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/rwsem.h>
22 #include <linux/capability.h>
23 #include <linux/semaphore.h>
24 #include <linux/fiemap.h>
25 #include <linux/rculist_bl.h>
26 #include <linux/atomic.h>
27 #include <linux/shrinker.h>
28 #include <linux/migrate_mode.h>
29 #include <linux/uidgid.h>
30 #include <linux/lockdep.h>
31 #include <linux/percpu-rwsem.h>
32 #include <linux/blk_types.h>
34 #include <asm/byteorder.h>
35 #include <uapi/linux/fs.h>
37 struct backing_dev_info;
38 struct export_operations;
44 struct pipe_inode_info;
45 struct poll_table_struct;
47 struct vm_area_struct;
50 struct swap_info_struct;
52 struct workqueue_struct;
56 extern void __init inode_init(void);
57 extern void __init inode_init_early(void);
58 extern void __init files_init(unsigned long);
60 extern struct files_stat_struct files_stat;
61 extern unsigned long get_max_files(void);
62 extern int sysctl_nr_open;
63 extern struct inodes_stat_t inodes_stat;
64 extern int leases_enable, lease_break_time;
65 extern int sysctl_protected_symlinks;
66 extern int sysctl_protected_hardlinks;
69 typedef int (get_block_t)(struct inode *inode, sector_t iblock,
70 struct buffer_head *bh_result, int create);
71 typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
72 ssize_t bytes, void *private);
74 #define MAY_EXEC 0x00000001
75 #define MAY_WRITE 0x00000002
76 #define MAY_READ 0x00000004
77 #define MAY_APPEND 0x00000008
78 #define MAY_ACCESS 0x00000010
79 #define MAY_OPEN 0x00000020
80 #define MAY_CHDIR 0x00000040
81 /* called from RCU mode, don't block */
82 #define MAY_NOT_BLOCK 0x00000080
85 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
86 * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
89 /* file is open for reading */
90 #define FMODE_READ ((__force fmode_t)0x1)
91 /* file is open for writing */
92 #define FMODE_WRITE ((__force fmode_t)0x2)
93 /* file is seekable */
94 #define FMODE_LSEEK ((__force fmode_t)0x4)
95 /* file can be accessed using pread */
96 #define FMODE_PREAD ((__force fmode_t)0x8)
97 /* file can be accessed using pwrite */
98 #define FMODE_PWRITE ((__force fmode_t)0x10)
99 /* File is opened for execution with sys_execve / sys_uselib */
100 #define FMODE_EXEC ((__force fmode_t)0x20)
101 /* File is opened with O_NDELAY (only set for block devices) */
102 #define FMODE_NDELAY ((__force fmode_t)0x40)
103 /* File is opened with O_EXCL (only set for block devices) */
104 #define FMODE_EXCL ((__force fmode_t)0x80)
105 /* File is opened using open(.., 3, ..) and is writeable only for ioctls
106 (specialy hack for floppy.c) */
107 #define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
108 /* 32bit hashes as llseek() offset (for directories) */
109 #define FMODE_32BITHASH ((__force fmode_t)0x200)
110 /* 64bit hashes as llseek() offset (for directories) */
111 #define FMODE_64BITHASH ((__force fmode_t)0x400)
114 * Don't update ctime and mtime.
116 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
117 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
119 #define FMODE_NOCMTIME ((__force fmode_t)0x800)
121 /* Expect random access pattern */
122 #define FMODE_RANDOM ((__force fmode_t)0x1000)
124 /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
125 #define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
127 /* File is opened with O_PATH; almost nothing can be done with it */
128 #define FMODE_PATH ((__force fmode_t)0x4000)
130 /* File needs atomic accesses to f_pos */
131 #define FMODE_ATOMIC_POS ((__force fmode_t)0x8000)
132 /* Write access to underlying fs */
133 #define FMODE_WRITER ((__force fmode_t)0x10000)
134 /* Has read method(s) */
135 #define FMODE_CAN_READ ((__force fmode_t)0x20000)
136 /* Has write method(s) */
137 #define FMODE_CAN_WRITE ((__force fmode_t)0x40000)
139 /* File was opened by fanotify and shouldn't generate fanotify events */
140 #define FMODE_NONOTIFY ((__force fmode_t)0x4000000)
143 * Flag for rw_copy_check_uvector and compat_rw_copy_check_uvector
144 * that indicates that they should check the contents of the iovec are
145 * valid, but not check the memory that the iovec elements
148 #define CHECK_IOVEC_ONLY -1
151 * The below are the various read and write types that we support. Some of
152 * them include behavioral modifiers that send information down to the
153 * block layer and IO scheduler. Terminology:
155 * The block layer uses device plugging to defer IO a little bit, in
156 * the hope that we will see more IO very shortly. This increases
157 * coalescing of adjacent IO and thus reduces the number of IOs we
158 * have to send to the device. It also allows for better queuing,
159 * if the IO isn't mergeable. If the caller is going to be waiting
160 * for the IO, then he must ensure that the device is unplugged so
161 * that the IO is dispatched to the driver.
163 * All IO is handled async in Linux. This is fine for background
164 * writes, but for reads or writes that someone waits for completion
165 * on, we want to notify the block layer and IO scheduler so that they
166 * know about it. That allows them to make better scheduling
167 * decisions. So when the below references 'sync' and 'async', it
168 * is referencing this priority hint.
170 * With that in mind, the available types are:
172 * READ A normal read operation. Device will be plugged.
173 * READ_SYNC A synchronous read. Device is not plugged, caller can
174 * immediately wait on this read without caring about
176 * READA Used for read-ahead operations. Lower priority, and the
177 * block layer could (in theory) choose to ignore this
178 * request if it runs into resource problems.
179 * WRITE A normal async write. Device will be plugged.
180 * WRITE_SYNC Synchronous write. Identical to WRITE, but passes down
181 * the hint that someone will be waiting on this IO
182 * shortly. The write equivalent of READ_SYNC.
183 * WRITE_ODIRECT Special case write for O_DIRECT only.
184 * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
185 * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
186 * non-volatile media on completion.
187 * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
188 * by a cache flush and data is guaranteed to be on
189 * non-volatile media on completion.
192 #define RW_MASK REQ_WRITE
193 #define RWA_MASK REQ_RAHEAD
196 #define WRITE RW_MASK
197 #define READA RWA_MASK
199 #define READ_SYNC (READ | REQ_SYNC)
200 #define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE)
201 #define WRITE_ODIRECT (WRITE | REQ_SYNC)
202 #define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
203 #define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
204 #define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
207 * Attribute flags. These should be or-ed together to figure out what
210 #define ATTR_MODE (1 << 0)
211 #define ATTR_UID (1 << 1)
212 #define ATTR_GID (1 << 2)
213 #define ATTR_SIZE (1 << 3)
214 #define ATTR_ATIME (1 << 4)
215 #define ATTR_MTIME (1 << 5)
216 #define ATTR_CTIME (1 << 6)
217 #define ATTR_ATIME_SET (1 << 7)
218 #define ATTR_MTIME_SET (1 << 8)
219 #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
220 #define ATTR_ATTR_FLAG (1 << 10)
221 #define ATTR_KILL_SUID (1 << 11)
222 #define ATTR_KILL_SGID (1 << 12)
223 #define ATTR_FILE (1 << 13)
224 #define ATTR_KILL_PRIV (1 << 14)
225 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
226 #define ATTR_TIMES_SET (1 << 16)
229 * Whiteout is represented by a char device. The following constants define the
230 * mode and device number to use.
232 #define WHITEOUT_MODE 0
233 #define WHITEOUT_DEV 0
236 * This is the Inode Attributes structure, used for notify_change(). It
237 * uses the above definitions as flags, to know which values have changed.
238 * Also, in this manner, a Filesystem can look at only the values it cares
239 * about. Basically, these are the attributes that the VFS layer can
240 * request to change from the FS layer.
242 * Derek Atkins <warlord@MIT.EDU> 94-10-20
245 unsigned int ia_valid;
250 struct timespec ia_atime;
251 struct timespec ia_mtime;
252 struct timespec ia_ctime;
255 * Not an attribute, but an auxiliary info for filesystems wanting to
256 * implement an ftruncate() like method. NOTE: filesystem should
257 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
259 struct file *ia_file;
263 * Includes for diskquotas.
265 #include <linux/quota.h>
268 * Maximum number of layers of fs stack. Needs to be limited to
269 * prevent kernel stack overflow
271 #define FILESYSTEM_MAX_STACK_DEPTH 2
274 * enum positive_aop_returns - aop return codes with specific semantics
276 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
277 * completed, that the page is still locked, and
278 * should be considered active. The VM uses this hint
279 * to return the page to the active list -- it won't
280 * be a candidate for writeback again in the near
281 * future. Other callers must be careful to unlock
282 * the page if they get this return. Returned by
285 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
286 * unlocked it and the page might have been truncated.
287 * The caller should back up to acquiring a new page and
288 * trying again. The aop will be taking reasonable
289 * precautions not to livelock. If the caller held a page
290 * reference, it should drop it before retrying. Returned
293 * address_space_operation functions return these large constants to indicate
294 * special semantics to the caller. These are much larger than the bytes in a
295 * page to allow for functions that return the number of bytes operated on in a
299 enum positive_aop_returns {
300 AOP_WRITEPAGE_ACTIVATE = 0x80000,
301 AOP_TRUNCATED_PAGE = 0x80001,
304 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
305 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
306 #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
307 * helper code (eg buffer layer)
308 * to clear GFP_FS from alloc */
311 * oh the beauties of C type declarations.
314 struct address_space;
315 struct writeback_control;
317 #define IOCB_EVENTFD (1 << 0)
320 struct file *ki_filp;
322 void (*ki_complete)(struct kiocb *iocb, long ret, long ret2);
327 static inline bool is_sync_kiocb(struct kiocb *kiocb)
329 return kiocb->ki_complete == NULL;
332 static inline void init_sync_kiocb(struct kiocb *kiocb, struct file *filp)
334 *kiocb = (struct kiocb) {
340 * "descriptor" for what we're up to with a read.
341 * This allows us to use the same read code yet
342 * have multiple different users of the data that
343 * we read from a file.
345 * The simplest case just copies the data to user
358 typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
359 unsigned long, unsigned long);
361 struct address_space_operations {
362 int (*writepage)(struct page *page, struct writeback_control *wbc);
363 int (*readpage)(struct file *, struct page *);
365 /* Write back some dirty pages from this mapping. */
366 int (*writepages)(struct address_space *, struct writeback_control *);
368 /* Set a page dirty. Return true if this dirtied it */
369 int (*set_page_dirty)(struct page *page);
371 int (*readpages)(struct file *filp, struct address_space *mapping,
372 struct list_head *pages, unsigned nr_pages);
374 int (*write_begin)(struct file *, struct address_space *mapping,
375 loff_t pos, unsigned len, unsigned flags,
376 struct page **pagep, void **fsdata);
377 int (*write_end)(struct file *, struct address_space *mapping,
378 loff_t pos, unsigned len, unsigned copied,
379 struct page *page, void *fsdata);
381 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
382 sector_t (*bmap)(struct address_space *, sector_t);
383 void (*invalidatepage) (struct page *, unsigned int, unsigned int);
384 int (*releasepage) (struct page *, gfp_t);
385 void (*freepage)(struct page *);
386 ssize_t (*direct_IO)(int, struct kiocb *, struct iov_iter *iter, loff_t offset);
388 * migrate the contents of a page to the specified target. If
389 * migrate_mode is MIGRATE_ASYNC, it must not block.
