Version:  2.0.40 2.2.26 2.4.37 3.9 3.10 3.11 3.12 3.13 3.14 3.15 3.16 3.17 3.18 3.19 4.0 4.1 4.2 4.3 4.4 4.5 4.6

Linux/include/linux/fs.h

  1 #ifndef _LINUX_FS_H
  2 #define _LINUX_FS_H
  3 
  4 #include <linux/linkage.h>
  5 #include <linux/wait.h>
  6 #include <linux/kdev_t.h>
  7 #include <linux/dcache.h>
  8 #include <linux/path.h>
  9 #include <linux/stat.h>
 10 #include <linux/cache.h>
 11 #include <linux/list.h>
 12 #include <linux/list_lru.h>
 13 #include <linux/llist.h>
 14 #include <linux/radix-tree.h>
 15 #include <linux/rbtree.h>
 16 #include <linux/init.h>
 17 #include <linux/pid.h>
 18 #include <linux/bug.h>
 19 #include <linux/mutex.h>
 20 #include <linux/rwsem.h>
 21 #include <linux/capability.h>
 22 #include <linux/semaphore.h>
 23 #include <linux/fiemap.h>
 24 #include <linux/rculist_bl.h>
 25 #include <linux/atomic.h>
 26 #include <linux/shrinker.h>
 27 #include <linux/migrate_mode.h>
 28 #include <linux/uidgid.h>
 29 #include <linux/lockdep.h>
 30 #include <linux/percpu-rwsem.h>
 31 #include <linux/blk_types.h>
 32 #include <linux/workqueue.h>
 33 #include <linux/percpu-rwsem.h>
 34 #include <linux/delayed_call.h>
 35 
 36 #include <asm/byteorder.h>
 37 #include <uapi/linux/fs.h>
 38 
 39 struct backing_dev_info;
 40 struct bdi_writeback;
 41 struct export_operations;
 42 struct hd_geometry;
 43 struct iovec;
 44 struct kiocb;
 45 struct kobject;
 46 struct pipe_inode_info;
 47 struct poll_table_struct;
 48 struct kstatfs;
 49 struct vm_area_struct;
 50 struct vfsmount;
 51 struct cred;
 52 struct swap_info_struct;
 53 struct seq_file;
 54 struct workqueue_struct;
 55 struct iov_iter;
 56 struct fscrypt_info;
 57 struct fscrypt_operations;
 58 
 59 extern void __init inode_init(void);
 60 extern void __init inode_init_early(void);
 61 extern void __init files_init(void);
 62 extern void __init files_maxfiles_init(void);
 63 
 64 extern struct files_stat_struct files_stat;
 65 extern unsigned long get_max_files(void);
 66 extern int sysctl_nr_open;
 67 extern struct inodes_stat_t inodes_stat;
 68 extern int leases_enable, lease_break_time;
 69 extern int sysctl_protected_symlinks;
 70 extern int sysctl_protected_hardlinks;
 71 
 72 struct buffer_head;
 73 typedef int (get_block_t)(struct inode *inode, sector_t iblock,
 74                         struct buffer_head *bh_result, int create);
 75 typedef int (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
 76                         ssize_t bytes, void *private);
 77 typedef void (dax_iodone_t)(struct buffer_head *bh_map, int uptodate);
 78 
 79 #define MAY_EXEC                0x00000001
 80 #define MAY_WRITE               0x00000002
 81 #define MAY_READ                0x00000004
 82 #define MAY_APPEND              0x00000008
 83 #define MAY_ACCESS              0x00000010
 84 #define MAY_OPEN                0x00000020
 85 #define MAY_CHDIR               0x00000040
 86 /* called from RCU mode, don't block */
 87 #define MAY_NOT_BLOCK           0x00000080
 88 
 89 /*
 90  * flags in file.f_mode.  Note that FMODE_READ and FMODE_WRITE must correspond
 91  * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
 92  */
 93 
 94 /* file is open for reading */
 95 #define FMODE_READ              ((__force fmode_t)0x1)
 96 /* file is open for writing */
 97 #define FMODE_WRITE             ((__force fmode_t)0x2)
 98 /* file is seekable */
 99 #define FMODE_LSEEK             ((__force fmode_t)0x4)
100 /* file can be accessed using pread */
101 #define FMODE_PREAD             ((__force fmode_t)0x8)
102 /* file can be accessed using pwrite */
103 #define FMODE_PWRITE            ((__force fmode_t)0x10)
104 /* File is opened for execution with sys_execve / sys_uselib */
105 #define FMODE_EXEC              ((__force fmode_t)0x20)
106 /* File is opened with O_NDELAY (only set for block devices) */
107 #define FMODE_NDELAY            ((__force fmode_t)0x40)
108 /* File is opened with O_EXCL (only set for block devices) */
109 #define FMODE_EXCL              ((__force fmode_t)0x80)
110 /* File is opened using open(.., 3, ..) and is writeable only for ioctls
111    (specialy hack for floppy.c) */
112 #define FMODE_WRITE_IOCTL       ((__force fmode_t)0x100)
113 /* 32bit hashes as llseek() offset (for directories) */
114 #define FMODE_32BITHASH         ((__force fmode_t)0x200)
115 /* 64bit hashes as llseek() offset (for directories) */
116 #define FMODE_64BITHASH         ((__force fmode_t)0x400)
117 
118 /*
119  * Don't update ctime and mtime.
120  *
121  * Currently a special hack for the XFS open_by_handle ioctl, but we'll
122  * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
123  */
124 #define FMODE_NOCMTIME          ((__force fmode_t)0x800)
125 
126 /* Expect random access pattern */
127 #define FMODE_RANDOM            ((__force fmode_t)0x1000)
128 
129 /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
130 #define FMODE_UNSIGNED_OFFSET   ((__force fmode_t)0x2000)
131 
132 /* File is opened with O_PATH; almost nothing can be done with it */
133 #define FMODE_PATH              ((__force fmode_t)0x4000)
134 
135 /* File needs atomic accesses to f_pos */
136 #define FMODE_ATOMIC_POS        ((__force fmode_t)0x8000)
137 /* Write access to underlying fs */
138 #define FMODE_WRITER            ((__force fmode_t)0x10000)
139 /* Has read method(s) */
140 #define FMODE_CAN_READ          ((__force fmode_t)0x20000)
141 /* Has write method(s) */
142 #define FMODE_CAN_WRITE         ((__force fmode_t)0x40000)
143 
144 /* File was opened by fanotify and shouldn't generate fanotify events */
145 #define FMODE_NONOTIFY          ((__force fmode_t)0x4000000)
146 
147 /*
148  * Flag for rw_copy_check_uvector and compat_rw_copy_check_uvector
149  * that indicates that they should check the contents of the iovec are
150  * valid, but not check the memory that the iovec elements
151  * points too.
152  */
153 #define CHECK_IOVEC_ONLY -1
154 
155 /*
156  * The below are the various read and write types that we support. Some of
157  * them include behavioral modifiers that send information down to the
158  * block layer and IO scheduler. Terminology:
159  *
160  *      The block layer uses device plugging to defer IO a little bit, in
161  *      the hope that we will see more IO very shortly. This increases
162  *      coalescing of adjacent IO and thus reduces the number of IOs we
163  *      have to send to the device. It also allows for better queuing,
164  *      if the IO isn't mergeable. If the caller is going to be waiting
165  *      for the IO, then he must ensure that the device is unplugged so
166  *      that the IO is dispatched to the driver.
167  *
168  *      All IO is handled async in Linux. This is fine for background
169  *      writes, but for reads or writes that someone waits for completion
170  *      on, we want to notify the block layer and IO scheduler so that they
171  *      know about it. That allows them to make better scheduling
172  *      decisions. So when the below references 'sync' and 'async', it
173  *      is referencing this priority hint.
