space *mapping,
2174 pgoff_t start, pgoff_t end);
2176 static inline void invalidate_remote_inode(struct inode *inode)
2178 if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
2179 S_ISLNK(inode->i_mode))
2180 invalidate_mapping_pages(inode->i_mapping, 0, -1);
2182 extern int invalidate_inode_pages2(struct address_space *mapping);
2183 extern int invalidate_inode_pages2_range(struct address_space *mapping,
2184 pgoff_t start, pgoff_t end);
2185 extern int write_inode_now(struct inode *, int);
2186 extern int filemap_fdatawrite(struct address_space *);
2187 extern int filemap_flush(struct address_space *);
2188 extern int filemap_fdatawait(struct address_space *);
2189 extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
2191 extern int filemap_write_and_wait(struct address_space *mapping);
2192 extern int filemap_write_and_wait_range(struct address_space *mapping,
2193 loff_t lstart, loff_t lend);
2194 extern int __filemap_fdatawrite_range(struct address_space *mapping,
2195 loff_t start, loff_t end, int sync_mode);
2196 extern int filemap_fdatawrite_range(struct address_space *mapping,
2197 loff_t start, loff_t end);
2199 extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
2201 extern int vfs_fsync(struct file *file, int datasync);
2202 static inline int generic_write_sync(struct file *file, loff_t pos, loff_t count)
2204 if (!(file->f_flags & O_DSYNC) && !IS_SYNC(file->f_mapping->host))
2206 return vfs_fsync_range(file, pos, pos + count - 1,
2207 (file->f_flags & __O_SYNC) ? 0 : 1);
2209 extern void emergency_sync(void);
2210 extern void emergency_remount(void);
2212 extern sector_t bmap(struct inode *, sector_t);
2214 extern int notify_change(struct dentry *, struct iattr *, struct inode **);
2215 extern int inode_permission(struct inode *, int);
2216 extern int generic_permission(struct inode *, int);
2218 static inline bool execute_ok(struct inode *inode)
2220 return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
2223 static inline struct inode *file_inode(struct file *f)
2228 static inline void file_start_write(struct file *file)
2230 if (!S_ISREG(file_inode(file)->i_mode))
2232 __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, true);
2235 static inline bool file_start_write_trylock(struct file *file)
2237 if (!S_ISREG(file_inode(file)->i_mode))
2239 return __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, false);
2242 static inline void file_end_write(struct file *file)
2244 if (!S_ISREG(file_inode(file)->i_mode))
2246 __sb_end_write(file_inode(file)->i_sb, SB_FREEZE_WRITE);
2250 * get_write_access() gets write permission for a file.
2251 * put_write_access() releases this write permission.
2252 * This is used for regular files.
2253 * We cannot support write (and maybe mmap read-write shared) accesses and
2254 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2255 * can have the following values:
2256 * 0: no writers, no VM_DENYWRITE mappings
2257 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2258 * > 0: (i_writecount) users are writing to the file.
2260 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2261 * except for the cases where we don't hold i_writecount yet. Then we need to
2262 * use {get,deny}_write_access() - these functions check the sign and refuse
2263 * to do the change if sign is wrong.