391 int (*migratepage) (struct address_space *,
392 struct page *, struct page *, enum migrate_mode);
393 int (*launder_page) (struct page *);
394 int (*is_partially_uptodate) (struct page *, unsigned long,
396 void (*is_dirty_writeback) (struct page *, bool *, bool *);
397 int (*error_remove_page)(struct address_space *, struct page *);
399 /* swapfile support */
400 int (*swap_activate)(struct swap_info_struct *sis, struct file *file,
402 void (*swap_deactivate)(struct file *file);
405 extern const struct address_space_operations empty_aops;
408 * pagecache_write_begin/pagecache_write_end must be used by general code
409 * to write into the pagecache.
411 int pagecache_write_begin(struct file *, struct address_space *mapping,
412 loff_t pos, unsigned len, unsigned flags,
413 struct page **pagep, void **fsdata);
415 int pagecache_write_end(struct file *, struct address_space *mapping,
416 loff_t pos, unsigned len, unsigned copied,
417 struct page *page, void *fsdata);
419 struct address_space {
420 struct inode *host; /* owner: inode, block_device */
421 struct radix_tree_root page_tree; /* radix tree of all pages */
422 spinlock_t tree_lock; /* and lock protecting it */
423 atomic_t i_mmap_writable;/* count VM_SHARED mappings */
424 struct rb_root i_mmap; /* tree of private and shared mappings */
425 struct rw_semaphore i_mmap_rwsem; /* protect tree, count, list */
426 /* Protected by tree_lock together with the radix tree */
427 unsigned long nrpages; /* number of total pages */
428 unsigned long nrshadows; /* number of shadow entries */
429 pgoff_t writeback_index;/* writeback starts here */
430 const struct address_space_operations *a_ops; /* methods */
431 unsigned long flags; /* error bits/gfp mask */
432 spinlock_t private_lock; /* for use by the address_space */
433 struct list_head private_list; /* ditto */
434 void *private_data; /* ditto */
435 } __attribute__((aligned(sizeof(long))));
437 * On most architectures that alignment is already the case; but
438 * must be enforced here for CRIS, to let the least significant bit
439 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
441 struct request_queue;
443 struct block_device {
444 dev_t bd_dev; /* not a kdev_t - it's a search key */
446 struct inode * bd_inode; /* will die */
447 struct super_block * bd_super;
448 struct mutex bd_mutex; /* open/close mutex */
449 struct list_head bd_inodes;
453 bool bd_write_holder;
455 struct list_head bd_holder_disks;
457 struct block_device * bd_contains;
458 unsigned bd_block_size;
459 struct hd_struct * bd_part;
460 /* number of times partitions within this device have been opened. */
461 unsigned bd_part_count;
463 struct gendisk * bd_disk;
464 struct request_queue * bd_queue;
465 struct list_head bd_list;
467 * Private data. You must have bd_claim'ed the block_device
468 * to use this. NOTE: bd_claim allows an owner to claim
469 * the same device multiple times, the owner must take special
470 * care to not mess up bd_private for that case.
472 unsigned long bd_private;
474 /* The counter of freeze processes */
475 int bd_fsfreeze_count;
476 /* Mutex for freeze */
477 struct mutex bd_fsfreeze_mutex;
481 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
484 #define PAGECACHE_TAG_DIRTY 0
485 #define PAGECACHE_TAG_WRITEBACK 1
486 #define PAGECACHE_TAG_TOWRITE 2
488 int mapping_tagged(struct address_space *mapping, int tag);
490 static inline void i_mmap_lock_write(struct address_space *mapping)
492 down_write(&mapping->i_mmap_rwsem);
495 static inline void i_mmap_unlock_write(struct address_space *mapping)
497 up_write(&mapping->i_mmap_rwsem);
500 static inline void i_mmap_lock_read(struct address_space *mapping)
502 down_read(&mapping->i_mmap_rwsem);
505 static inline void i_mmap_unlock_read(struct address_space *mapping)
507 up_read(&mapping->i_mmap_rwsem);
511 * Might pages of this file be mapped into userspace?
513 static inline int mapping_mapped(struct address_space *mapping)
515 return !RB_EMPTY_ROOT(&mapping->i_mmap);
519 * Might pages of this file have been modified in userspace?
520 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
521 * marks vma as VM_SHARED if it is shared, and the file was opened for
522 * writing i.e. vma may be mprotected writable even if now readonly.
524 * If i_mmap_writable is negative, no new writable mappings are allowed. You
525 * can only deny writable mappings, if none exists right now.
527 static inline int mapping_writably_mapped(struct address_space *mapping)
529 return atomic_read(&mapping->i_mmap_writable) > 0;
532 static inline int mapping_map_writable(struct address_space *mapping)
534 return atomic_inc_unless_negative(&mapping->i_mmap_writable) ?
538 static inline void mapping_unmap_writable(struct address_space *mapping)
540 atomic_dec(&mapping->i_mmap_writable);
543 static inline int mapping_deny_writable(struct address_space *mapping)
545 return atomic_dec_unless_positive(&mapping->i_mmap_writable) ?
549 static inline void mapping_allow_writable(struct address_space *mapping)
551 atomic_inc(&mapping->i_mmap_writable);
555 * Use sequence counter to get consistent i_size on 32-bit processors.
557 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
558 #include <linux/seqlock.h>
559 #define __NEED_I_SIZE_ORDERED
560 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
562 #define i_size_ordered_init(inode) do { } while (0)
566 #define ACL_NOT_CACHED ((void *)(-1))
568 #define IOP_FASTPERM 0x0001
569 #define IOP_LOOKUP 0x0002
570 #define IOP_NOFOLLOW 0x0004
573 * Keep mostly read-only and often accessed (especially for
574 * the RCU path lookup and 'stat' data) fields at the beginning
575 * of the 'struct inode'
579 unsigned short i_opflags;
582 unsigned int i_flags;
584 #ifdef CONFIG_FS_POSIX_ACL
585 struct posix_acl *i_acl;
586 struct posix_acl *i_default_acl;
589 const struct inode_operations *i_op;
590 struct super_block *i_sb;
591 struct address_space *i_mapping;
593 #ifdef CONFIG_SECURITY
597 /* Stat data, not accessed from path walking */
600 * Filesystems may only read i_nlink directly. They shall use the
601 * following functions for modification:
603 * (set|clear|inc|drop)_nlink
604 * inode_(inc|dec)_link_count
607 const unsigned int i_nlink;
608 unsigned int __i_nlink;
612 struct timespec i_atime;
613 struct timespec i_mtime;
614 struct timespec i_ctime;
615 spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
616 unsigned short i_bytes;
617 unsigned int i_blkbits;
620 #ifdef __NEED_I_SIZE_ORDERED
621 seqcount_t i_size_seqcount;
625 unsigned long i_state;
626 struct mutex i_mutex;
628 unsigned long dirtied_when; /* jiffies of first dirtying */
630 struct hlist_node i_hash;
631 struct list_head i_wb_list; /* backing dev IO list */
632 struct list_head i_lru; /* inode LRU list */
633 struct list_head i_sb_list;
635 struct hlist_head i_dentry;
636 struct rcu_head i_rcu;
640 atomic_t i_dio_count;
641 atomic_t i_writecount;
643 atomic_t i_readcount; /* struct files open RO */
645 const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
646 struct file_lock_context *i_flctx;
647 struct address_space i_data;
648 struct list_head i_devices;
650 struct pipe_inode_info *i_pipe;
651 struct block_device *i_bdev;
657 #ifdef CONFIG_FSNOTIFY
658 __u32 i_fsnotify_mask; /* all events this inode cares about */
659 struct hlist_head i_fsnotify_marks;
662 void *i_private; /* fs or device private pointer */
665 static inline int inode_unhashed(struct inode *inode)
667 return hlist_unhashed(&inode->i_hash);
671 * inode->i_mutex nesting subclasses for the lock validator:
673 * 0: the object of the current VFS operation
677 * 4: second non-directory
678 * 5: second parent (when locking independent directories in rename)
680 * I_MUTEX_NONDIR2 is for certain operations (such as rename) which lock two
681 * non-directories at once.
683 * The locking order between these classes is
684 * parent[2] -> child -> grandchild -> normal -> xattr -> second non-directory
686 enum inode_i_mutex_lock_class
696 void lock_two_nondirectories(struct inode *, struct inode*);
697 void unlock_two_nondirectories(struct inode *, struct inode*);
700 * NOTE: in a 32bit arch with a preemptable kernel and
701 * an UP compile the i_size_read/write must be atomic
702 * with respect to the local cpu (unlike with preempt disabled),
703 * but they don't need to be atomic with respect to other cpus like in
704 * true SMP (so they need either to either locally disable irq around
705 * the read or for example on x86 they can be still implemented as a
706 * cmpxchg8b without the need of the lock prefix). For SMP compiles
707 * and 64bit archs it makes no difference if preempt is enabled or not.
709 static inline loff_t i_size_read(const struct inode *inode)
711 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
716 seq = read_seqcount_begin(&inode->i_size_seqcount);
717 i_size = inode->i_size;
718 } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
720 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
724 i_size = inode->i_size;
728 return inode->i_size;
733 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
734 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
735 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
737 static inline void i_size_write(struct inode *inode, loff_t i_size)
739 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
741 write_seqcount_begin(&inode->i_size_seqcount);
742 inode->i_size = i_size;
743 write_seqcount_end(&inode->i_size_seqcount);
745 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
747 inode->i_size = i_size;
750 inode->i_size = i_size;
754 /* Helper functions so that in most cases filesystems will
755 * not need to deal directly with kuid_t and kgid_t and can
756 * instead deal with the raw numeric values that are stored
759 static inline uid_t i_uid_read(const struct inode *inode)
761 return from_kuid(&init_user_ns, inode->i_uid);
764 static inline gid_t i_gid_read(const struct inode *inode)
766 return from_kgid(&init_user_ns, inode->i_gid);
769 static inline void i_uid_write(struct inode *inode, uid_t uid)
771 inode->i_uid = make_kuid(&init_user_ns, uid);
774 static inline void i_gid_write(struct inode *inode, gid_t gid)
776 inode->i_gid = make_kgid(&init_user_ns, gid);
779 static inline unsigned iminor(const struct inode *inode)
781 return MINOR(inode->i_rdev);
784 static inline unsigned imajor(const struct inode *inode)
786 return MAJOR(inode->i_rdev);
789 extern struct block_device *I_BDEV(struct inode *inode);
792 rwlock_t lock; /* protects pid, uid, euid fields */
793 struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
794 enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
795 kuid_t uid, euid; /* uid/euid of process setting the owner */
796 int signum; /* posix.1b rt signal to be delivered on IO */
800 * Track a single file's readahead state
802 struct file_ra_state {
803 pgoff_t start; /* where readahead started */
804 unsigned int size; /* # of readahead pages */
805 unsigned int async_size; /* do asynchronous readahead when
806 there are only # of pages ahead */
808 unsigned int ra_pages; /* Maximum readahead window */
809 unsigned int mmap_miss; /* Cache miss stat for mmap accesses */
810 loff_t prev_pos; /* Cache last read() position */
814 * Check if @index falls in the readahead windows.
816 static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
818 return (index >= ra->start &&
819 index < ra->start + ra->size);
824 struct llist_node fu_llist;
825 struct rcu_head fu_rcuhead;
828 struct inode *f_inode; /* cached value */
829 const struct file_operations *f_op;
832 * Protects f_ep_links, f_flags.
833 * Must not be taken from IRQ context.