174  *
175  * With that in mind, the available types are:
176  *
177  * READ                 A normal read operation. Device will be plugged.
178  * READ_SYNC            A synchronous read. Device is not plugged, caller can
179  *                      immediately wait on this read without caring about
180  *                      unplugging.
181  * READA                Used for read-ahead operations. Lower priority, and the
182  *                      block layer could (in theory) choose to ignore this
183  *                      request if it runs into resource problems.
184  * WRITE                A normal async write. Device will be plugged.
185  * WRITE_SYNC           Synchronous write. Identical to WRITE, but passes down
186  *                      the hint that someone will be waiting on this IO
187  *                      shortly. The write equivalent of READ_SYNC.
188  * WRITE_ODIRECT        Special case write for O_DIRECT only.
189  * WRITE_FLUSH          Like WRITE_SYNC but with preceding cache flush.
190  * WRITE_FUA            Like WRITE_SYNC but data is guaranteed to be on
191  *                      non-volatile media on completion.
192  * WRITE_FLUSH_FUA      Combination of WRITE_FLUSH and FUA. The IO is preceded
193  *                      by a cache flush and data is guaranteed to be on
194  *                      non-volatile media on completion.
195  *
196  */
197 #define RW_MASK                 REQ_WRITE
198 #define RWA_MASK                REQ_RAHEAD
199 
200 #define READ                    0
201 #define WRITE                   RW_MASK
202 #define READA                   RWA_MASK
203 
204 #define READ_SYNC               (READ | REQ_SYNC)
205 #define WRITE_SYNC              (WRITE | REQ_SYNC | REQ_NOIDLE)
206 #define WRITE_ODIRECT           (WRITE | REQ_SYNC)
207 #define WRITE_FLUSH             (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
208 #define WRITE_FUA               (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
209 #define WRITE_FLUSH_FUA         (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
210 
211 /*
212  * Attribute flags.  These should be or-ed together to figure out what
213  * has been changed!
214  */
215 #define ATTR_MODE       (1 << 0)
216 #define ATTR_UID        (1 << 1)
217 #define ATTR_GID        (1 << 2)
218 #define ATTR_SIZE       (1 << 3)
219 #define ATTR_ATIME      (1 << 4)
220 #define ATTR_MTIME      (1 << 5)
221 #define ATTR_CTIME      (1 << 6)
222 #define ATTR_ATIME_SET  (1 << 7)
223 #define ATTR_MTIME_SET  (1 << 8)
224 #define ATTR_FORCE      (1 << 9) /* Not a change, but a change it */
225 #define ATTR_ATTR_FLAG  (1 << 10)
226 #define ATTR_KILL_SUID  (1 << 11)
227 #define ATTR_KILL_SGID  (1 << 12)
228 #define ATTR_FILE       (1 << 13)
229 #define ATTR_KILL_PRIV  (1 << 14)
230 #define ATTR_OPEN       (1 << 15) /* Truncating from open(O_TRUNC) */
231 #define ATTR_TIMES_SET  (1 << 16)
232 
233 /*
234  * Whiteout is represented by a char device.  The following constants define the
235  * mode and device number to use.
236  */
237 #define WHITEOUT_MODE 0
238 #define WHITEOUT_DEV 0
239 
240 /*
241  * This is the Inode Attributes structure, used for notify_change().  It
242  * uses the above definitions as flags, to know which values have changed.
243  * Also, in this manner, a Filesystem can look at only the values it cares
244  * about.  Basically, these are the attributes that the VFS layer can
245  * request to change from the FS layer.
246  *
247  * Derek Atkins <warlord@MIT.EDU> 94-10-20
248  */
249 struct iattr {
250         unsigned int    ia_valid;
251         umode_t         ia_mode;
252         kuid_t          ia_uid;
253         kgid_t          ia_gid;
254         loff_t          ia_size;
255         struct timespec ia_atime;
256         struct timespec ia_mtime;
257         struct timespec ia_ctime;
258 
259         /*
260          * Not an attribute, but an auxiliary info for filesystems wanting to
261          * implement an ftruncate() like method.  NOTE: filesystem should
262          * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
263          */
264         struct file     *ia_file;
265 };
266 
267 /*
268  * Includes for diskquotas.
269  */
270 #include <linux/quota.h>
271 
272 /*
273  * Maximum number of layers of fs stack.  Needs to be limited to
274  * prevent kernel stack overflow
275  */
276 #define FILESYSTEM_MAX_STACK_DEPTH 2
277 
278 /** 
279  * enum positive_aop_returns - aop return codes with specific semantics
280  *
281  * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
282  *                          completed, that the page is still locked, and
283  *                          should be considered active.  The VM uses this hint
284  *                          to return the page to the active list -- it won't
285  *                          be a candidate for writeback again in the near
286  *                          future.  Other callers must be careful to unlock
287  *                          the page if they get this return.  Returned by
288  *                          writepage(); 
289  *
290  * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
291  *                      unlocked it and the page might have been truncated.
292  *                      The caller should back up to acquiring a new page and
293  *                      trying again.  The aop will be taking reasonable
294  *                      precautions not to livelock.  If the caller held a page
295  *                      reference, it should drop it before retrying.  Returned
296  *                      by readpage().
297  *
298  * address_space_operation functions return these large constants to indicate
299  * special semantics to the caller.  These are much larger than the bytes in a
300  * page to allow for functions that return the number of bytes operated on in a
301  * given page.
302  */
303 
304 enum positive_aop_returns {
305         AOP_WRITEPAGE_ACTIVATE  = 0x80000,
306         AOP_TRUNCATED_PAGE      = 0x80001,
307 };
308 
309 #define AOP_FLAG_UNINTERRUPTIBLE        0x0001 /* will not do a short write */
310 #define AOP_FLAG_CONT_EXPAND            0x0002 /* called from cont_expand */
311 #define AOP_FLAG_NOFS                   0x0004 /* used by filesystem to direct
312                                                 * helper code (eg buffer layer)
313                                                 * to clear GFP_FS from alloc */
314 
315 /*
316  * oh the beauties of C type declarations.
317  */
318 struct page;
319 struct address_space;
320 struct writeback_control;
321 
322 #define IOCB_EVENTFD            (1 << 0)
323 #define IOCB_APPEND             (1 << 1)
324 #define IOCB_DIRECT             (1 << 2)
325 #define IOCB_HIPRI              (1 << 3)
326 
327 struct kiocb {
328         struct file             *ki_filp;
329         loff_t                  ki_pos;
330         void (*ki_complete)(struct kiocb *iocb, long ret, long ret2);
331         void                    *private;
332         int                     ki_flags;
333 };
334 
335 static inline bool is_sync_kiocb(struct kiocb *kiocb)
336 {
337         return kiocb->ki_complete == NULL;
338 }
339 
340 static inline int iocb_flags(struct file *file);
341 
342 static inline void init_sync_kiocb(struct kiocb *kiocb, struct file *filp)
343 {
344         *kiocb = (struct kiocb) {
345                 .ki_filp = filp,
346                 .ki_flags = iocb_flags(filp),
347         };
348 }
349 
350 /*
351  * "descriptor" for what we're up to with a read.
352  * This allows us to use the same read code yet
353  * have multiple different users of the data that
354  * we read from a file.
355  *
356  * The simplest case just copies the data to user
357  * mode.