2265 static inline int get_write_access(struct inode *inode)
2267 return atomic_inc_unless_negative(&inode->i_writecount) ? 0 : -ETXTBSY;
2269 static inline int deny_write_access(struct file *file)
2271 struct inode *inode = file_inode(file);
2272 return atomic_dec_unless_positive(&inode->i_writecount) ? 0 : -ETXTBSY;
2274 static inline void put_write_access(struct inode * inode)
2276 atomic_dec(&inode->i_writecount);
2278 static inline void allow_write_access(struct file *file)
2281 atomic_inc(&file_inode(file)->i_writecount);
2283 static inline bool inode_is_open_for_write(const struct inode *inode)
2285 return atomic_read(&inode->i_writecount) > 0;
2289 static inline void i_readcount_dec(struct inode *inode)
2291 BUG_ON(!atomic_read(&inode->i_readcount));
2292 atomic_dec(&inode->i_readcount);
2294 static inline void i_readcount_inc(struct inode *inode)
2296 atomic_inc(&inode->i_readcount);
2299 static inline void i_readcount_dec(struct inode *inode)
2303 static inline void i_readcount_inc(struct inode *inode)
2308 extern int do_pipe_flags(int *, int);
2310 extern int kernel_read(struct file *, loff_t, char *, unsigned long);
2311 extern ssize_t kernel_write(struct file *, const char *, size_t, loff_t);
2312 extern struct file * open_exec(const char *);
2314 /* fs/dcache.c -- generic fs support functions */
2315 extern int is_subdir(struct dentry *, struct dentry *);
2316 extern int path_is_under(struct path *, struct path *);
2318 #include <linux/err.h>
2320 /* needed for stackable file system support */
2321 extern loff_t default_llseek(struct file *file, loff_t offset, int whence);
2323 extern loff_t vfs_llseek(struct file *file, loff_t offset, int whence);
2325 extern int inode_init_always(struct super_block *, struct inode *);
2326 extern void inode_init_once(struct inode *);
2327 extern void address_space_init_once(struct address_space *mapping);
2328 extern struct inode * igrab(struct inode *);
2329 extern ino_t iunique(struct super_block *, ino_t);
2330 extern int inode_needs_sync(struct inode *inode);
2331 extern int generic_delete_inode(struct inode *inode);
2332 static inline int generic_drop_inode(struct inode *inode)
2334 return !inode->i_nlink || inode_unhashed(inode);
2337 extern struct inode *ilookup5_nowait(struct super_block *sb,
2338 unsigned long hashval, int (*test)(struct inode *, void *),
2340 extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
2341 int (*test)(struct inode *, void *), void *data);
2342 extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
2344 extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
2345 extern struct inode * iget_locked(struct super_block *, unsigned long);
2346 extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
2347 extern int insert_inode_locked(struct inode *);
2348 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2349 extern void lockdep_annotate_inode_mutex_key(struct inode *inode);
2351 static inline void lockdep_annotate_inode_mutex_key(struct inode *inode) { };
2353 extern void unlock_new_inode(struct inode *);
2354 extern unsigned int get_next_ino(void);
2356 extern void __iget(struct inode * inode);
2357 extern void iget_failed(struct inode *);
2358 extern void clear_inode(struct inode *);
2359 extern void __destroy_inode(struct inode *);
2360 extern struct inode *new_inode_pseudo(struct super_block *sb);
2361 extern struct inode *new_inode(struct super_block *sb);
2362 extern void free_inode_nonrcu(struct inode *inode);
2363 extern int should_remove_suid(struct dentry *);
2364 extern int file_remove_suid(struct file *);
2366 extern void __insert_inode_hash(struct inode *, unsigned long hashval);
2367 static inline void insert_inode_hash(struct inode *inode)
2369 __insert_inode_hash(inode, inode->i_ino);
2372 extern void __remove_inode_hash(struct inode *);
2373 static inline void remove_inode_hash(struct inode *inode)
2375 if (!inode_unhashed(inode))
2376 __remove_inode_hash(inode);
2379 extern void inode_sb_list_add(struct inode *inode);
2382 extern void submit_bio(int, struct bio *);
2383 extern int bdev_read_only(struct block_device *);
2385 extern int set_blocksize(struct block_device *, int);
2386 extern int sb_set_blocksize(struct super_block *, int);
2387 extern int sb_min_blocksize(struct super_block *, int);
2389 extern int generic_file_mmap(struct file *, struct vm_area_struct *);
2390 extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
2391 extern int generic_file_remap_pages(struct vm_area_struct *, unsigned long addr,
2392 unsigned long size, pgoff_t pgoff);
2393 int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
2394 extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2395 extern ssize_t __generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long,