836 atomic_long_t f_count;
837 unsigned int f_flags;
839 struct mutex f_pos_lock;
841 struct fown_struct f_owner;
842 const struct cred *f_cred;
843 struct file_ra_state f_ra;
846 #ifdef CONFIG_SECURITY
849 /* needed for tty driver, and maybe others */
853 /* Used by fs/eventpoll.c to link all the hooks to this file */
854 struct list_head f_ep_links;
855 struct list_head f_tfile_llink;
856 #endif /* #ifdef CONFIG_EPOLL */
857 struct address_space *f_mapping;
858 } __attribute__((aligned(4))); /* lest something weird decides that 2 is OK */
863 /* file identifier */
864 unsigned char f_handle[0];
867 static inline struct file *get_file(struct file *f)
869 atomic_long_inc(&f->f_count);
872 #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
873 #define file_count(x) atomic_long_read(&(x)->f_count)
875 #define MAX_NON_LFS ((1UL<<31) - 1)
877 /* Page cache limit. The filesystems should put that into their s_maxbytes
878 limits, otherwise bad things can happen in VM. */
879 #if BITS_PER_LONG==32
880 #define MAX_LFS_FILESIZE (((loff_t)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
881 #elif BITS_PER_LONG==64
882 #define MAX_LFS_FILESIZE ((loff_t)0x7fffffffffffffffLL)
887 #define FL_DELEG 4 /* NFSv4 delegation */
888 #define FL_ACCESS 8 /* not trying to lock, just looking */
889 #define FL_EXISTS 16 /* when unlocking, test for existence */
890 #define FL_LEASE 32 /* lease held on this file */
891 #define FL_CLOSE 64 /* unlock on close */
892 #define FL_SLEEP 128 /* A blocking lock */
893 #define FL_DOWNGRADE_PENDING 256 /* Lease is being downgraded */
894 #define FL_UNLOCK_PENDING 512 /* Lease is being broken */
895 #define FL_OFDLCK 1024 /* lock is "owned" by struct file */
896 #define FL_LAYOUT 2048 /* outstanding pNFS layout */
899 * Special return value from posix_lock_file() and vfs_lock_file() for
900 * asynchronous locking.
902 #define FILE_LOCK_DEFERRED 1
904 /* legacy typedef, should eventually be removed */
905 typedef void *fl_owner_t;
909 struct file_lock_operations {
910 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
911 void (*fl_release_private)(struct file_lock *);
914 struct lock_manager_operations {
915 int (*lm_compare_owner)(struct file_lock *, struct file_lock *);
916 unsigned long (*lm_owner_key)(struct file_lock *);
917 void (*lm_get_owner)(struct file_lock *, struct file_lock *);
918 void (*lm_put_owner)(struct file_lock *);
919 void (*lm_notify)(struct file_lock *); /* unblock callback */
920 int (*lm_grant)(struct file_lock *, int);
921 bool (*lm_break)(struct file_lock *);
922 int (*lm_change)(struct file_lock *, int, struct list_head *);
923 void (*lm_setup)(struct file_lock *, void **);
926 struct lock_manager {
927 struct list_head list;
931 void locks_start_grace(struct net *, struct lock_manager *);
932 void locks_end_grace(struct lock_manager *);
933 int locks_in_grace(struct net *);
935 /* that will die - we need it for nfs_lock_info */
936 #include <linux/nfs_fs_i.h>
939 * struct file_lock represents a generic "file lock". It's used to represent
940 * POSIX byte range locks, BSD (flock) locks, and leases. It's important to
941 * note that the same struct is used to represent both a request for a lock and
942 * the lock itself, but the same object is never used for both.
944 * FIXME: should we create a separate "struct lock_request" to help distinguish
947 * The varous i_flctx lists are ordered by:
950 * 2) lock range start
953 * Obviously, the last two criteria only matter for POSIX locks.
956 struct file_lock *fl_next; /* singly linked list for this inode */
957 struct list_head fl_list; /* link into file_lock_context */
958 struct hlist_node fl_link; /* node in global lists */
959 struct list_head fl_block; /* circular list of blocked processes */
961 unsigned int fl_flags;
962 unsigned char fl_type;
964 int fl_link_cpu; /* what cpu's list is this on? */
965 struct pid *fl_nspid;
966 wait_queue_head_t fl_wait;
967 struct file *fl_file;
971 struct fasync_struct * fl_fasync; /* for lease break notifications */
972 /* for lease breaks: */
973 unsigned long fl_break_time;
974 unsigned long fl_downgrade_time;
976 const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
977 const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
979 struct nfs_lock_info nfs_fl;
980 struct nfs4_lock_info nfs4_fl;
982 struct list_head link; /* link in AFS vnode's pending_locks list */
983 int state; /* state of grant or error if -ve */
988 struct file_lock_context {
990 struct list_head flc_flock;
991 struct list_head flc_posix;
992 struct list_head flc_lease;
995 /* The following constant reflects the upper bound of the file/locking space */
997 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
998 #define OFFSET_MAX INT_LIMIT(loff_t)
999 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
1002 #include <linux/fcntl.h>
1004 extern void send_sigio(struct fown_struct *fown, int fd, int band);
1006 #ifdef CONFIG_FILE_LOCKING
1007 extern int fcntl_getlk(struct file *, unsigned int, struct flock __user *);
1008 extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
1009 struct flock __user *);
1011 #if BITS_PER_LONG == 32
1012 extern int fcntl_getlk64(struct file *, unsigned int, struct flock64 __user *);
1013 extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
1014 struct flock64 __user *);
1017 extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
1018 extern int fcntl_getlease(struct file *filp);
1021 void locks_free_lock_context(struct file_lock_context *ctx);
1022 void locks_free_lock(struct file_lock *fl);
1023 extern void locks_init_lock(struct file_lock *);
1024 extern struct file_lock * locks_alloc_lock(void);
1025 extern void locks_copy_lock(struct file_lock *, struct file_lock *);
1026 extern void locks_copy_conflock(struct file_lock *, struct file_lock *);
1027 extern void locks_remove_posix(struct file *, fl_owner_t);
1028 extern void locks_remove_file(struct file *);
1029 extern void locks_release_private(struct file_lock *);
1030 extern void posix_test_lock(struct file *, struct file_lock *);
1031 extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
1032 extern int posix_lock_file_wait(struct file *, struct file_lock *);
1033 extern int posix_unblock_lock(struct file_lock *);
1034 extern int vfs_test_lock(struct file *, struct file_lock *);
1035 extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
1036 extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
1037 extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
1038 extern int __break_lease(struct inode *inode, unsigned int flags, unsigned int type);
1039 extern void lease_get_mtime(struct inode *, struct timespec *time);
1040 extern int generic_setlease(struct file *, long, struct file_lock **, void **priv);
1041 extern int vfs_setlease(struct file *, long, struct file_lock **, void **);
1042 extern int lease_modify(struct file_lock *, int, struct list_head *);
1043 #else /* !CONFIG_FILE_LOCKING */
1044 static inline int fcntl_getlk(struct file *file, unsigned int cmd,
1045 struct flock __user *user)
1050 static inline int fcntl_setlk(unsigned int fd, struct file *file,
1051 unsigned int cmd, struct flock __user *user)
1056 #if BITS_PER_LONG == 32
1057 static inline int fcntl_getlk64(struct file *file, unsigned int cmd,
1058 struct flock64 __user *user)
1063 static inline int fcntl_setlk64(unsigned int fd, struct file *file,
1064 unsigned int cmd, struct flock64 __user *user)
1069 static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
1074 static inline int fcntl_getlease(struct file *filp)
1080 locks_free_lock_context(struct file_lock_context *ctx)
1084 static inline void locks_init_lock(struct file_lock *fl)
1089 static inline void locks_copy_conflock(struct file_lock *new, struct file_lock *fl)
1094 static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1099 static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
1104 static inline void locks_remove_file(struct file *filp)
1109 static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
1114 static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
1115 struct file_lock *conflock)
1120 static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
1125 static inline int posix_unblock_lock(struct file_lock *waiter)
1130 static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
1135 static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
1136 struct file_lock *fl, struct file_lock *conf)
1141 static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
1146 static inline int flock_lock_file_wait(struct file *filp,
1147 struct file_lock *request)
1152 static inline int __break_lease(struct inode *inode, unsigned int mode, unsigned int type)
1157 static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
1162 static inline int generic_setlease(struct file *filp, long arg,
1163 struct file_lock **flp, void **priv)
1168 static inline int vfs_setlease(struct file *filp, long arg,
1169 struct file_lock **lease, void **priv)
1174 static inline int lease_modify(struct file_lock *fl, int arg,
1175 struct list_head *dispose)
1179 #endif /* !CONFIG_FILE_LOCKING */
1182 struct fasync_struct {
1186 struct fasync_struct *fa_next; /* singly linked list */
1187 struct file *fa_file;
1188 struct rcu_head fa_rcu;
1191 #define FASYNC_MAGIC 0x4601
1193 /* SMP safe fasync helpers: */
1194 extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
1195 extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
1196 extern int fasync_remove_entry(struct file *, struct fasync_struct **);
1197 extern struct fasync_struct *fasync_alloc(void);
1198 extern void fasync_free(struct fasync_struct *);
1200 /* can be called from interrupts */
1201 extern void kill_fasync(struct fasync_struct **, int, int);
1203 extern void __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
1204 extern void f_setown(struct file *filp, unsigned long arg, int force);
1205 extern void f_delown(struct file *filp);
1206 extern pid_t f_getown(struct file *filp);
1207 extern int send_sigurg(struct fown_struct *fown);
1215 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1216 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1217 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1218 #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1219 #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
1222 /* Possible states of 'frozen' field */
1224 SB_UNFROZEN = 0, /* FS is unfrozen */
1225 SB_FREEZE_WRITE = 1, /* Writes, dir ops, ioctls frozen */
1226 SB_FREEZE_PAGEFAULT = 2, /* Page faults stopped as well */
1227 SB_FREEZE_FS = 3, /* For internal FS use (e.g. to stop
1228 * internal threads if needed) */
1229 SB_FREEZE_COMPLETE = 4, /* ->freeze_fs finished successfully */
1232 #define SB_FREEZE_LEVELS (SB_FREEZE_COMPLETE - 1)
1235 /* Counters for counting writers at each level */
1236 struct percpu_counter counter[SB_FREEZE_LEVELS];
1237 wait_queue_head_t wait; /* queue for waiting for
1238 writers / faults to finish */
1239 int frozen; /* Is sb frozen? */
1240 wait_queue_head_t wait_unfrozen; /* queue for waiting for
1242 #ifdef CONFIG_DEBUG_LOCK_ALLOC
1243 struct lockdep_map lock_map[SB_FREEZE_LEVELS];
1247 struct super_block {
1248 struct list_head s_list; /* Keep this first */
1249 dev_t s_dev; /* search index; _not_ kdev_t */
1250 unsigned char s_blocksize_bits;
1251 unsigned long s_blocksize;
1252 loff_t s_maxbytes; /* Max file size */
1253 struct file_system_type *s_type;
1254 const struct super_operations *s_op;
1255 const struct dquot_operations *dq_op;
1256 const struct quotactl_ops *s_qcop;
1257 const struct export_operations *s_export_op;
1258 unsigned long s_flags;
1259 unsigned long s_magic;
1260 struct dentry *s_root;
1261 struct rw_semaphore s_umount;
1264 #ifdef CONFIG_SECURITY
1267 const struct xattr_handler **s_xattr;
1269 struct list_head s_inodes; /* all inodes */
1270 struct hlist_bl_head s_anon; /* anonymous dentries for (nfs) exporting */
1271 struct list_head s_mounts; /* list of mounts; _not_ for fs use */
1272 struct block_device *s_bdev;
1273 struct backing_dev_info *s_bdi;
1274 struct mtd_info *s_mtd;
1275 struct hlist_node s_instances;
1276 unsigned int s_quota_types; /* Bitmask of supported quota types */
1277 struct quota_info s_dquot; /* Diskquota specific options */
1279 struct sb_writers s_writers;
1281 char s_id[32]; /* Informational name */
1282 u8 s_uuid[16]; /* UUID */
1284 void *s_fs_info; /* Filesystem private info */
1285 unsigned int s_max_links;
1288 /* Granularity of c/m/atime in ns.