358  */
359 typedef struct {
360         size_t written;
361         size_t count;
362         union {
363                 char __user *buf;
364                 void *data;
365         } arg;
366         int error;
367 } read_descriptor_t;
368 
369 typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
370                 unsigned long, unsigned long);
371 
372 struct address_space_operations {
373         int (*writepage)(struct page *page, struct writeback_control *wbc);
374         int (*readpage)(struct file *, struct page *);
375 
376         /* Write back some dirty pages from this mapping. */
377         int (*writepages)(struct address_space *, struct writeback_control *);
378 
379         /* Set a page dirty.  Return true if this dirtied it */
380         int (*set_page_dirty)(struct page *page);
381 
382         int (*readpages)(struct file *filp, struct address_space *mapping,
383                         struct list_head *pages, unsigned nr_pages);
384 
385         int (*write_begin)(struct file *, struct address_space *mapping,
386                                 loff_t pos, unsigned len, unsigned flags,
387                                 struct page **pagep, void **fsdata);
388         int (*write_end)(struct file *, struct address_space *mapping,
389                                 loff_t pos, unsigned len, unsigned copied,
390                                 struct page *page, void *fsdata);
391 
392         /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
393         sector_t (*bmap)(struct address_space *, sector_t);
394         void (*invalidatepage) (struct page *, unsigned int, unsigned int);
395         int (*releasepage) (struct page *, gfp_t);
396         void (*freepage)(struct page *);
397         ssize_t (*direct_IO)(struct kiocb *, struct iov_iter *iter, loff_t offset);
398         /*
399          * migrate the contents of a page to the specified target. If
400          * migrate_mode is MIGRATE_ASYNC, it must not block.
401          */
402         int (*migratepage) (struct address_space *,
403                         struct page *, struct page *, enum migrate_mode);
404         int (*launder_page) (struct page *);
405         int (*is_partially_uptodate) (struct page *, unsigned long,
406                                         unsigned long);
407         void (*is_dirty_writeback) (struct page *, bool *, bool *);
408         int (*error_remove_page)(struct address_space *, struct page *);
409 
410         /* swapfile support */
411         int (*swap_activate)(struct swap_info_struct *sis, struct file *file,
412                                 sector_t *span);
413         void (*swap_deactivate)(struct file *file);
414 };
415 
416 extern const struct address_space_operations empty_aops;
417 
418 /*
419  * pagecache_write_begin/pagecache_write_end must be used by general code
420  * to write into the pagecache.
421  */
422 int pagecache_write_begin(struct file *, struct address_space *mapping,
423                                 loff_t pos, unsigned len, unsigned flags,
424                                 struct page **pagep, void **fsdata);
425 
426 int pagecache_write_end(struct file *, struct address_space *mapping,
427                                 loff_t pos, unsigned len, unsigned copied,
428                                 struct page *page, void *fsdata);
429 
430 struct address_space {
431         struct inode            *host;          /* owner: inode, block_device */
432         struct radix_tree_root  page_tree;      /* radix tree of all pages */
433         spinlock_t              tree_lock;      /* and lock protecting it */
434         atomic_t                i_mmap_writable;/* count VM_SHARED mappings */
435         struct rb_root          i_mmap;         /* tree of private and shared mappings */
436         struct rw_semaphore     i_mmap_rwsem;   /* protect tree, count, list */
437         /* Protected by tree_lock together with the radix tree */
438         unsigned long           nrpages;        /* number of total pages */
439         /* number of shadow or DAX exceptional entries */
440         unsigned long           nrexceptional;
441         pgoff_t                 writeback_index;/* writeback starts here */
442         const struct address_space_operations *a_ops;   /* methods */
443         unsigned long           flags;          /* error bits/gfp mask */
444         spinlock_t              private_lock;   /* for use by the address_space */
445         struct list_head        private_list;   /* ditto */
446         void                    *private_data;  /* ditto */
447 } __attribute__((aligned(sizeof(long))));
448         /*
449          * On most architectures that alignment is already the case; but
450          * must be enforced here for CRIS, to let the least significant bit
451          * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
452          */
453 struct request_queue;
454 
455 struct block_device {
456         dev_t                   bd_dev;  /* not a kdev_t - it's a search key */
457         int                     bd_openers;
458         struct inode *          bd_inode;       /* will die */
459         struct super_block *    bd_super;
460         struct mutex            bd_mutex;       /* open/close mutex */
461         struct list_head        bd_inodes;
462         void *                  bd_claiming;
463         void *                  bd_holder;
464         int                     bd_holders;
465         bool                    bd_write_holder;
466 #ifdef CONFIG_SYSFS
467         struct list_head        bd_holder_disks;
468 #endif
469         struct block_device *   bd_contains;
470         unsigned                bd_block_size;
471         struct hd_struct *      bd_part;
472         /* number of times partitions within this device have been opened. */
473         unsigned                bd_part_count;
474         int                     bd_invalidated;
475         struct gendisk *        bd_disk;
476         struct request_queue *  bd_queue;
477         struct list_head        bd_list;
478         /*
479          * Private data.  You must have bd_claim'ed the block_device
480          * to use this.  NOTE:  bd_claim allows an owner to claim
481          * the same device multiple times, the owner must take special
482          * care to not mess up bd_private for that case.
483          */
484         unsigned long           bd_private;
485 
486         /* The counter of freeze processes */
487         int                     bd_fsfreeze_count;
488         /* Mutex for freeze */
489         struct mutex            bd_fsfreeze_mutex;
490 };
491 
492 /*
493  * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
494  * radix trees
495  */
496 #define PAGECACHE_TAG_DIRTY     0
497 #define PAGECACHE_TAG_WRITEBACK 1
498 #define PAGECACHE_TAG_TOWRITE   2
499 
500 int mapping_tagged(struct address_space *mapping, int tag);
501 
502 static inline void i_mmap_lock_write(struct address_space *mapping)
503 {
504         down_write(&mapping->i_mmap_rwsem);
505 }
506 
507 static inline void i_mmap_unlock_write(struct address_space *mapping)
508 {
509         up_write(&mapping->i_mmap_rwsem);
510 }
511 
512 static inline void i_mmap_lock_read(struct address_space *mapping)
513 {
514         down_read(&mapping->i_mmap_rwsem);
515 }
516 
517 static inline void i_mmap_unlock_read(struct address_space *mapping)
518 {
519         up_read(&mapping->i_mmap_rwsem);
520 }
521 
522 /*
523  * Might pages of this file be mapped into userspace?
524  */
525 static inline int mapping_mapped(struct address_space *mapping)
526 {
527         return  !RB_EMPTY_ROOT(&mapping->i_mmap);
528 }
529 
530 /*
531  * Might pages of this file have been modified in userspace?
532  * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
533  * marks vma as VM_SHARED if it is shared, and the file was opened for
534  * writing i.e. vma may be mprotected writable even if now readonly.
535  *
536  * If i_mmap_writable is negative, no new writable mappings are allowed. You
537  * can only deny writable mappings, if none exists right now.
538  */
539 static inline int mapping_writably_mapped(struct address_space *mapping)
540 {
541         return atomic_read(&mapping->i_mmap_writable) > 0;
542 }
543 
544 static inline int mapping_map_writable(struct address_space *mapping)
545 {
546         return atomic_inc_unless_negative(&mapping->i_mmap_writable) ?
547                 0 : -EPERM;
548 }
549 
550 static inline void mapping_unmap_writable(struct address_space *mapping)
551 {
552         atomic_dec(&mapping->i_mmap_writable);
553 }
554 
555 static inline int mapping_deny_writable(struct address_space *mapping)
556 {
557         return atomic_dec_unless_positive(&mapping->i_mmap_writable) ?
558                 0 : -EBUSY;
559 }
560 
561 static inline void mapping_allow_writable(struct address_space *mapping)
562 {
563         atomic_inc(&mapping->i_mmap_writable);
564 }
565 
566 /*
567  * Use sequence counter to get consistent i_size on 32-bit processors.