2397 extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2398 extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
2399 unsigned long *, loff_t, loff_t *, size_t, size_t);
2400 extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
2401 unsigned long, loff_t, loff_t *, size_t, ssize_t);
2402 extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
2403 extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
2404 extern int generic_segment_checks(const struct iovec *iov,
2405 unsigned long *nr_segs, size_t *count, int access_flags);
2407 /* fs/block_dev.c */
2408 extern ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
2409 unsigned long nr_segs, loff_t pos);
2410 extern int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
2412 extern void block_sync_page(struct page *page);
2415 extern ssize_t generic_file_splice_read(struct file *, loff_t *,
2416 struct pipe_inode_info *, size_t, unsigned int);
2417 extern ssize_t default_file_splice_read(struct file *, loff_t *,
2418 struct pipe_inode_info *, size_t, unsigned int);
2419 extern ssize_t generic_file_splice_write(struct pipe_inode_info *,
2420 struct file *, loff_t *, size_t, unsigned int);
2421 extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
2422 struct file *out, loff_t *, size_t len, unsigned int flags);
2425 file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
2426 extern loff_t noop_llseek(struct file *file, loff_t offset, int whence);
2427 extern loff_t no_llseek(struct file *file, loff_t offset, int whence);
2428 extern loff_t vfs_setpos(struct file *file, loff_t offset, loff_t maxsize);
2429 extern loff_t generic_file_llseek(struct file *file, loff_t offset, int whence);
2430 extern loff_t generic_file_llseek_size(struct file *file, loff_t offset,
2431 int whence, loff_t maxsize, loff_t eof);
2432 extern loff_t fixed_size_llseek(struct file *file, loff_t offset,
2433 int whence, loff_t size);
2434 extern int generic_file_open(struct inode * inode, struct file * filp);
2435 extern int nonseekable_open(struct inode * inode, struct file * filp);
2437 #ifdef CONFIG_FS_XIP
2438 extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
2440 extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
2441 extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
2442 size_t len, loff_t *ppos);
2443 extern int xip_truncate_page(struct address_space *mapping, loff_t from);
2445 static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
2452 typedef void (dio_submit_t)(int rw, struct bio *bio, struct inode *inode,
2453 loff_t file_offset);
2456 /* need locking between buffered and direct access */
2459 /* filesystem does not support filling holes */
2460 DIO_SKIP_HOLES = 0x02,
2463 void dio_end_io(struct bio *bio, int error);
2465 ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
2466 struct block_device *bdev, const struct iovec *iov, loff_t offset,
2467 unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
2468 dio_submit_t submit_io, int flags);
2470 static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
2471 struct inode *inode, const struct iovec *iov, loff_t offset,
2472 unsigned long nr_segs, get_block_t get_block)
2474 return __blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
2475 offset, nr_segs, get_block, NULL, NULL,
2476 DIO_LOCKING | DIO_SKIP_HOLES);
2480 void inode_dio_wait(struct inode *inode);
2481 void inode_dio_done(struct inode *inode);
2483 extern const struct file_operations generic_ro_fops;
2485 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2487 extern int readlink_copy(char __user *, int, const char *);
2488 extern int page_readlink(struct dentry *, char __user *, int);
2489 extern void *page_follow_link_light(struct dentry *, struct nameidata *);
2490 extern void page_put_link(struct dentry *, struct nameidata *, void *);
2491 extern int __page_symlink(struct inode *inode, const char *symname, int len,
2493 extern int page_symlink(struct inode *inode, const char *symname, int len);
2494 extern const struct inode_operations page_symlink_inode_operations;
2495 extern void kfree_put_link(struct dentry *, struct nameidata *, void *);
2496 extern int generic_readlink(struct dentry *, char __user *, int);
2497 extern void generic_fillattr(struct inode *, struct kstat *);
2498 int vfs_getattr_nosec(struct path *path, struct kstat *stat);
2499 extern int vfs_getattr(struct path *, struct kstat *);
2500 void __inode_add_bytes(struct inode *inode, loff_t bytes);
2501 void inode_add_bytes(struct inode *inode, loff_t bytes);
2502 void __inode_sub_bytes(struct inode *inode, loff_t bytes);
2503 void inode_sub_bytes(struct inode *inode, loff_t bytes);
2504 loff_t inode_get_bytes(struct inode *inode);
2505 void inode_set_bytes(struct inode *inode, loff_t bytes);
2507 extern int vfs_readdir(struct file *, filldir_t, void *);