1289 Cannot be worse than a second */
1293 * The next field is for VFS *only*. No filesystems have any business
1294 * even looking at it. You had been warned.
1296 struct mutex s_vfs_rename_mutex; /* Kludge */
1299 * Filesystem subtype. If non-empty the filesystem type field
1300 * in /proc/mounts will be "type.subtype"
1305 * Saved mount options for lazy filesystems using
1306 * generic_show_options()
1308 char __rcu *s_options;
1309 const struct dentry_operations *s_d_op; /* default d_op for dentries */
1312 * Saved pool identifier for cleancache (-1 means none)
1314 int cleancache_poolid;
1316 struct shrinker s_shrink; /* per-sb shrinker handle */
1318 /* Number of inodes with nlink == 0 but still referenced */
1319 atomic_long_t s_remove_count;
1321 /* Being remounted read-only */
1322 int s_readonly_remount;
1324 /* AIO completions deferred from interrupt context */
1325 struct workqueue_struct *s_dio_done_wq;
1326 struct hlist_head s_pins;
1329 * Keep the lru lists last in the structure so they always sit on their
1330 * own individual cachelines.
1332 struct list_lru s_dentry_lru ____cacheline_aligned_in_smp;
1333 struct list_lru s_inode_lru ____cacheline_aligned_in_smp;
1334 struct rcu_head rcu;
1337 * Indicates how deep in a filesystem stack this SB is
1342 extern struct timespec current_fs_time(struct super_block *sb);
1345 * Snapshotting support.
1348 void __sb_end_write(struct super_block *sb, int level);
1349 int __sb_start_write(struct super_block *sb, int level, bool wait);
1352 * sb_end_write - drop write access to a superblock
1353 * @sb: the super we wrote to
1355 * Decrement number of writers to the filesystem. Wake up possible waiters
1356 * wanting to freeze the filesystem.
1358 static inline void sb_end_write(struct super_block *sb)
1360 __sb_end_write(sb, SB_FREEZE_WRITE);
1364 * sb_end_pagefault - drop write access to a superblock from a page fault
1365 * @sb: the super we wrote to
1367 * Decrement number of processes handling write page fault to the filesystem.
1368 * Wake up possible waiters wanting to freeze the filesystem.
1370 static inline void sb_end_pagefault(struct super_block *sb)
1372 __sb_end_write(sb, SB_FREEZE_PAGEFAULT);
1376 * sb_end_intwrite - drop write access to a superblock for internal fs purposes
1377 * @sb: the super we wrote to
1379 * Decrement fs-internal number of writers to the filesystem. Wake up possible
1380 * waiters wanting to freeze the filesystem.
1382 static inline void sb_end_intwrite(struct super_block *sb)
1384 __sb_end_write(sb, SB_FREEZE_FS);
1388 * sb_start_write - get write access to a superblock
1389 * @sb: the super we write to
1391 * When a process wants to write data or metadata to a file system (i.e. dirty
1392 * a page or an inode), it should embed the operation in a sb_start_write() -
1393 * sb_end_write() pair to get exclusion against file system freezing. This
1394 * function increments number of writers preventing freezing. If the file
1395 * system is already frozen, the function waits until the file system is
1398 * Since freeze protection behaves as a lock, users have to preserve
1399 * ordering of freeze protection and other filesystem locks. Generally,
1400 * freeze protection should be the outermost lock. In particular, we have:
1403 * -> i_mutex (write path, truncate, directory ops, ...)
1404 * -> s_umount (freeze_super, thaw_super)
1406 static inline void sb_start_write(struct super_block *sb)
1408 __sb_start_write(sb, SB_FREEZE_WRITE, true);
1411 static inline int sb_start_write_trylock(struct super_block *sb)
1413 return __sb_start_write(sb, SB_FREEZE_WRITE, false);
1417 * sb_start_pagefault - get write access to a superblock from a page fault
1418 * @sb: the super we write to
1420 * When a process starts handling write page fault, it should embed the
1421 * operation into sb_start_pagefault() - sb_end_pagefault() pair to get
1422 * exclusion against file system freezing. This is needed since the page fault
1423 * is going to dirty a page. This function increments number of running page
1424 * faults preventing freezing. If the file system is already frozen, the
1425 * function waits until the file system is thawed.
1427 * Since page fault freeze protection behaves as a lock, users have to preserve
1428 * ordering of freeze protection and other filesystem locks. It is advised to
1429 * put sb_start_pagefault() close to mmap_sem in lock ordering. Page fault
1430 * handling code implies lock dependency:
1433 * -> sb_start_pagefault
1435 static inline void sb_start_pagefault(struct super_block *sb)
1437 __sb_start_write(sb, SB_FREEZE_PAGEFAULT, true);
1441 * sb_start_intwrite - get write access to a superblock for internal fs purposes
1442 * @sb: the super we write to
1444 * This is the third level of protection against filesystem freezing. It is
1445 * free for use by a filesystem. The only requirement is that it must rank
1446 * below sb_start_pagefault.
1448 * For example filesystem can call sb_start_intwrite() when starting a
1449 * transaction which somewhat eases handling of freezing for internal sources
1450 * of filesystem changes (internal fs threads, discarding preallocation on file
1453 static inline void sb_start_intwrite(struct super_block *sb)
1455 __sb_start_write(sb, SB_FREEZE_FS, true);
1459 extern bool inode_owner_or_capable(const struct inode *inode);
1462 * VFS helper functions..
1464 extern int vfs_create(struct inode *, struct dentry *, umode_t, bool);
1465 extern int vfs_mkdir(struct inode *, struct dentry *, umode_t);
1466 extern int vfs_mknod(struct inode *, struct dentry *, umode_t, dev_t);
1467 extern int vfs_symlink(struct inode *, struct dentry *, const char *);
1468 extern int vfs_link(struct dentry *, struct inode *, struct dentry *, struct inode **);
1469 extern int vfs_rmdir(struct inode *, struct dentry *);
1470 extern int vfs_unlink(struct inode *, struct dentry *, struct inode **);
1471 extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *, struct inode **, unsigned int);
1472 extern int vfs_whiteout(struct inode *, struct dentry *);
1475 * VFS dentry helper functions.
1477 extern void dentry_unhash(struct dentry *dentry);
1480 * VFS file helper functions.
1482 extern void inode_init_owner(struct inode *inode, const struct inode *dir,
1485 * VFS FS_IOC_FIEMAP helper definitions.
1487 struct fiemap_extent_info {
1488 unsigned int fi_flags; /* Flags as passed from user */
1489 unsigned int fi_extents_mapped; /* Number of mapped extents */
1490 unsigned int fi_extents_max; /* Size of fiemap_extent array */
1491 struct fiemap_extent __user *fi_extents_start; /* Start of
1492 fiemap_extent array */
1494 int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
1495 u64 phys, u64 len, u32 flags);
1496 int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
1501 * NOTE! These match bits 12..15 of stat.st_mode
1502 * (ie "(i_mode >> 12) & 15").
1504 #define DT_UNKNOWN 0
1515 * This is the "filldir" function type, used by readdir() to let
1516 * the kernel specify what kind of dirent layout it wants to have.
1517 * This allows the kernel to read directories into kernel space or
1518 * to have different dirent layouts depending on the binary type.
1521 typedef int (*filldir_t)(struct dir_context *, const char *, int, loff_t, u64,
1524 struct dir_context {
1525 const filldir_t actor;
1529 struct block_device_operations;
1531 /* These macros are for out of kernel modules to test that
1532 * the kernel supports the unlocked_ioctl and compat_ioctl
1533 * fields in struct file_operations. */
1534 #define HAVE_COMPAT_IOCTL 1
1535 #define HAVE_UNLOCKED_IOCTL 1
1538 * These flags let !MMU mmap() govern direct device mapping vs immediate
1539 * copying more easily for MAP_PRIVATE, especially for ROM filesystems.