568  */
569 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
570 #include <linux/seqlock.h>
571 #define __NEED_I_SIZE_ORDERED
572 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
573 #else
574 #define i_size_ordered_init(inode) do { } while (0)
575 #endif
576 
577 struct posix_acl;
578 #define ACL_NOT_CACHED ((void *)(-1))
579 
580 #define IOP_FASTPERM    0x0001
581 #define IOP_LOOKUP      0x0002
582 #define IOP_NOFOLLOW    0x0004
583 
584 /*
585  * Keep mostly read-only and often accessed (especially for
586  * the RCU path lookup and 'stat' data) fields at the beginning
587  * of the 'struct inode'
588  */
589 struct inode {
590         umode_t                 i_mode;
591         unsigned short          i_opflags;
592         kuid_t                  i_uid;
593         kgid_t                  i_gid;
594         unsigned int            i_flags;
595 
596 #ifdef CONFIG_FS_POSIX_ACL
597         struct posix_acl        *i_acl;
598         struct posix_acl        *i_default_acl;
599 #endif
600 
601         const struct inode_operations   *i_op;
602         struct super_block      *i_sb;
603         struct address_space    *i_mapping;
604 
605 #ifdef CONFIG_SECURITY
606         void                    *i_security;
607 #endif
608 
609         /* Stat data, not accessed from path walking */
610         unsigned long           i_ino;
611         /*
612          * Filesystems may only read i_nlink directly.  They shall use the
613          * following functions for modification:
614          *
615          *    (set|clear|inc|drop)_nlink
616          *    inode_(inc|dec)_link_count
617          */
618         union {
619                 const unsigned int i_nlink;
620                 unsigned int __i_nlink;
621         };
622         dev_t                   i_rdev;
623         loff_t                  i_size;
624         struct timespec         i_atime;
625         struct timespec         i_mtime;
626         struct timespec         i_ctime;
627         spinlock_t              i_lock; /* i_blocks, i_bytes, maybe i_size */
628         unsigned short          i_bytes;
629         unsigned int            i_blkbits;
630         blkcnt_t                i_blocks;
631 
632 #ifdef __NEED_I_SIZE_ORDERED
633         seqcount_t              i_size_seqcount;
634 #endif
635 
636         /* Misc */
637         unsigned long           i_state;
638         struct mutex            i_mutex;
639 
640         unsigned long           dirtied_when;   /* jiffies of first dirtying */
641         unsigned long           dirtied_time_when;
642 
643         struct hlist_node       i_hash;
644         struct list_head        i_io_list;      /* backing dev IO list */
645 #ifdef CONFIG_CGROUP_WRITEBACK
646         struct bdi_writeback    *i_wb;          /* the associated cgroup wb */
647 
648         /* foreign inode detection, see wbc_detach_inode() */
649         int                     i_wb_frn_winner;
650         u16                     i_wb_frn_avg_time;
651         u16                     i_wb_frn_history;
652 #endif
653         struct list_head        i_lru;          /* inode LRU list */
654         struct list_head        i_sb_list;
655         union {
656                 struct hlist_head       i_dentry;
657                 struct rcu_head         i_rcu;
658         };
659         u64                     i_version;
660         atomic_t                i_count;
661         atomic_t                i_dio_count;
662         atomic_t                i_writecount;
663 #ifdef CONFIG_IMA
664         atomic_t                i_readcount; /* struct files open RO */
665 #endif
666         const struct file_operations    *i_fop; /* former ->i_op->default_file_ops */
667         struct file_lock_context        *i_flctx;
668         struct address_space    i_data;
669         struct list_head        i_devices;
670         union {
671                 struct pipe_inode_info  *i_pipe;
672                 struct block_device     *i_bdev;
673                 struct cdev             *i_cdev;
674                 char                    *i_link;
675         };
676 
677         __u32                   i_generation;
678 
679 #ifdef CONFIG_FSNOTIFY
680         __u32                   i_fsnotify_mask; /* all events this inode cares about */
681         struct hlist_head       i_fsnotify_marks;
682 #endif
683 
684 #if IS_ENABLED(CONFIG_FS_ENCRYPTION)
685         struct fscrypt_info     *i_crypt_info;
686 #endif
687 
688         void                    *i_private; /* fs or device private pointer */
689 };
690 
691 static inline int inode_unhashed(struct inode *inode)
692 {
693         return hlist_unhashed(&inode->i_hash);
694 }
695 
696 /*
697  * inode->i_mutex nesting subclasses for the lock validator:
698  *
699  * 0: the object of the current VFS operation
700  * 1: parent
701  * 2: child/target
702  * 3: xattr
703  * 4: second non-directory
704  * 5: second parent (when locking independent directories in rename)
705  *
706  * I_MUTEX_NONDIR2 is for certain operations (such as rename) which lock two
707  * non-directories at once.
708  *
709  * The locking order between these classes is
710  * parent[2] -> child -> grandchild -> normal -> xattr -> second non-directory
711  */
712 enum inode_i_mutex_lock_class
713 {
714         I_MUTEX_NORMAL,
715         I_MUTEX_PARENT,
716         I_MUTEX_CHILD,
717         I_MUTEX_XATTR,
718         I_MUTEX_NONDIR2,
719         I_MUTEX_PARENT2,
720 };
721 
722 static inline void inode_lock(struct inode *inode)
723 {
724         mutex_lock(&inode->i_mutex);
725 }
726 
727 static inline void inode_unlock(struct inode *inode)
728 {
729         mutex_unlock(&inode->i_mutex);
730 }
731 
732 static inline int inode_trylock(struct inode *inode)
733 {
734         return mutex_trylock(&inode->i_mutex);
735 }
736 
737 static inline int inode_is_locked(struct inode *inode)
738 {
739         return mutex_is_locked(&inode->i_mutex);
740 }
741 
742 static inline void inode_lock_nested(struct inode *inode, unsigned subclass)
743 {
744         mutex_lock_nested(&inode->i_mutex, subclass);
745 }
746 
747 void lock_two_nondirectories(struct inode *, struct inode*);
748 void unlock_two_nondirectories(struct inode *, struct inode*);
749 
750 /*
751  * NOTE: in a 32bit arch with a preemptable kernel and
752  * an UP compile the i_size_read/write must be atomic
753  * with respect to the local cpu (unlike with preempt disabled),
754  * but they don't need to be atomic with respect to other cpus like in
755  * true SMP (so they need either to either locally disable irq around
756  * the read or for example on x86 they can be still implemented as a
757  * cmpxchg8b without the need of the lock prefix). For SMP compiles
758  * and 64bit archs it makes no difference if preempt is enabled or not.
759  */
760 static inline loff_t i_size_read(const struct inode *inode)
761 {
762 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
763         loff_t i_size;
764         unsigned int seq;
765 
766         do {
767                 seq = read_seqcount_begin(&inode->i_size_seqcount);
768                 i_size = inode->i_size;
769         } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
770         return i_size;
771 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
772         loff_t i_size;
773 
774         preempt_disable();
775         i_size = inode->i_size;
776         preempt_enable();
777         return i_size;
778 #else
779         return inode->i_size;
780 #endif
781 }
782 
783 /*
784  * NOTE: unlike i_size_read(), i_size_write() does need locking around it
785  * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
786  * can be lost, resulting in subsequent i_size_read() calls spinning forever.