2508 extern int iterate_dir(struct file *, struct dir_context *);
2510 extern int vfs_stat(const char __user *, struct kstat *);
2511 extern int vfs_lstat(const char __user *, struct kstat *);
2512 extern int vfs_fstat(unsigned int, struct kstat *);
2513 extern int vfs_fstatat(int , const char __user *, struct kstat *, int);
2515 extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
2517 extern int __generic_block_fiemap(struct inode *inode,
2518 struct fiemap_extent_info *fieinfo,
2519 loff_t start, loff_t len,
2520 get_block_t *get_block);
2521 extern int generic_block_fiemap(struct inode *inode,
2522 struct fiemap_extent_info *fieinfo, u64 start,
2523 u64 len, get_block_t *get_block);
2525 extern void get_filesystem(struct file_system_type *fs);
2526 extern void put_filesystem(struct file_system_type *fs);
2527 extern struct file_system_type *get_fs_type(const char *name);
2528 extern struct super_block *get_super(struct block_device *);
2529 extern struct super_block *get_super_thawed(struct block_device *);
2530 extern struct super_block *get_active_super(struct block_device *bdev);
2531 extern void drop_super(struct super_block *sb);
2532 extern void iterate_supers(void (*)(struct super_block *, void *), void *);
2533 extern void iterate_supers_type(struct file_system_type *,
2534 void (*)(struct super_block *, void *), void *);
2536 extern int dcache_dir_open(struct inode *, struct file *);
2537 extern int dcache_dir_close(struct inode *, struct file *);
2538 extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
2539 extern int dcache_readdir(struct file *, struct dir_context *);
2540 extern int simple_setattr(struct dentry *, struct iattr *);
2541 extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2542 extern int simple_statfs(struct dentry *, struct kstatfs *);
2543 extern int simple_open(struct inode *inode, struct file *file);
2544 extern int simple_link(struct dentry *, struct inode *, struct dentry *);
2545 extern int simple_unlink(struct inode *, struct dentry *);
2546 extern int simple_rmdir(struct inode *, struct dentry *);
2547 extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
2548 extern int noop_fsync(struct file *, loff_t, loff_t, int);
2549 extern int simple_empty(struct dentry *);
2550 extern int simple_readpage(struct file *file, struct page *page);
2551 extern int simple_write_begin(struct file *file, struct address_space *mapping,
2552 loff_t pos, unsigned len, unsigned flags,
2553 struct page **pagep, void **fsdata);
2554 extern int simple_write_end(struct file *file, struct address_space *mapping,
2555 loff_t pos, unsigned len, unsigned copied,
2556 struct page *page, void *fsdata);
2557 extern int always_delete_dentry(const struct dentry *);
2558 extern struct inode *alloc_anon_inode(struct super_block *);
2559 extern const struct dentry_operations simple_dentry_operations;
2561 extern struct dentry *simple_lookup(struct inode *, struct dentry *, unsigned int flags);
2562 extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
2563 extern const struct file_operations simple_dir_operations;
2564 extern const struct inode_operations simple_dir_inode_operations;
2565 struct tree_descr { char *name; const struct file_operations *ops; int mode; };
2566 struct dentry *d_alloc_name(struct dentry *, const char *);
2567 extern int simple_fill_super(struct super_block *, unsigned long, struct tree_descr *);
2568 extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
2569 extern void simple_release_fs(struct vfsmount **mount, int *count);
2571 extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
2572 loff_t *ppos, const void *from, size_t available);
2573 extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
2574 const void __user *from, size_t count);
2576 extern int generic_file_fsync(struct file *, loff_t, loff_t, int);
2578 extern int generic_check_addressable(unsigned, u64);
2580 #ifdef CONFIG_MIGRATION
2581 extern int buffer_migrate_page(struct address_space *,
2582 struct page *, struct page *,
2585 #define buffer_migrate_page NULL
2588 extern int inode_change_ok(const struct inode *, struct iattr *);
2589 extern int inode_newsize_ok(const struct inode *, loff_t offset);
2590 extern void setattr_copy(struct inode *inode, const struct iattr *attr);
2592 extern int file_update_time(struct file *file);
2594 extern int generic_show_options(struct seq_file *m, struct dentry *root);
2595 extern void save_mount_options(struct super_block *sb, char *options);
2596 extern void replace_mount_options(struct super_block *sb, char *options);
2598 static inline ino_t parent_ino(struct dentry *dentry)
2603 * Don't strictly need d_lock here? If the parent ino could change
2604 * then surely we'd have a deeper race in the caller?