1541 * NOMMU_MAP_COPY: Copy can be mapped (MAP_PRIVATE)
1542 * NOMMU_MAP_DIRECT: Can be mapped directly (MAP_SHARED)
1543 * NOMMU_MAP_READ: Can be mapped for reading
1544 * NOMMU_MAP_WRITE: Can be mapped for writing
1545 * NOMMU_MAP_EXEC: Can be mapped for execution
1547 #define NOMMU_MAP_COPY 0x00000001
1548 #define NOMMU_MAP_DIRECT 0x00000008
1549 #define NOMMU_MAP_READ VM_MAYREAD
1550 #define NOMMU_MAP_WRITE VM_MAYWRITE
1551 #define NOMMU_MAP_EXEC VM_MAYEXEC
1553 #define NOMMU_VMFLAGS \
1554 (NOMMU_MAP_READ | NOMMU_MAP_WRITE | NOMMU_MAP_EXEC)
1559 struct file_operations {
1560 struct module *owner;
1561 loff_t (*llseek) (struct file *, loff_t, int);
1562 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1563 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
1564 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1565 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1566 ssize_t (*read_iter) (struct kiocb *, struct iov_iter *);
1567 ssize_t (*write_iter) (struct kiocb *, struct iov_iter *);
1568 int (*iterate) (struct file *, struct dir_context *);
1569 unsigned int (*poll) (struct file *, struct poll_table_struct *);
1570 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1571 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1572 int (*mmap) (struct file *, struct vm_area_struct *);
1573 void (*mremap)(struct file *, struct vm_area_struct *);
1574 int (*open) (struct inode *, struct file *);
1575 int (*flush) (struct file *, fl_owner_t id);
1576 int (*release) (struct inode *, struct file *);
1577 int (*fsync) (struct file *, loff_t, loff_t, int datasync);
1578 int (*aio_fsync) (struct kiocb *, int datasync);
1579 int (*fasync) (int, struct file *, int);
1580 int (*lock) (struct file *, int, struct file_lock *);
1581 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1582 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1583 int (*check_flags)(int);
1584 int (*flock) (struct file *, int, struct file_lock *);
1585 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1586 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
1587 int (*setlease)(struct file *, long, struct file_lock **, void **);
1588 long (*fallocate)(struct file *file, int mode, loff_t offset,
1590 void (*show_fdinfo)(struct seq_file *m, struct file *f);
1592 unsigned (*mmap_capabilities)(struct file *);
1596 struct inode_operations {
1597 struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int);
1598 void * (*follow_link) (struct dentry *, struct nameidata *);
1599 int (*permission) (struct inode *, int);
1600 struct posix_acl * (*get_acl)(struct inode *, int);
1602 int (*readlink) (struct dentry *, char __user *,int);
1603 void (*put_link) (struct dentry *, struct nameidata *, void *);
1605 int (*create) (struct inode *,struct dentry *, umode_t, bool);
1606 int (*link) (struct dentry *,struct inode *,struct dentry *);
1607 int (*unlink) (struct inode *,struct dentry *);
1608 int (*symlink) (struct inode *,struct dentry *,const char *);
1609 int (*mkdir) (struct inode *,struct dentry *,umode_t);
1610 int (*rmdir) (struct inode *,struct dentry *);
1611 int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t);
1612 int (*rename) (struct inode *, struct dentry *,
1613 struct inode *, struct dentry *);
1614 int (*rename2) (struct inode *, struct dentry *,
1615 struct inode *, struct dentry *, unsigned int);
1616 int (*setattr) (struct dentry *, struct iattr *);
1617 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1618 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1619 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1620 ssize_t (*listxattr) (struct dentry *, char *, size_t);
1621 int (*removexattr) (struct dentry *, const char *);
1622 int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
1624 int (*update_time)(struct inode *, struct timespec *, int);
1625 int (*atomic_open)(struct inode *, struct dentry *,
1626 struct file *, unsigned open_flag,
1627 umode_t create_mode, int *opened);
1628 int (*tmpfile) (struct inode *, struct dentry *, umode_t);
1629 int (*set_acl)(struct inode *, struct posix_acl *, int);
1631 /* WARNING: probably going away soon, do not use! */
1632 int (*dentry_open)(struct dentry *, struct file *, const struct cred *);
1633 } ____cacheline_aligned;
1635 ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
1636 unsigned long nr_segs, unsigned long fast_segs,
1637 struct iovec *fast_pointer,
1638 struct iovec **ret_pointer);
1640 extern ssize_t __vfs_read(struct file *, char __user *, size_t, loff_t *);
1641 extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1642 extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1643 extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1644 unsigned long, loff_t *);
1645 extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1646 unsigned long, loff_t *);
1648 struct super_operations {
1649 struct inode *(*alloc_inode)(struct super_block *sb);
1650 void (*destroy_inode)(struct inode *);
1652 void (*dirty_inode) (struct inode *, int flags);
1653 int (*write_inode) (struct inode *, struct writeback_control *wbc);
1654 int (*drop_inode) (struct inode *);
1655 void (*evict_inode) (struct inode *);
1656 void (*put_super) (struct super_block *);
1657 int (*sync_fs)(struct super_block *sb, int wait);
1658 int (*freeze_super) (struct super_block *);
1659 int (*freeze_fs) (struct super_block *);
1660 int (*thaw_super) (struct super_block *);
1661 int (*unfreeze_fs) (struct super_block *);
1662 int (*statfs) (struct dentry *, struct kstatfs *);
1663 int (*remount_fs) (struct super_block *, int *, char *);
1664 void (*umount_begin) (struct super_block *);
1666 int (*show_options)(struct seq_file *, struct dentry *);
1667 int (*show_devname)(struct seq_file *, struct dentry *);
1668 int (*show_path)(struct seq_file *, struct dentry *);
1669 int (*show_stats)(struct seq_file *, struct dentry *);
1671 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1672 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
1673 struct dquot **(*get_dquots)(struct inode *);
1675 int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
1676 long (*nr_cached_objects)(struct super_block *,
1677 struct shrink_control *);
1678 long (*free_cached_objects)(struct super_block *,
1679 struct shrink_control *);
1683 * Inode flags - they have no relation to superblock flags now
1685 #define S_SYNC 1 /* Writes are synced at once */
1686 #define S_NOATIME 2 /* Do not update access times */
1687 #define S_APPEND 4 /* Append-only file */
1688 #define S_IMMUTABLE 8 /* Immutable file */
1689 #define S_DEAD 16 /* removed, but still open directory */
1690 #define S_NOQUOTA 32 /* Inode is not counted to quota */
1691 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
1692 #define S_NOCMTIME 128 /* Do not update file c/mtime */
1693 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
1694 #define S_PRIVATE 512 /* Inode is fs-internal */
1695 #define S_IMA 1024 /* Inode has an associated IMA struct */
1696 #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
1697 #define S_NOSEC 4096 /* no suid or xattr security attributes */
1698 #ifdef CONFIG_FS_DAX
1699 #define S_DAX 8192 /* Direct Access, avoiding the page cache */
1701 #define S_DAX 0 /* Make all the DAX code disappear */
1705 * Note that nosuid etc flags are inode-specific: setting some file-system
1706 * flags just means all the inodes inherit those flags by default. It might be
1707 * possible to override it selectively if you really wanted to with some
1708 * ioctl() that is not currently implemented.
1710 * Exception: MS_RDONLY is always applied to the entire file system.
1712 * Unfortunately, it is possible to change a filesystems flags with it mounted
1713 * with files in use. This means that all of the inodes will not have their
1714 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
1715 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
1717 #define __IS_FLG(inode, flg) ((inode)->i_sb->s_flags & (flg))
1719 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
1720 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
1721 ((inode)->i_flags & S_SYNC))
1722 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
1723 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
1724 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
1725 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
1726 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
1728 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
1729 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
1730 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
1731 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
1733 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
1734 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
1735 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
1736 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
1737 #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
1738 #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
1739 #define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
1740 #define IS_DAX(inode) ((inode)->i_flags & S_DAX)
1742 #define IS_WHITEOUT(inode) (S_ISCHR(inode->i_mode) && \
1743 (inode)->i_rdev == WHITEOUT_DEV)
1746 * Inode state bits. Protected by inode->i_lock
1748 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1749 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1751 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1752 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1753 * various stages of removing an inode.
1755 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1757 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1758 * fdatasync(). i_atime is the usual cause.
1759 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1760 * these changes separately from I_DIRTY_SYNC so that we
1761 * don't have to write inode on fdatasync() when only
1762 * mtime has changed in it.
1763 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1764 * I_NEW Serves as both a mutex and completion notification.
1765 * New inodes set I_NEW. If two processes both create
1766 * the same inode, one of them will release its inode and
1767 * wait for I_NEW to be released before returning.
1768 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1769 * also cause waiting on I_NEW, without I_NEW actually
1770 * being set. find_inode() uses this to prevent returning
1771 * nearly-dead inodes.
1772 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1773 * is zero. I_FREEING must be set when I_WILL_FREE is
1775 * I_FREEING Set when inode is about to be freed but still has dirty
1776 * pages or buffers attached or the inode itself is still
1778 * I_CLEAR Added by clear_inode(). In this state the inode is
1779 * clean and can be destroyed. Inode keeps I_FREEING.
1781 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1782 * prohibited for many purposes. iget() must wait for
1783 * the inode to be completely released, then create it
1784 * anew. Other functions will just ignore such inodes,
1785 * if appropriate. I_NEW is used for waiting.
1787 * I_SYNC Writeback of inode is running. The bit is set during
1788 * data writeback, and cleared with a wakeup on the bit
1789 * address once it is done. The bit is also used to pin
1790 * the inode in memory for flusher thread.
1792 * I_REFERENCED Marks the inode as recently references on the LRU list.
1794 * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit().
1796 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1798 #define I_DIRTY_SYNC (1 << 0)
1799 #define I_DIRTY_DATASYNC (1 << 1)
1800 #define I_DIRTY_PAGES (1 << 2)
1802 #define I_NEW (1 << __I_NEW)
1803 #define I_WILL_FREE (1 << 4)
1804 #define I_FREEING (1 << 5)
1805 #define I_CLEAR (1 << 6)
1807 #define I_SYNC (1 << __I_SYNC)
1808 #define I_REFERENCED (1 << 8)
1809 #define __I_DIO_WAKEUP 9
1810 #define I_DIO_WAKEUP (1 << I_DIO_WAKEUP)
1811 #define I_LINKABLE (1 << 10)
1812 #define I_DIRTY_TIME (1 << 11)
1813 #define __I_DIRTY_TIME_EXPIRED 12
1814 #define I_DIRTY_TIME_EXPIRED (1 << __I_DIRTY_TIME_EXPIRED)
1816 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1817 #define I_DIRTY_ALL (I_DIRTY | I_DIRTY_TIME)
1819 extern void __mark_inode_dirty(struct inode *, int);
1820 static inline void mark_inode_dirty(struct inode *inode)
1822 __mark_inode_dirty(inode, I_DIRTY);
1825 static inline void mark_inode_dirty_sync(struct inode *inode)
1827 __mark_inode_dirty(inode, I_DIRTY_SYNC);
1830 extern void inc_nlink(struct inode *inode);
1831 extern void drop_nlink(struct inode *inode);
1832 extern void clear_nlink(struct inode *inode);
1833 extern void set_nlink(struct inode *inode, unsigned int nlink);
1835 static inline void inode_inc_link_count(struct inode *inode)
1838 mark_inode_dirty(inode);
1841 static inline void inode_dec_link_count(struct inode *inode)
1844 mark_inode_dirty(inode);
1848 * inode_inc_iversion - increments i_version
1849 * @inode: inode that need to be updated
1851 * Every time the inode is modified, the i_version field will be incremented.
1852 * The filesystem has to be mounted with i_version flag
1855 static inline void inode_inc_iversion(struct inode *inode)
1857 spin_lock(&inode->i_lock);
1859 spin_unlock(&inode->i_lock);
1862 enum file_time_flags {
1869 extern void touch_atime(const struct path *);
1870 static inline void file_accessed(struct file *file)
1872 if (!(file->f_flags & O_NOATIME))
1873 touch_atime(&file->f_path);
1876 int sync_inode(struct inode *inode, struct writeback_control *wbc);
1877 int sync_inode_metadata(struct inode *inode, int wait);
1879 struct file_system_type {
1882 #define FS_REQUIRES_DEV 1
1883 #define FS_BINARY_MOUNTDATA 2
1884 #define FS_HAS_SUBTYPE 4
1885 #define FS_USERNS_MOUNT 8 /* Can be mounted by userns root */
1886 #define FS_USERNS_DEV_MOUNT 16 /* A userns mount does not imply MNT_NODEV */
1887 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() during rename() internally. */
1888 struct dentry *(*mount) (struct file_system_type *, int,
1889 const char *, void *);
1890 void (*kill_sb) (struct super_block *);
1891 struct module *owner;
1892 struct file_system_type * next;
1893 struct hlist_head fs_supers;
1895 struct lock_class_key s_lock_key;
1896 struct lock_class_key s_umount_key;
1897 struct lock_class_key s_vfs_rename_key;
1898 struct lock_class_key s_writers_key[SB_FREEZE_LEVELS];
1900 struct lock_class_key i_lock_key;
1901 struct lock_class_key i_mutex_key;
1902 struct lock_class_key i_mutex_dir_key;
1905 #define MODULE_ALIAS_FS(NAME) MODULE_ALIAS("fs-" NAME)
1907 extern struct dentry *mount_ns(struct file_system_type *fs_type, int flags,
1908 void *data, int (*fill_super)(struct super_block *, void *, int));
1909 extern struct dentry *mount_bdev(struct file_system_type *fs_type,
1910 int flags, const char *dev_name, void *data,
1911 int (*fill_super)(struct super_block *, void *, int));
1912 extern struct dentry *mount_single(struct file_system_type *fs_type,
1913 int flags, void *data,
1914 int (*fill_super)(struct super_block *, void *, int));
1915 extern struct dentry *mount_nodev(struct file_system_type *fs_type,
1916 int flags, void *data,
1917 int (*fill_super)(struct super_block *, void *, int));
1918 extern struct dentry *mount_subtree(struct vfsmount *mnt, const char *path);
1919 void generic_shutdown_super(struct super_block *sb);
1920 void kill_block_super(struct super_block *sb);
1921 void kill_anon_super(struct super_block *sb);
1922 void kill_litter_super(struct super_block *sb);
1923 void deactivate_super(struct super_block *sb);
1924 void deactivate_locked_super(struct super_block *sb);
1925 int set_anon_super(struct super_block *s, void *data);
1926 int get_anon_bdev(dev_t *);
1927 void free_anon_bdev(dev_t);
1928 struct super_block *sget(struct file_system_type *type,
1929 int (*test)(struct super_block *,void *),
1930 int (*set)(struct super_block *,void *),
1931 int flags, void *data);
1932 extern struct dentry *mount_pseudo(struct file_system_type *, char *,
1933 const struct super_operations *ops,
1934 const struct dentry_operations *dops,
1937 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1938 #define fops_get(fops) \
1939 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1940 #define fops_put(fops) \
1941 do { if (fops) module_put((fops)->owner); } while(0)
1943 * This one is to be used *ONLY* from ->open() instances.