787  */
788 static inline void i_size_write(struct inode *inode, loff_t i_size)
789 {
790 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
791         preempt_disable();
792         write_seqcount_begin(&inode->i_size_seqcount);
793         inode->i_size = i_size;
794         write_seqcount_end(&inode->i_size_seqcount);
795         preempt_enable();
796 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
797         preempt_disable();
798         inode->i_size = i_size;
799         preempt_enable();
800 #else
801         inode->i_size = i_size;
802 #endif
803 }
804 
805 /* Helper functions so that in most cases filesystems will
806  * not need to deal directly with kuid_t and kgid_t and can
807  * instead deal with the raw numeric values that are stored
808  * in the filesystem.
809  */
810 static inline uid_t i_uid_read(const struct inode *inode)
811 {
812         return from_kuid(&init_user_ns, inode->i_uid);
813 }
814 
815 static inline gid_t i_gid_read(const struct inode *inode)
816 {
817         return from_kgid(&init_user_ns, inode->i_gid);
818 }
819 
820 static inline void i_uid_write(struct inode *inode, uid_t uid)
821 {
822         inode->i_uid = make_kuid(&init_user_ns, uid);
823 }
824 
825 static inline void i_gid_write(struct inode *inode, gid_t gid)
826 {
827         inode->i_gid = make_kgid(&init_user_ns, gid);
828 }
829 
830 static inline unsigned iminor(const struct inode *inode)
831 {
832         return MINOR(inode->i_rdev);
833 }
834 
835 static inline unsigned imajor(const struct inode *inode)
836 {
837         return MAJOR(inode->i_rdev);
838 }
839 
840 extern struct block_device *I_BDEV(struct inode *inode);
841 
842 struct fown_struct {
843         rwlock_t lock;          /* protects pid, uid, euid fields */
844         struct pid *pid;        /* pid or -pgrp where SIGIO should be sent */
845         enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
846         kuid_t uid, euid;       /* uid/euid of process setting the owner */
847         int signum;             /* posix.1b rt signal to be delivered on IO */
848 };
849 
850 /*
851  * Track a single file's readahead state
852  */
853 struct file_ra_state {
854         pgoff_t start;                  /* where readahead started */
855         unsigned int size;              /* # of readahead pages */
856         unsigned int async_size;        /* do asynchronous readahead when
857                                            there are only # of pages ahead */
858 
859         unsigned int ra_pages;          /* Maximum readahead window */
860         unsigned int mmap_miss;         /* Cache miss stat for mmap accesses */
861         loff_t prev_pos;                /* Cache last read() position */
862 };
863 
864 /*
865  * Check if @index falls in the readahead windows.
866  */
867 static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
868 {
869         return (index >= ra->start &&
870                 index <  ra->start + ra->size);
871 }
872 
873 struct file {
874         union {
875                 struct llist_node       fu_llist;
876                 struct rcu_head         fu_rcuhead;
877         } f_u;
878         struct path             f_path;
879         struct inode            *f_inode;       /* cached value */
880         const struct file_operations    *f_op;
881 
882         /*
883          * Protects f_ep_links, f_flags.
884          * Must not be taken from IRQ context.
885          */
886         spinlock_t              f_lock;
887         atomic_long_t           f_count;
888         unsigned int            f_flags;
889         fmode_t                 f_mode;
890         struct mutex            f_pos_lock;
891         loff_t                  f_pos;
892         struct fown_struct      f_owner;
893         const struct cred       *f_cred;
894         struct file_ra_state    f_ra;
895 
896         u64                     f_version;
897 #ifdef CONFIG_SECURITY
898         void                    *f_security;
899 #endif
900         /* needed for tty driver, and maybe others */
901         void                    *private_data;
902 
903 #ifdef CONFIG_EPOLL
904         /* Used by fs/eventpoll.c to link all the hooks to this file */
905         struct list_head        f_ep_links;
906         struct list_head        f_tfile_llink;
907 #endif /* #ifdef CONFIG_EPOLL */
908         struct address_space    *f_mapping;
909 } __attribute__((aligned(4)));  /* lest something weird decides that 2 is OK */
910 
911 struct file_handle {
912         __u32 handle_bytes;
913         int handle_type;
914         /* file identifier */
915         unsigned char f_handle[0];
916 };
917 
918 static inline struct file *get_file(struct file *f)
919 {
920         atomic_long_inc(&f->f_count);
921         return f;
922 }
923 #define get_file_rcu(x) atomic_long_inc_not_zero(&(x)->f_count)
924 #define fput_atomic(x)  atomic_long_add_unless(&(x)->f_count, -1, 1)
925 #define file_count(x)   atomic_long_read(&(x)->f_count)
926 
927 #define MAX_NON_LFS     ((1UL<<31) - 1)
928 
929 /* Page cache limit. The filesystems should put that into their s_maxbytes 
930    limits, otherwise bad things can happen in VM. */ 
931 #if BITS_PER_LONG==32
932 #define MAX_LFS_FILESIZE        (((loff_t)PAGE_SIZE << (BITS_PER_LONG-1))-1)
933 #elif BITS_PER_LONG==64
934 #define MAX_LFS_FILESIZE        ((loff_t)0x7fffffffffffffffLL)
935 #endif
936 
937 #define FL_POSIX        1
938 #define FL_FLOCK        2
939 #define FL_DELEG        4       /* NFSv4 delegation */
940 #define FL_ACCESS       8       /* not trying to lock, just looking */
941 #define FL_EXISTS       16      /* when unlocking, test for existence */
942 #define FL_LEASE        32      /* lease held on this file */
943 #define FL_CLOSE        64      /* unlock on close */
944 #define FL_SLEEP        128     /* A blocking lock */
945 #define FL_DOWNGRADE_PENDING    256 /* Lease is being downgraded */
946 #define FL_UNLOCK_PENDING       512 /* Lease is being broken */
947 #define FL_OFDLCK       1024    /* lock is "owned" by struct file */
948 #define FL_LAYOUT       2048    /* outstanding pNFS layout */
949 
950 /*
951  * Special return value from posix_lock_file() and vfs_lock_file() for
952  * asynchronous locking.
953  */
954 #define FILE_LOCK_DEFERRED 1
955 
956 /* legacy typedef, should eventually be removed */
957 typedef void *fl_owner_t;
958 
959 struct file_lock;
960 
961 struct file_lock_operations {
962         void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
963         void (*fl_release_private)(struct file_lock *);
964 };
965 
966 struct lock_manager_operations {
967         int (*lm_compare_owner)(struct file_lock *, struct file_lock *);
968         unsigned long (*lm_owner_key)(struct file_lock *);
969         fl_owner_t (*lm_get_owner)(fl_owner_t);
970         void (*lm_put_owner)(fl_owner_t);
971         void (*lm_notify)(struct file_lock *);  /* unblock callback */
972         int (*lm_grant)(struct file_lock *, int);
973         bool (*lm_break)(struct file_lock *);
974         int (*lm_change)(struct file_lock *, int, struct list_head *);
975         void (*lm_setup)(struct file_lock *, void **);
976 };
977 
978 struct lock_manager {
979         struct list_head list;
980         /*
981          * NFSv4 and up also want opens blocked during the grace period;
982          * NLM doesn't care:
983          */
984         bool block_opens;
985 };
986 
987 struct net;
988 void locks_start_grace(struct net *, struct lock_manager *);
989 void locks_end_grace(struct lock_manager *);
990 int locks_in_grace(struct net *);
991 int opens_in_grace(struct net *);
992 
993 /* that will die - we need it for nfs_lock_info */
994 #include <linux/nfs_fs_i.h>
995 
996 /*
997  * struct file_lock represents a generic "file lock". It's used to represent
998  * POSIX byte range locks, BSD (flock) locks, and leases. It's important to
999  * note that the same struct is used to represent both a request for a lock and
1000  * the lock itself, but the same object is never used for both.
1001  *
1002  * FIXME: should we create a separate "struct lock_request" to help distinguish
1003  * these two uses?