2606 spin_lock(&dentry->d_lock);
2607 res = dentry->d_parent->d_inode->i_ino;
2608 spin_unlock(&dentry->d_lock);
2612 /* Transaction based IO helpers */
2615 * An argresp is stored in an allocated page and holds the
2616 * size of the argument or response, along with its content
2618 struct simple_transaction_argresp {
2623 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2625 char *simple_transaction_get(struct file *file, const char __user *buf,
2627 ssize_t simple_transaction_read(struct file *file, char __user *buf,
2628 size_t size, loff_t *pos);
2629 int simple_transaction_release(struct inode *inode, struct file *file);
2631 void simple_transaction_set(struct file *file, size_t n);
2634 * simple attribute files
2636 * These attributes behave similar to those in sysfs:
2638 * Writing to an attribute immediately sets a value, an open file can be
2639 * written to multiple times.
2641 * Reading from an attribute creates a buffer from the value that might get
2642 * read with multiple read calls. When the attribute has been read
2643 * completely, no further read calls are possible until the file is opened
2646 * All attributes contain a text representation of a numeric value
2647 * that are accessed with the get() and set() functions.
2649 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2650 static int __fops ## _open(struct inode *inode, struct file *file) \
2652 __simple_attr_check_format(__fmt, 0ull); \
2653 return simple_attr_open(inode, file, __get, __set, __fmt); \
2655 static const struct file_operations __fops = { \
2656 .owner = THIS_MODULE, \
2657 .open = __fops ## _open, \
2658 .release = simple_attr_release, \
2659 .read = simple_attr_read, \
2660 .write = simple_attr_write, \
2661 .llseek = generic_file_llseek, \
2664 static inline __printf(1, 2)
2665 void __simple_attr_check_format(const char *fmt, ...)
2667 /* don't do anything, just let the compiler check the arguments; */
2670 int simple_attr_open(struct inode *inode, struct file *file,
2671 int (*get)(void *, u64 *), int (*set)(void *, u64),
2673 int simple_attr_release(struct inode *inode, struct file *file);
2674 ssize_t simple_attr_read(struct file *file, char __user *buf,
2675 size_t len, loff_t *ppos);
2676 ssize_t simple_attr_write(struct file *file, const char __user *buf,
2677 size_t len, loff_t *ppos);
2680 int proc_nr_files(struct ctl_table *table, int write,
2681 void __user *buffer, size_t *lenp, loff_t *ppos);
2682 int proc_nr_dentry(struct ctl_table *table, int write,
2683 void __user *buffer, size_t *lenp, loff_t *ppos);
2684 int proc_nr_inodes(struct ctl_table *table, int write,
2685 void __user *buffer, size_t *lenp, loff_t *ppos);
2686 int __init get_filesystem_list(char *buf);
2688 #define __FMODE_EXEC ((__force int) FMODE_EXEC)
2689 #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
2691 #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
2692 #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
2693 (flag & __FMODE_NONOTIFY)))
2695 static inline int is_sxid(umode_t mode)
2697 return (mode & S_ISUID) || ((mode & S_ISGID) && (mode & S_IXGRP));
2700 static inline void inode_has_no_xattr(struct inode *inode)
2702 if (!is_sxid(inode->i_mode) && (inode->i_sb->s_flags & MS_NOSEC))
2703 inode->i_flags |= S_NOSEC;
2706 static inline bool dir_emit(struct dir_context *ctx,
2707 const char *name, int namelen,
2708 u64 ino, unsigned type)
2710 return ctx->actor(ctx, name, namelen, ctx->pos, ino, type) == 0;
2712 static inline bool dir_emit_dot(struct file *file, struct dir_context *ctx)
2714 return ctx->actor(ctx, ".", 1, ctx->pos,
2715 file->f_path.dentry->d_inode->i_ino, DT_DIR) == 0;
2717 static inline bool dir_emit_dotdot(struct file *file, struct dir_context *ctx)
2719 return ctx->actor(ctx, "..", 2, ctx->pos,
2720 parent_ino(file->f_path.dentry), DT_DIR) == 0;
2722 static inline bool dir_emit_dots(struct file *file, struct dir_context *ctx)
2724 if (ctx->pos == 0) {
2725 if (!dir_emit_dot(file, ctx))
2729 if (ctx->pos == 1) {
2730 if (!dir_emit_dotdot(file, ctx))
2736 static inline bool dir_relax(struct inode *inode)
2738 mutex_unlock(&inode->i_mutex);
2739 mutex_lock(&inode->i_mutex);
2740 return !IS_DEADDIR(inode);
2743 #endif /* _LINUX_FS_H */