1944 * fops must be non-NULL, pinned down *and* module dependencies
1945 * should be sufficient to pin the caller down as well.
1947 #define replace_fops(f, fops) \
1949 struct file *__file = (f); \
1950 fops_put(__file->f_op); \
1951 BUG_ON(!(__file->f_op = (fops))); \
1954 extern int register_filesystem(struct file_system_type *);
1955 extern int unregister_filesystem(struct file_system_type *);
1956 extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
1957 #define kern_mount(type) kern_mount_data(type, NULL)
1958 extern void kern_unmount(struct vfsmount *mnt);
1959 extern int may_umount_tree(struct vfsmount *);
1960 extern int may_umount(struct vfsmount *);
1961 extern long do_mount(const char *, const char __user *,
1962 const char *, unsigned long, void *);
1963 extern struct vfsmount *collect_mounts(struct path *);
1964 extern void drop_collected_mounts(struct vfsmount *);
1965 extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
1967 extern int vfs_statfs(struct path *, struct kstatfs *);
1968 extern int user_statfs(const char __user *, struct kstatfs *);
1969 extern int fd_statfs(int, struct kstatfs *);
1970 extern int vfs_ustat(dev_t, struct kstatfs *);
1971 extern int freeze_super(struct super_block *super);
1972 extern int thaw_super(struct super_block *super);
1973 extern bool our_mnt(struct vfsmount *mnt);
1974 extern bool fs_fully_visible(struct file_system_type *);
1976 extern int current_umask(void);
1978 extern void ihold(struct inode * inode);
1979 extern void iput(struct inode *);
1980 extern int generic_update_time(struct inode *, struct timespec *, int);
1982 static inline struct inode *file_inode(const struct file *f)
1988 extern struct kobject *fs_kobj;
1990 #define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
1992 #define FLOCK_VERIFY_READ 1
1993 #define FLOCK_VERIFY_WRITE 2
1995 #ifdef CONFIG_FILE_LOCKING
1996 extern int locks_mandatory_locked(struct file *);
1997 extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
2000 * Candidates for mandatory locking have the setgid bit set
2001 * but no group execute bit - an otherwise meaningless combination.
2004 static inline int __mandatory_lock(struct inode *ino)
2006 return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
2010 * ... and these candidates should be on MS_MANDLOCK mounted fs,
2011 * otherwise these will be advisory locks
2014 static inline int mandatory_lock(struct inode *ino)
2016 return IS_MANDLOCK(ino) && __mandatory_lock(ino);
2019 static inline int locks_verify_locked(struct file *file)
2021 if (mandatory_lock(file_inode(file)))
2022 return locks_mandatory_locked(file);
2026 static inline int locks_verify_truncate(struct inode *inode,
2030 if (inode->i_flctx && mandatory_lock(inode))
2031 return locks_mandatory_area(
2032 FLOCK_VERIFY_WRITE, inode, filp,
2033 size < inode->i_size ? size : inode->i_size,
2034 (size < inode->i_size ? inode->i_size - size
2035 : size - inode->i_size)
2040 static inline int break_lease(struct inode *inode, unsigned int mode)
2043 * Since this check is lockless, we must ensure that any refcounts
2044 * taken are done before checking i_flctx->flc_lease. Otherwise, we
2045 * could end up racing with tasks trying to set a new lease on this
2049 if (inode->i_flctx && !list_empty_careful(&inode->i_flctx->flc_lease))
2050 return __break_lease(inode, mode, FL_LEASE);
2054 static inline int break_deleg(struct inode *inode, unsigned int mode)
2057 * Since this check is lockless, we must ensure that any refcounts
2058 * taken are done before checking i_flctx->flc_lease. Otherwise, we
2059 * could end up racing with tasks trying to set a new lease on this
2063 if (inode->i_flctx && !list_empty_careful(&inode->i_flctx->flc_lease))
2064 return __break_lease(inode, mode, FL_DELEG);
2068 static inline int try_break_deleg(struct inode *inode, struct inode **delegated_inode)
2072 ret = break_deleg(inode, O_WRONLY|O_NONBLOCK);
2073 if (ret == -EWOULDBLOCK && delegated_inode) {
2074 *delegated_inode = inode;
2080 static inline int break_deleg_wait(struct inode **delegated_inode)
2084 ret = break_deleg(*delegated_inode, O_WRONLY);
2085 iput(*delegated_inode);
2086 *delegated_inode = NULL;
2090 static inline int break_layout(struct inode *inode, bool wait)
2093 if (inode->i_flctx && !list_empty_careful(&inode->i_flctx->flc_lease))
2094 return __break_lease(inode,
2095 wait ? O_WRONLY : O_WRONLY | O_NONBLOCK,
2100 #else /* !CONFIG_FILE_LOCKING */
2101 static inline int locks_mandatory_locked(struct file *file)
2106 static inline int locks_mandatory_area(int rw, struct inode *inode,
2107 struct file *filp, loff_t offset,
2113 static inline int __mandatory_lock(struct inode *inode)
2118 static inline int mandatory_lock(struct inode *inode)
2123 static inline int locks_verify_locked(struct file *file)
2128 static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
2134 static inline int break_lease(struct inode *inode, unsigned int mode)
2139 static inline int break_deleg(struct inode *inode, unsigned int mode)
2144 static inline int try_break_deleg(struct inode *inode, struct inode **delegated_inode)
2149 static inline int break_deleg_wait(struct inode **delegated_inode)
2155 static inline int break_layout(struct inode *inode, bool wait)
2160 #endif /* CONFIG_FILE_LOCKING */
2165 const char *name; /* pointer to actual string */
2166 const __user char *uptr; /* original userland pointer */
2167 struct audit_names *aname;
2172 extern long vfs_truncate(struct path *, loff_t);
2173 extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
2175 extern int vfs_fallocate(struct file *file, int mode, loff_t offset,
2177 extern long do_sys_open(int dfd, const char __user *filename, int flags,
2179 extern struct file *file_open_name(struct filename *, int, umode_t);
2180 extern struct file *filp_open(const char *, int, umode_t);
2181 extern struct file *file_open_root(struct dentry *, struct vfsmount *,
2183 extern int vfs_open(const struct path *, struct file *, const struct cred *);
2184 extern struct file * dentry_open(const struct path *, int, const struct cred *);
2185 extern int filp_close(struct file *, fl_owner_t id);
2187 extern struct filename *getname_flags(const char __user *, int, int *);
2188 extern struct filename *getname(const char __user *);
2189 extern struct filename *getname_kernel(const char *);
2190 extern void putname(struct filename *name);
2196 extern int finish_open(struct file *file, struct dentry *dentry,
2197 int (*open)(struct inode *, struct file *),
2199 extern int finish_no_open(struct file *file, struct dentry *dentry);
2203 extern int ioctl_preallocate(struct file *filp, void __user *argp);
2206 extern void __init vfs_caches_init_early(void);
2207 extern void __init vfs_caches_init(unsigned long);
2209 extern struct kmem_cache *names_cachep;
2211 #define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL)
2212 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
2215 extern int register_blkdev(unsigned int, const char *);
2216 extern void unregister_blkdev(unsigned int, const char *);
2217 extern struct block_device *bdget(dev_t);
2218 extern struct block_device *bdgrab(struct block_device *bdev);
2219 extern void bd_set_size(struct block_device *, loff_t size);
2220 extern void bd_forget(struct inode *inode);
2221 extern void bdput(struct block_device *);
2222 extern void invalidate_bdev(struct block_device *);
2223 extern void iterate_bdevs(void (*)(struct block_device *, void *), void *);
2224 extern int sync_blockdev(struct block_device *bdev);
2225 extern void kill_bdev(struct block_device *);
2226 extern struct super_block *freeze_bdev(struct block_device *);
2227 extern void emergency_thaw_all(void);
2228 extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
2229 extern int fsync_bdev(struct block_device *);
2230 extern int sb_is_blkdev_sb(struct super_block *sb);
2232 static inline void bd_forget(struct inode *inode) {}
2233 static inline int sync_blockdev(struct block_device *bdev) { return 0; }
2234 static inline void kill_bdev(struct block_device *bdev) {}
2235 static inline void invalidate_bdev(struct block_device *bdev) {}
2237 static inline struct super_block *freeze_bdev(struct block_device *sb)
2242 static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
2247 static inline void iterate_bdevs(void (*f)(struct block_device *, void *), void *arg)
2251 static inline int sb_is_blkdev_sb(struct super_block *sb)
2256 extern int sync_filesystem(struct super_block *);
2257 extern const struct file_operations def_blk_fops;
2258 extern const struct file_operations def_chr_fops;
2260 extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
2261 extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
2262 extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
2263 extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
2264 extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
2266 extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
2268 extern void blkdev_put(struct block_device *bdev, fmode_t mode);
2270 extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
2271 extern void bd_unlink_disk_holder(struct block_device *bdev,
2272 struct gendisk *disk);
2274 static inline int bd_link_disk_holder(struct block_device *bdev,
2275 struct gendisk *disk)
2279 static inline void bd_unlink_disk_holder(struct block_device *bdev,
2280 struct gendisk *disk)
2287 #define CHRDEV_MAJOR_HASH_SIZE 255
2288 extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
2289 extern int register_chrdev_region(dev_t, unsigned, const char *);
2290 extern int __register_chrdev(unsigned int major, unsigned int baseminor,
2291 unsigned int count, const char *name,
2292 const struct file_operations *fops);
2293 extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
2294 unsigned int count, const char *name);
2295 extern void unregister_chrdev_region(dev_t, unsigned);
2296 extern void chrdev_show(struct seq_file *,off_t);
2298 static inline int register_chrdev(unsigned int major, const char *name,
2299 const struct file_operations *fops)
2301 return __register_chrdev(major, 0, 256, name, fops);
2304 static inline void unregister_chrdev(unsigned int major, const char *name)
2306 __unregister_chrdev(major, 0, 256, name);
2309 /* fs/block_dev.c */
2310 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
2311 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
2314 #define BLKDEV_MAJOR_HASH_SIZE 255
2315 extern const char *__bdevname(dev_t, char *buffer);
2316 extern const char *bdevname(struct block_device *bdev, char *buffer);
2317 extern struct block_device *lookup_bdev(const char *);
2318 extern void blkdev_show(struct seq_file *,off_t);
2321 #define BLKDEV_MAJOR_HASH_SIZE 0
2324 extern void init_special_inode(struct inode *, umode_t, dev_t);
2326 /* Invalid inode operations -- fs/bad_inode.c */
2327 extern void make_bad_inode(struct inode *);
2328 extern int is_bad_inode(struct inode *);
2332 * return READ, READA, or WRITE
2334 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2337 * return data direction, READ or WRITE
2339 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
2341 extern void check_disk_size_change(struct gendisk *disk,
2342 struct block_device *bdev);
2343 extern int revalidate_disk(struct gendisk *);
2344 extern int check_disk_change(struct block_device *);
2345 extern int __invalidate_device(struct block_device *, bool);
2346 extern int invalidate_partition(struct gendisk *, int);
2348 unsigned long invalidate_mapping_pages(struct address_space *mapping,
2349 pgoff_t start, pgoff_t end);
2351 static inline void invalidate_remote_inode(struct inode *inode)
2353 if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
2354 S_ISLNK(inode->i_mode))
2355 invalidate_mapping_pages(inode->i_mapping, 0, -1);
2357 extern int invalidate_inode_pages2(struct address_space *mapping);
2358 extern int invalidate_inode_pages2_range(struct address_space *mapping,
2359 pgoff_t start, pgoff_t end);
2360 extern int write_inode_now(struct inode *, int);
2361 extern int filemap_fdatawrite(struct address_space *);
2362 extern int filemap_flush(struct address_space *);
2363 extern int filemap_fdatawait(struct address_space *);
2364 extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
2366 extern int filemap_write_and_wait(struct address_space *mapping);
2367 extern int filemap_write_and_wait_range(struct address_space *mapping,
2368 loff_t lstart, loff_t lend);
2369 extern int __filemap_fdatawrite_range(struct address_space *mapping,
2370 loff_t start, loff_t end, int sync_mode);
2371 extern int filemap_fdatawrite_range(struct address_space *mapping,
2372 loff_t start, loff_t end);
2374 extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
2376 extern int vfs_fsync(struct file *file, int datasync);
2377 static inline int generic_write_sync(struct file *file, loff_t pos, loff_t count)
2379 if (!(file->f_flags & O_DSYNC) && !IS_SYNC(file->f_mapping->host))
2381 return vfs_fsync_range(file, pos, pos + count - 1,
2382 (file->f_flags & __O_SYNC) ? 0 : 1);
2384 extern void emergency_sync(void);
2385 extern void emergency_remount(void);
2387 extern sector_t bmap(struct inode *, sector_t);
2389 extern int notify_change(struct dentry *, struct iattr *, struct inode **);
2390 extern int inode_permission(struct inode *, int);
2391 extern int __inode_permission(struct inode *, int);
2392 extern int generic_permission(struct inode *, int);
2393 extern int __check_sticky(struct inode *dir, struct inode *inode);
2395 static inline bool execute_ok(struct inode *inode)
2397 return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
2400 static inline void file_start_write(struct file *file)
2402 if (!S_ISREG(file_inode(file)->i_mode))
2404 __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, true);
2407 static inline bool file_start_write_trylock(struct file *file)
2409 if (!S_ISREG(file_inode(file)->i_mode))
2411 return __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, false);
2414 static inline void file_end_write(struct file *file)
2416 if (!S_ISREG(file_inode(file)->i_mode))
2418 __sb_end_write(file_inode(file)->i_sb, SB_FREEZE_WRITE);
2422 * get_write_access() gets write permission for a file.