1004  *
1005  * The varous i_flctx lists are ordered by:
1006  *
1007  * 1) lock owner
1008  * 2) lock range start
1009  * 3) lock range end
1010  *
1011  * Obviously, the last two criteria only matter for POSIX locks.
1012  */
1013 struct file_lock {
1014         struct file_lock *fl_next;      /* singly linked list for this inode  */
1015         struct list_head fl_list;       /* link into file_lock_context */
1016         struct hlist_node fl_link;      /* node in global lists */
1017         struct list_head fl_block;      /* circular list of blocked processes */
1018         fl_owner_t fl_owner;
1019         unsigned int fl_flags;
1020         unsigned char fl_type;
1021         unsigned int fl_pid;
1022         int fl_link_cpu;                /* what cpu's list is this on? */
1023         struct pid *fl_nspid;
1024         wait_queue_head_t fl_wait;
1025         struct file *fl_file;
1026         loff_t fl_start;
1027         loff_t fl_end;
1028 
1029         struct fasync_struct *  fl_fasync; /* for lease break notifications */
1030         /* for lease breaks: */
1031         unsigned long fl_break_time;
1032         unsigned long fl_downgrade_time;
1033 
1034         const struct file_lock_operations *fl_ops;      /* Callbacks for filesystems */
1035         const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
1036         union {
1037                 struct nfs_lock_info    nfs_fl;
1038                 struct nfs4_lock_info   nfs4_fl;
1039                 struct {
1040                         struct list_head link;  /* link in AFS vnode's pending_locks list */
1041                         int state;              /* state of grant or error if -ve */
1042                 } afs;
1043         } fl_u;
1044 };
1045 
1046 struct file_lock_context {
1047         spinlock_t              flc_lock;
1048         struct list_head        flc_flock;
1049         struct list_head        flc_posix;
1050         struct list_head        flc_lease;
1051 };
1052 
1053 /* The following constant reflects the upper bound of the file/locking space */
1054 #ifndef OFFSET_MAX
1055 #define INT_LIMIT(x)    (~((x)1 << (sizeof(x)*8 - 1)))
1056 #define OFFSET_MAX      INT_LIMIT(loff_t)
1057 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
1058 #endif
1059 
1060 #include <linux/fcntl.h>
1061 
1062 extern void send_sigio(struct fown_struct *fown, int fd, int band);
1063 
1064 #ifdef CONFIG_FILE_LOCKING
1065 extern int fcntl_getlk(struct file *, unsigned int, struct flock __user *);
1066 extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
1067                         struct flock __user *);
1068 
1069 #if BITS_PER_LONG == 32
1070 extern int fcntl_getlk64(struct file *, unsigned int, struct flock64 __user *);
1071 extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
1072                         struct flock64 __user *);
1073 #endif
1074 
1075 extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
1076 extern int fcntl_getlease(struct file *filp);
1077 
1078 /* fs/locks.c */
1079 void locks_free_lock_context(struct inode *inode);
1080 void locks_free_lock(struct file_lock *fl);
1081 extern void locks_init_lock(struct file_lock *);
1082 extern struct file_lock * locks_alloc_lock(void);
1083 extern void locks_copy_lock(struct file_lock *, struct file_lock *);
1084 extern void locks_copy_conflock(struct file_lock *, struct file_lock *);
1085 extern void locks_remove_posix(struct file *, fl_owner_t);
1086 extern void locks_remove_file(struct file *);
1087 extern void locks_release_private(struct file_lock *);
1088 extern void posix_test_lock(struct file *, struct file_lock *);
1089 extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
1090 extern int posix_unblock_lock(struct file_lock *);
1091 extern int vfs_test_lock(struct file *, struct file_lock *);
1092 extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
1093 extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
1094 extern int locks_lock_inode_wait(struct inode *inode, struct file_lock *fl);
1095 extern int __break_lease(struct inode *inode, unsigned int flags, unsigned int type);
1096 extern void lease_get_mtime(struct inode *, struct timespec *time);
1097 extern int generic_setlease(struct file *, long, struct file_lock **, void **priv);
1098 extern int vfs_setlease(struct file *, long, struct file_lock **, void **);
1099 extern int lease_modify(struct file_lock *, int, struct list_head *);
1100 struct files_struct;
1101 extern void show_fd_locks(struct seq_file *f,
1102                          struct file *filp, struct files_struct *files);
1103 #else /* !CONFIG_FILE_LOCKING */
1104 static inline int fcntl_getlk(struct file *file, unsigned int cmd,
1105                               struct flock __user *user)
1106 {
1107         return -EINVAL;
1108 }
1109 
1110 static inline int fcntl_setlk(unsigned int fd, struct file *file,
1111                               unsigned int cmd, struct flock __user *user)
1112 {
1113         return -EACCES;
1114 }
1115 
1116 #if BITS_PER_LONG == 32
1117 static inline int fcntl_getlk64(struct file *file, unsigned int cmd,
1118                                 struct flock64 __user *user)
1119 {
1120         return -EINVAL;
1121 }
1122 
1123 static inline int fcntl_setlk64(unsigned int fd, struct file *file,
1124                                 unsigned int cmd, struct flock64 __user *user)
1125 {
1126         return -EACCES;
1127 }
1128 #endif
1129 static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
1130 {
1131         return -EINVAL;
1132 }
1133 
1134 static inline int fcntl_getlease(struct file *filp)
1135 {
1136         return F_UNLCK;
1137 }
1138 
1139 static inline void
1140 locks_free_lock_context(struct inode *inode)
1141 {
1142 }
1143 
1144 static inline void locks_init_lock(struct file_lock *fl)
1145 {
1146         return;
1147 }
1148 
1149 static inline void locks_copy_conflock(struct file_lock *new, struct file_lock *fl)
1150 {
1151         return;
1152 }
1153 
1154 static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1155 {
1156         return;
1157 }
1158 
1159 static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
1160 {
1161         return;
1162 }
1163 
1164 static inline void locks_remove_file(struct file *filp)
1165 {
1166         return;
1167 }
1168 
1169 static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
1170 {
1171         return;
1172 }
1173 
1174 static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
1175                                   struct file_lock *conflock)
1176 {
1177         return -ENOLCK;
1178 }
1179 
1180 static inline int posix_unblock_lock(struct file_lock *waiter)
1181 {
1182         return -ENOENT;
1183 }
1184 
1185 static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
1186 {
1187         return 0;
1188 }
1189 
1190 static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
1191                                 struct file_lock *fl, struct file_lock *conf)
1192 {
1193         return -ENOLCK;
1194 }
1195 
1196 static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
1197 {
1198         return 0;
1199 }
1200 
1201 static inline int locks_lock_inode_wait(struct inode *inode, struct file_lock *fl)
1202 {
1203         return -ENOLCK;
1204 }
1205 
1206 static inline int __break_lease(struct inode *inode, unsigned int mode, unsigned int type)
1207 {
1208         return 0;
1209 }
1210 
1211 static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
1212 {
1213         return;
1214 }
1215 
1216 static inline int generic_setlease(struct file *filp, long arg,
1217                                     struct file_lock **flp, void **priv)
1218 {
1219         return -EINVAL;
1220 }
1221 
1222 static inline int vfs_setlease(struct file *filp, long arg,
1223                                struct file_lock **lease, void **priv)
1224 {
1225         return -EINVAL;
1226 }
1227 
1228 static inline int lease_modify(struct file_lock *fl, int arg,
1229                                struct list_head *dispose)
1230 {
1231         return -EINVAL;
1232 }
1233 
1234 struct files_struct;