2423 * put_write_access() releases this write permission.
2424 * This is used for regular files.
2425 * We cannot support write (and maybe mmap read-write shared) accesses and
2426 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2427 * can have the following values:
2428 * 0: no writers, no VM_DENYWRITE mappings
2429 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2430 * > 0: (i_writecount) users are writing to the file.
2432 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2433 * except for the cases where we don't hold i_writecount yet. Then we need to
2434 * use {get,deny}_write_access() - these functions check the sign and refuse
2435 * to do the change if sign is wrong.
2437 static inline int get_write_access(struct inode *inode)
2439 return atomic_inc_unless_negative(&inode->i_writecount) ? 0 : -ETXTBSY;
2441 static inline int deny_write_access(struct file *file)
2443 struct inode *inode = file_inode(file);
2444 return atomic_dec_unless_positive(&inode->i_writecount) ? 0 : -ETXTBSY;
2446 static inline void put_write_access(struct inode * inode)
2448 atomic_dec(&inode->i_writecount);
2450 static inline void allow_write_access(struct file *file)
2453 atomic_inc(&file_inode(file)->i_writecount);
2455 static inline bool inode_is_open_for_write(const struct inode *inode)
2457 return atomic_read(&inode->i_writecount) > 0;
2461 static inline void i_readcount_dec(struct inode *inode)
2463 BUG_ON(!atomic_read(&inode->i_readcount));
2464 atomic_dec(&inode->i_readcount);
2466 static inline void i_readcount_inc(struct inode *inode)
2468 atomic_inc(&inode->i_readcount);
2471 static inline void i_readcount_dec(struct inode *inode)
2475 static inline void i_readcount_inc(struct inode *inode)
2480 extern int do_pipe_flags(int *, int);
2482 extern int kernel_read(struct file *, loff_t, char *, unsigned long);
2483 extern ssize_t kernel_write(struct file *, const char *, size_t, loff_t);
2484 extern ssize_t __kernel_write(struct file *, const char *, size_t, loff_t *);
2485 extern struct file * open_exec(const char *);
2487 /* fs/dcache.c -- generic fs support functions */
2488 extern int is_subdir(struct dentry *, struct dentry *);
2489 extern int path_is_under(struct path *, struct path *);
2491 #include <linux/err.h>
2493 /* needed for stackable file system support */
2494 extern loff_t default_llseek(struct file *file, loff_t offset, int whence);
2496 extern loff_t vfs_llseek(struct file *file, loff_t offset, int whence);
2498 extern int inode_init_always(struct super_block *, struct inode *);
2499 extern void inode_init_once(struct inode *);
2500 extern void address_space_init_once(struct address_space *mapping);
2501 extern struct inode * igrab(struct inode *);
2502 extern ino_t iunique(struct super_block *, ino_t);
2503 extern int inode_needs_sync(struct inode *inode);
2504 extern int generic_delete_inode(struct inode *inode);
2505 static inline int generic_drop_inode(struct inode *inode)
2507 return !inode->i_nlink || inode_unhashed(inode);
2510 extern struct inode *ilookup5_nowait(struct super_block *sb,
2511 unsigned long hashval, int (*test)(struct inode *, void *),
2513 extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
2514 int (*test)(struct inode *, void *), void *data);
2515 extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
2517 extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
2518 extern struct inode * iget_locked(struct super_block *, unsigned long);
2519 extern struct inode *find_inode_nowait(struct super_block *,
2521 int (*match)(struct inode *,
2522 unsigned long, void *),
2524 extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
2525 extern int insert_inode_locked(struct inode *);
2526 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2527 extern void lockdep_annotate_inode_mutex_key(struct inode *inode);
2529 static inline void lockdep_annotate_inode_mutex_key(struct inode *inode) { };
2531 extern void unlock_new_inode(struct inode *);
2532 extern unsigned int get_next_ino(void);
2534 extern void __iget(struct inode * inode);
2535 extern void iget_failed(struct inode *);
2536 extern void clear_inode(struct inode *);
2537 extern void __destroy_inode(struct inode *);
2538 extern struct inode *new_inode_pseudo(struct super_block *sb);
2539 extern struct inode *new_inode(struct super_block *sb);
2540 extern void free_inode_nonrcu(struct inode *inode);
2541 extern int should_remove_suid(struct dentry *);
2542 extern int file_remove_suid(struct file *);
2544 extern void __insert_inode_hash(struct inode *, unsigned long hashval);
2545 static inline void insert_inode_hash(struct inode *inode)
2547 __insert_inode_hash(inode, inode->i_ino);
2550 extern void __remove_inode_hash(struct inode *);
2551 static inline void remove_inode_hash(struct inode *inode)
2553 if (!inode_unhashed(inode))
2554 __remove_inode_hash(inode);
2557 extern void inode_sb_list_add(struct inode *inode);
2560 extern void submit_bio(int, struct bio *);
2561 extern int bdev_read_only(struct block_device *);
2563 extern int set_blocksize(struct block_device *, int);
2564 extern int sb_set_blocksize(struct super_block *, int);
2565 extern int sb_min_blocksize(struct super_block *, int);
2567 extern int generic_file_mmap(struct file *, struct vm_area_struct *);
2568 extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
2569 int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
2570 extern ssize_t generic_file_read_iter(struct kiocb *, struct iov_iter *);
2571 extern ssize_t __generic_file_write_iter(struct kiocb *, struct iov_iter *);
2572 extern ssize_t generic_file_write_iter(struct kiocb *, struct iov_iter *);
2573 extern ssize_t generic_file_direct_write(struct kiocb *, struct iov_iter *, loff_t);
2574 extern ssize_t generic_perform_write(struct file *, struct iov_iter *, loff_t);
2575 extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
2576 extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
2577 extern ssize_t new_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
2578 extern ssize_t new_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
2580 ssize_t vfs_iter_read(struct file *file, struct iov_iter *iter, loff_t *ppos);
2581 ssize_t vfs_iter_write(struct file *file, struct iov_iter *iter, loff_t *ppos);
2583 /* fs/block_dev.c */
2584 extern ssize_t blkdev_read_iter(struct kiocb *iocb, struct iov_iter *to);
2585 extern ssize_t blkdev_write_iter(struct kiocb *iocb, struct iov_iter *from);
2586 extern int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
2588 extern void block_sync_page(struct page *page);
2591 extern ssize_t generic_file_splice_read(struct file *, loff_t *,
2592 struct pipe_inode_info *, size_t, unsigned int);
2593 extern ssize_t default_file_splice_read(struct file *, loff_t *,
2594 struct pipe_inode_info *, size_t, unsigned int);
2595 extern ssize_t iter_file_splice_write(struct pipe_inode_info *,
2596 struct file *, loff_t *, size_t, unsigned int);
2597 extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
2598 struct file *out, loff_t *, size_t len, unsigned int flags);
2599 extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
2600 loff_t *opos, size_t len, unsigned int flags);
2604 file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
2605 extern loff_t noop_llseek(struct file *file, loff_t offset, int whence);
2606 extern loff_t no_llseek(struct file *file, loff_t offset, int whence);
2607 extern loff_t vfs_setpos(struct file *file, loff_t offset, loff_t maxsize);
2608 extern loff_t generic_file_llseek(struct file *file, loff_t offset, int whence);
2609 extern loff_t generic_file_llseek_size(struct file *file, loff_t offset,
2610 int whence, loff_t maxsize, loff_t eof);
2611 extern loff_t fixed_size_llseek(struct file *file, loff_t offset,
2612 int whence, loff_t size);
2613 extern int generic_file_open(struct inode * inode, struct file * filp);
2614 extern int nonseekable_open(struct inode * inode, struct file * filp);
2616 ssize_t dax_do_io(int rw, struct kiocb *, struct inode *, struct iov_iter *,
2617 loff_t, get_block_t, dio_iodone_t, int flags);
2618 int dax_clear_blocks(struct inode *, sector_t block, long size);
2619 int dax_zero_page_range(struct inode *, loff_t from, unsigned len, get_block_t);
2620 int dax_truncate_page(struct inode *, loff_t from, get_block_t);
2621 int dax_fault(struct vm_area_struct *, struct vm_fault *, get_block_t);
2622 #define dax_mkwrite(vma, vmf, gb) dax_fault(vma, vmf, gb)
2625 typedef void (dio_submit_t)(int rw, struct bio *bio, struct inode *inode,
2626 loff_t file_offset);
2629 /* need locking between buffered and direct access */
2632 /* filesystem does not support filling holes */
2633 DIO_SKIP_HOLES = 0x02,
2635 /* filesystem can handle aio writes beyond i_size */
2636 DIO_ASYNC_EXTEND = 0x04,
2639 void dio_end_io(struct bio *bio, int error);
2641 ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
2642 struct block_device *bdev, struct iov_iter *iter, loff_t offset,
2643 get_block_t get_block, dio_iodone_t end_io,
2644 dio_submit_t submit_io, int flags);
2646 static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
2647 struct inode *inode, struct iov_iter *iter, loff_t offset,
2648 get_block_t get_block)
2650 return __blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iter,
2651 offset, get_block, NULL, NULL,
2652 DIO_LOCKING | DIO_SKIP_HOLES);
2656 void inode_dio_wait(struct inode *inode);
2657 void inode_dio_done(struct inode *inode);
2659 extern void inode_set_flags(struct inode *inode, unsigned int flags,
2662 extern const struct file_operations generic_ro_fops;
2664 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2666 extern int readlink_copy(char __user *, int, const char *);
2667 extern int page_readlink(struct dentry *, char __user *, int);
2668 extern void *page_follow_link_light(struct dentry *, struct nameidata *);
2669 extern void page_put_link(struct dentry *, struct nameidata *, void *);
2670 extern int __page_symlink(struct inode *inode, const char *symname, int len,
2672 extern int page_symlink(struct inode *inode, const char *symname, int len);
2673 extern const struct inode_operations page_symlink_inode_operations;
2674 extern void kfree_put_link(struct dentry *, struct nameidata *, void *);
2675 extern int generic_readlink(struct dentry *, char __user *, int);
2676 extern void generic_fillattr(struct inode *, struct kstat *);
2677 int vfs_getattr_nosec(struct path *path, struct kstat *stat);
2678 extern int vfs_getattr(struct path *, struct kstat *);
2679 void __inode_add_bytes(struct inode *inode, loff_t bytes);
2680 void inode_add_bytes(struct inode *inode, loff_t bytes);
2681 void __inode_sub_bytes(struct inode *inode, loff_t bytes);
2682 void inode_sub_bytes(struct inode *inode, loff_t bytes);
2683 loff_t inode_get_bytes(struct inode *inode);
2684 void inode_set_bytes(struct inode *inode, loff_t bytes);
2686 extern int vfs_readdir(struct file *, filldir_t, void *);
2687 extern int iterate_dir(struct file *, struct dir_context *);
2689 extern int vfs_stat(const char __user *, struct kstat *);
2690 extern int vfs_lstat(const char __user *, struct kstat *);
2691 extern int vfs_fstat(unsigned int, struct kstat *);
2692 extern int vfs_fstatat(int , const char __user *, struct kstat *, int);
2694 extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
2696 extern int __generic_block_fiemap(struct inode *inode,
2697 struct fiemap_extent_info *fieinfo,
2698 loff_t start, loff_t len,
2699 get_block_t *get_block);
2700 extern int generic_block_fiemap(struct inode *inode,
2701 struct fiemap_extent_info *fieinfo, u64 start,
2702 u64 len, get_block_t *get_block);
2704 extern void get_filesystem(struct file_system_type *fs);
2705 extern void put_filesystem(struct file_system_type *fs);
2706 extern struct file_system_type *get_fs_type(const char *name);
2707 extern struct super_block *get_super(struct block_device *);
2708 extern struct super_block *get_super_thawed(struct block_device *);
2709 extern struct super_block *get_active_super(struct block_device *bdev);
2710 extern void drop_super(struct super_block *sb);
2711 extern void iterate_supers(void (*)(struct super_block *, void *), void *);
2712 extern void iterate_supers_type(struct file_system_type *,
2713 void (*)(struct super_block *, void *), void *);
2715 extern int dcache_dir_open(struct inode *, struct file *);
2716 extern int dcache_dir_close(struct inode *, struct file *);
2717 extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
2718 extern int dcache_readdir(struct file *, struct dir_context *);
2719 extern int simple_setattr(struct dentry *, struct iattr *);
2720 extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2721 extern int simple_statfs(struct dentry *, struct kstatfs *);
2722 extern int simple_open(struct inode *inode, struct file *file);
2723 extern int simple_link(struct dentry *, struct inode *, struct dentry *);
2724 extern int simple_unlink(struct inode *, struct dentry *);
2725 extern int simple_rmdir(struct inode *, struct dentry *);
2726 extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
2727 extern int noop_fsync(struct file *, loff_t, loff_t, int);
2728 extern int simple_empty(struct dentry *);
2729 extern int simple_readpage(struct file *file, struct page *page);
2730 extern int simple_write_begin(struct file *file, struct address_space *mapping,
2731 loff_t pos, unsigned len, unsigned flags,
2732 struct page **pagep, void **fsdata);
2733 extern int simple_write_end(struct file *file, struct address_space *mapping,
2734 loff_t pos, unsigned len, unsigned copied,
2735 struct page *page, void *fsdata);
2736 extern int always_delete_dentry(const struct dentry *);
2737 extern struct inode *alloc_anon_inode(struct super_block *);
2738 extern int simple_nosetlease(struct file *, long, struct file_lock **, void **);
2739 extern const struct dentry_operations simple_dentry_operations;
2741 extern struct dentry *simple_lookup(struct inode *, struct dentry *, unsigned int flags);
2742 extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
2743 extern const struct file_operations simple_dir_operations;
2744 extern const struct inode_operations simple_dir_inode_operations;
2745 struct tree_descr { char *name; const struct file_operations *ops; int mode; };
2746 struct dentry *d_alloc_name(struct dentry *, const char *);
2747 extern int simple_fill_super(struct super_block *, unsigned long, struct tree_descr *);
2748 extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
2749 extern void simple_release_fs(struct vfsmount **mount, int *count);
2751 extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
2752 loff_t *ppos, const void *from, size_t available);
2753 extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
2754 const void __user *from, size_t count);
2756 extern int __generic_file_fsync(struct file *, loff_t, loff_t, int);
2757 extern int generic_file_fsync(struct file *, loff_t, loff_t, int);
2759 extern int generic_check_addressable(unsigned, u64);
2761 #ifdef CONFIG_MIGRATION
2762 extern int buffer_migrate_page(struct address_space *,
2763 struct page *, struct page *,
2766 #define buffer_migrate_page NULL
2769 extern int inode_change_ok(const struct inode *, struct iattr *);
2770 extern int inode_newsize_ok(const struct inode *, loff_t offset);
2771 extern void setattr_copy(struct inode *inode, const struct iattr *attr);
2773 extern int file_update_time(struct file *file);
2775 extern int generic_show_options(struct seq_file *m, struct dentry *root);
2776 extern void save_mount_options(struct super_block *sb, char *options);
2777 extern void replace_mount_options(struct super_block *sb, char *options);
2779 static inline bool io_is_direct(struct file *filp)
2781 return (filp->f_flags & O_DIRECT) || IS_DAX(file_inode(filp));
2784 static inline ino_t parent_ino(struct dentry *dentry)
2789 * Don't strictly need d_lock here? If the parent ino could change
2790 * then surely we'd have a deeper race in the caller?
2792 spin_lock(&dentry->d_lock);
2793 res = dentry->d_parent->d_inode->i_ino;
2794 spin_unlock(&dentry->d_lock);
2798 /* Transaction based IO helpers */
2801 * An argresp is stored in an allocated page and holds the
2802 * size of the argument or response, along with its content
2804 struct simple_transaction_argresp {
2809 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2811 char *simple_transaction_get(struct file *file, const char __user *buf,
2813 ssize_t simple_transaction_read(struct file *file, char __user *buf,
2814 size_t size, loff_t *pos);
2815 int simple_transaction_release(struct inode *inode, struct file *file);
2817 void simple_transaction_set(struct file *file, size_t n);
2820 * simple attribute files
2822 * These attributes behave similar to those in sysfs:
2824 * Writing to an attribute immediately sets a value, an open file can be
2825 * written to multiple times.
2827 * Reading from an attribute creates a buffer from the value that might get
2828 * read with multiple read calls. When the attribute has been read
2829 * completely, no further read calls are possible until the file is opened
2832 * All attributes contain a text representation of a numeric value
2833 * that are accessed with the get() and set() functions.
2835 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2836 static int __fops ## _open(struct inode *inode, struct file *file) \
2838 __simple_attr_check_format(__fmt, 0ull); \
2839 return simple_attr_open(inode, file, __get, __set, __fmt); \
2841 static const struct file_operations __fops = { \
2842 .owner = THIS_MODULE, \
2843 .open = __fops ## _open, \
2844 .release = simple_attr_release, \
2845 .read = simple_attr_read, \
2846 .write = simple_attr_write, \
2847 .llseek = generic_file_llseek, \
2850 static inline __printf(1, 2)
2851 void __simple_attr_check_format(const char *fmt, ...)
2853 /* don't do anything, just let the compiler check the arguments; */
2856 int simple_attr_open(struct inode *inode, struct file *file,
2857 int (*get)(void *, u64 *), int (*set)(void *, u64),
2859 int simple_attr_release(struct inode *inode, struct file *file);
2860 ssize_t simple_attr_read(struct file *file, char __user *buf,
2861 size_t len, loff_t *ppos);
2862 ssize_t simple_attr_write(struct file *file, const char __user *buf,
2863 size_t len, loff_t *ppos);
2866 int proc_nr_files(struct ctl_table *table, int write,
2867 void __user *buffer, size_t *lenp, loff_t *ppos);
2868 int proc_nr_dentry(struct ctl_table *table, int write,
2869 void __user *buffer, size_t *lenp, loff_t *ppos);
2870 int proc_nr_inodes(struct ctl_table *table, int write,
2871 void __user *buffer, size_t *lenp, loff_t *ppos);
2872 int __init get_filesystem_list(char *buf);
2874 #define __FMODE_EXEC ((__force int) FMODE_EXEC)
2875 #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
2877 #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
2878 #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
2879 (flag & __FMODE_NONOTIFY)))
2881 static inline int is_sxid(umode_t mode)
2883 return (mode & S_ISUID) || ((mode & S_ISGID) && (mode & S_IXGRP));
2886 static inline int check_sticky(struct inode *dir, struct inode *inode)
2888 if (!(dir->i_mode & S_ISVTX))
2891 return __check_sticky(dir, inode);
2894 static inline void inode_has_no_xattr(struct inode *inode)
2896 if (!is_sxid(inode->i_mode) && (inode->i_sb->s_flags & MS_NOSEC))
2897 inode->i_flags |= S_NOSEC;
2900 static inline bool is_root_inode(struct inode *inode)
2902 return inode == inode->i_sb->s_root->d_inode;
2905 static inline bool dir_emit(struct dir_context *ctx,
2906 const char *name, int namelen,
2907 u64 ino, unsigned type)
2909 return ctx->actor(ctx, name, namelen, ctx->pos, ino, type) == 0;
2911 static inline bool dir_emit_dot(struct file *file, struct dir_context *ctx)
2913 return ctx->actor(ctx, ".", 1, ctx->pos,
2914 file->f_path.dentry->d_inode->i_ino, DT_DIR) == 0;
2916 static inline bool dir_emit_dotdot(struct file *file, struct dir_context *ctx)
2918 return ctx->actor(ctx, "..", 2, ctx->pos,
2919 parent_ino(file->f_path.dentry), DT_DIR) == 0;
2921 static inline bool dir_emit_dots(struct file *file, struct dir_context *ctx)
2923 if (ctx->pos == 0) {
2924 if (!dir_emit_dot(file, ctx))
2928 if (ctx->pos == 1) {
2929 if (!dir_emit_dotdot(file, ctx))
2935 static inline bool dir_relax(struct inode *inode)
2937 mutex_unlock(&inode->i_mutex);
2938 mutex_lock(&inode->i_mutex);
2939 return !IS_DEADDIR(inode);
2942 #endif /* _LINUX_FS_H */