1235 static inline void show_fd_locks(struct seq_file *f,
1236                         struct file *filp, struct files_struct *files) {}
1237 #endif /* !CONFIG_FILE_LOCKING */
1238 
1239 static inline struct inode *file_inode(const struct file *f)
1240 {
1241         return f->f_inode;
1242 }
1243 
1244 static inline struct dentry *file_dentry(const struct file *file)
1245 {
1246         struct dentry *dentry = file->f_path.dentry;
1247 
1248         if (unlikely(dentry->d_flags & DCACHE_OP_REAL))
1249                 return dentry->d_op->d_real(dentry, file_inode(file));
1250         else
1251                 return dentry;
1252 }
1253 
1254 static inline int locks_lock_file_wait(struct file *filp, struct file_lock *fl)
1255 {
1256         return locks_lock_inode_wait(file_inode(filp), fl);
1257 }
1258 
1259 struct fasync_struct {
1260         spinlock_t              fa_lock;
1261         int                     magic;
1262         int                     fa_fd;
1263         struct fasync_struct    *fa_next; /* singly linked list */
1264         struct file             *fa_file;
1265         struct rcu_head         fa_rcu;
1266 };
1267 
1268 #define FASYNC_MAGIC 0x4601
1269 
1270 /* SMP safe fasync helpers: */
1271 extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
1272 extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
1273 extern int fasync_remove_entry(struct file *, struct fasync_struct **);
1274 extern struct fasync_struct *fasync_alloc(void);
1275 extern void fasync_free(struct fasync_struct *);
1276 
1277 /* can be called from interrupts */
1278 extern void kill_fasync(struct fasync_struct **, int, int);
1279 
1280 extern void __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
1281 extern void f_setown(struct file *filp, unsigned long arg, int force);
1282 extern void f_delown(struct file *filp);
1283 extern pid_t f_getown(struct file *filp);
1284 extern int send_sigurg(struct fown_struct *fown);
1285 
1286 struct mm_struct;
1287 
1288 /*
1289  *      Umount options
1290  */
1291 
1292 #define MNT_FORCE       0x00000001      /* Attempt to forcibily umount */
1293 #define MNT_DETACH      0x00000002      /* Just detach from the tree */
1294 #define MNT_EXPIRE      0x00000004      /* Mark for expiry */
1295 #define UMOUNT_NOFOLLOW 0x00000008      /* Don't follow symlink on umount */
1296 #define UMOUNT_UNUSED   0x80000000      /* Flag guaranteed to be unused */
1297 
1298 /* sb->s_iflags */
1299 #define SB_I_CGROUPWB   0x00000001      /* cgroup-aware writeback enabled */
1300 #define SB_I_NOEXEC     0x00000002      /* Ignore executables on this fs */
1301 
1302 /* Possible states of 'frozen' field */
1303 enum {
1304         SB_UNFROZEN = 0,                /* FS is unfrozen */
1305         SB_FREEZE_WRITE = 1,            /* Writes, dir ops, ioctls frozen */
1306         SB_FREEZE_PAGEFAULT = 2,        /* Page faults stopped as well */
1307         SB_FREEZE_FS = 3,               /* For internal FS use (e.g. to stop
1308                                          * internal threads if needed) */
1309         SB_FREEZE_COMPLETE = 4,         /* ->freeze_fs finished successfully */
1310 };
1311 
1312 #define SB_FREEZE_LEVELS (SB_FREEZE_COMPLETE - 1)
1313 
1314 struct sb_writers {
1315         int                             frozen;         /* Is sb frozen? */
1316         wait_queue_head_t               wait_unfrozen;  /* for get_super_thawed() */
1317         struct percpu_rw_semaphore      rw_sem[SB_FREEZE_LEVELS];
1318 };
1319 
1320 struct super_block {
1321         struct list_head        s_list;         /* Keep this first */
1322         dev_t                   s_dev;          /* search index; _not_ kdev_t */
1323         unsigned char           s_blocksize_bits;
1324         unsigned long           s_blocksize;
1325         loff_t                  s_maxbytes;     /* Max file size */
1326         struct file_system_type *s_type;
1327         const struct super_operations   *s_op;
1328         const struct dquot_operations   *dq_op;
1329         const struct quotactl_ops       *s_qcop;
1330         const struct export_operations *s_export_op;
1331         unsigned long           s_flags;
1332         unsigned long           s_iflags;       /* internal SB_I_* flags */
1333         unsigned long           s_magic;
1334         struct dentry           *s_root;
1335         struct rw_semaphore     s_umount;
1336         int                     s_count;
1337         atomic_t                s_active;
1338 #ifdef CONFIG_SECURITY
1339         void                    *s_security;
1340 #endif
1341         const struct xattr_handler **s_xattr;
1342 
1343         const struct fscrypt_operations *s_cop;
1344 
1345         struct hlist_bl_head    s_anon;         /* anonymous dentries for (nfs) exporting */
1346         struct list_head        s_mounts;       /* list of mounts; _not_ for fs use */
1347         struct block_device     *s_bdev;
1348         struct backing_dev_info *s_bdi;
1349         struct mtd_info         *s_mtd;
1350         struct hlist_node       s_instances;
1351         unsigned int            s_quota_types;  /* Bitmask of supported quota types */
1352         struct quota_info       s_dquot;        /* Diskquota specific options */
1353 
1354         struct sb_writers       s_writers;
1355 
1356         char s_id[32];                          /* Informational name */
1357         u8 s_uuid[16];                          /* UUID */
1358 
1359         void                    *s_fs_info;     /* Filesystem private info */
1360         unsigned int            s_max_links;
1361         fmode_t                 s_mode;
1362 
1363         /* Granularity of c/m/atime in ns.
1364            Cannot be worse than a second */
1365         u32                s_time_gran;
1366 
1367         /*
1368          * The next field is for VFS *only*. No filesystems have any business
1369          * even looking at it. You had been warned.
1370          */
1371         struct mutex s_vfs_rename_mutex;        /* Kludge */
1372 
1373         /*
1374          * Filesystem subtype.  If non-empty the filesystem type field
1375          * in /proc/mounts will be "type.subtype"
1376          */
1377         char *s_subtype;
1378 
1379         /*
1380          * Saved mount options for lazy filesystems using
1381          * generic_show_options()
1382          */
1383         char __rcu *s_options;
1384         const struct dentry_operations *s_d_op; /* default d_op for dentries */
1385 
1386         /*
1387          * Saved pool identifier for cleancache (-1 means none)
1388          */
1389         int cleancache_poolid;
1390 
1391         struct shrinker s_shrink;       /* per-sb shrinker handle */
1392 
1393         /* Number of inodes with nlink == 0 but still referenced */
1394         atomic_long_t s_remove_count;
1395 
1396         /* Being remounted read-only */
1397         int s_readonly_remount;
1398 
1399         /* AIO completions deferred from interrupt context */
1400         struct workqueue_struct *s_dio_done_wq;
1401         struct hlist_head s_pins;
1402 
1403         /*
1404          * Keep the lru lists last in the structure so they always sit on their
1405          * own individual cachelines.
1406          */
1407         struct list_lru         s_dentry_lru ____cacheline_aligned_in_smp;
1408         struct list_lru         s_inode_lru ____cacheline_aligned_in_smp;
1409         struct rcu_head         rcu;
1410         struct work_struct      destroy_work;
1411 
1412         struct mutex            s_sync_lock;    /* sync serialisation lock */
1413 
1414         /*
1415          * Indicates how deep in a filesystem stack this SB is
1416          */
1417         int s_stack_depth;
1418 
1419         /* s_inode_list_lock protects s_inodes */
1420         spinlock_t              s_inode_list_lock ____cacheline_aligned_in_smp;
1421         struct list_head        s_inodes;       /* all inodes */
1422 };
1423 
1424 extern struct timespec current_fs_time(struct super_block *sb);
1425 
1426 /*
1427  * Snapshotting support.
1428  */
1429 
1430 void __sb_end_write(struct super_block *sb, int level);
1431 int __sb_start_write(struct super_block *sb, int level, bool wait);
1432 
1433 #define __sb_writers_acquired(sb, lev)  \
1434         percpu_rwsem_acquire(&(sb)->s_writers.rw_sem[(lev)-1], 1, _THIS_IP_)
1435 #define __sb_writers_release(sb, lev)   \
1436         percpu_rwsem_release(&(sb)->s_writers.rw_sem[(lev)-1], 1, _THIS_IP_)
1437 
1438 /**
1439  * sb_end_write - drop write access to a superblock
1440  * @sb: the super we wrote to
1441  *
1442  * Decrement number of writers to the filesystem. Wake up possible waiters
1443  * wanting to freeze the filesystem.
1444  */
1445 static inline void sb_end_write(struct super_block *sb)
1446 {
1447         __sb_end_write(sb, SB_FREEZE_WRITE);
1448 }
1449 
1450 /**
1451  * sb_end_pagefault - drop write access to a superblock from a page fault
1452  * @sb: the super we wrote to
1453  *
1454  * Decrement number of processes handling write page fault to the filesystem.
1455  * Wake up possible waiters wanting to freeze the filesystem.
1456  */
1457 static inline void sb_end_pagefault(struct super_block *sb)
1458 {
1459         __sb_end_write(sb, SB_FREEZE_PAGEFAULT);
1460 }
1461 
1462 /**
1463  * sb_end_intwrite - drop write access to a superblock for internal fs purposes
1464  * @sb: the super we wrote to
1465  *
1466  * Decrement fs-internal number of writers to the filesystem.  Wake up possible
1467  * waiters wanting to freeze the filesystem.
1468  */
1469 static inline void sb_end_intwrite(struct super_block *sb)
1470 {
1471         __sb_end_write(sb, SB_FREEZE_FS);
1472 }
1473 
1474 /**
1475  * sb_start_write - get write access to a superblock
1476  * @sb: the super we write to
1477  *
1478  * When a process wants to write data or metadata to a file system (i.e. dirty
1479  * a page or an inode), it should embed the operation in a sb_start_write() -
1480  * sb_end_write() pair to get exclusion against file system freezing. This
1481  * function increments number of writers preventing freezing. If the file
1482  * system is already frozen, the function waits until the file system is
1483  * thawed.
1484  *
1485  * Since freeze protection behaves as a lock, users have to preserve
1486  * ordering of freeze protection and other filesystem locks. Generally,
1487  * freeze protection should be the outermost lock. In particular, we have:
1488  *
1489  * sb_start_write
1490  *   -> i_mutex                 (write path, truncate, directory ops, ...)
1491  *   -> s_umount                (freeze_super, thaw_super)
1492  */
1493 static inline void sb_start_write(struct super_block *sb)
1494 {
1495         __sb_start_write(sb, SB_FREEZE_WRITE, true);
1496 }
1497 
1498 static inline int sb_start_write_trylock(struct super_block *sb)
1499 {
1500         return __sb_start_write(sb, SB_FREEZE_WRITE, false);
1501 }
1502 
1503 /**
1504  * sb_start_pagefault - get write access to a superblock from a page fault
1505  * @sb: the super we write to
1506  *
1507  * When a process starts handling write page fault, it should embed the
1508  * operation into sb_start_pagefault() - sb_end_pagefault() pair to get
1509  * exclusion against file system freezing. This is needed since the page fault
1510  * is going to dirty a page. This function increments number of running page
1511  * faults preventing freezing. If the file system is already frozen, the
1512  * function waits until the file system is thawed.
1513  *
1514  * Since page fault freeze protection behaves as a lock, users have to preserve
1515  * ordering of freeze protection and other filesystem locks. It is advised to
1516  * put sb_start_pagefault() close to mmap_sem in lock ordering. Page fault
1517  * handling code implies lock dependency:
1518  *
1519  * mmap_sem
1520  *   -> sb_start_pagefault
1521  */
1522 static inline void sb_start_pagefault(struct super_block *sb)
1523 {
1524         __sb_start_write(sb, SB_FREEZE_PAGEFAULT, true);
1525 }
1526 
1527 /*
1528  * sb_start_intwrite - get write access to a superblock for internal fs purposes
1529  * @sb: the super we write to
1530  *
1531  * This is the third level of protection against filesystem freezing. It is
1532  * free for use by a filesystem. The only requirement is that it must rank
1533  * below sb_start_pagefault.
1534  *
1535  * For example filesystem can call sb_start_intwrite() when starting a
1536  * transaction which somewhat eases handling of freezing for internal sources
1537  * of filesystem changes (internal fs threads, discarding preallocation on file
1538  * close, etc.).
1539  */
1540 static inline void sb_start_intwrite(struct super_block *sb)
1541 {
1542         __sb_start_write(sb, SB_FREEZE_FS, true);
1543 }
1544 
1545 
1546 extern bool inode_owner_or_capable(const struct inode *inode);
1547 
1548 /*
1549  * VFS helper functions..
1550  */
1551 extern int vfs_create(struct inode *, struct dentry *, umode_t, bool);
1552 extern int vfs_mkdir(struct inode *, struct dentry *, umode_t);
1553 extern int vfs_mknod(struct inode *, struct dentry *, umode_t, dev_t);
1554 extern int vfs_symlink(struct inode *, struct dentry *, const char *);
1555 extern int vfs_link(struct dentry *, struct inode *, struct dentry *, struct inode **);
1556 extern int vfs_rmdir(struct inode *, struct dentry *);
1557 extern int vfs_unlink(struct inode *, struct dentry *, struct inode **);
1558 extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *, struct inode **, unsigned int);
1559 extern int vfs_whiteout(struct inode *, struct dentry *);
1560 
1561 /*
1562  * VFS file helper functions.
1563  */
1564 extern void inode_init_owner(struct inode *inode, const struct inode *dir,
1565                         umode_t mode);
1566 /*
1567  * VFS FS_IOC_FIEMAP helper definitions.
1568  */
1569 struct fiemap_extent_info {
1570         unsigned int fi_flags;          /* Flags as passed from user */
1571         unsigned int fi_extents_mapped; /* Number of mapped extents */
1572         unsigned int fi_extents_max;    /* Size of fiemap_extent array */
1573         struct fiemap_extent __user *fi_extents_start; /* Start of
1574                                                         fiemap_extent array */
1575 };
1576 int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
1577                             u64 phys, u64 len, u32 flags);
1578 int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
1579 
1580 /*
1581  * File types
1582  *
1583  * NOTE! These match bits 12..15 of stat.st_mode
1584  * (ie "(i_mode >> 12) & 15").
1585  */
1586 #define DT_UNKNOWN      0
1587 #define DT_FIFO         1
1588 #define DT_CHR          2
1589 #define DT_DIR          4
1590 #define DT_BLK          6
1591 #define DT_REG          8
1592 #define DT_LNK          10
1593 #define DT_SOCK         12
1594 #define DT_WHT          14
1595 
1596 /*
1597  * This is the "filldir" function type, used by readdir() to let
1598  * the kernel specify what kind of dirent layout it wants to have.
1599  * This allows the kernel to read directories into kernel space or
1600  * to have different dirent layouts depending on the binary type.
1601  */
1602 struct dir_context;
1603 typedef int (*filldir_t)(struct dir_context *, const char *, int, loff_t, u64,
1604                          unsigned);
1605 
1606 struct dir_context {
1607         const filldir_t actor;
1608         loff_t pos;
1609 };
1610 
1611 struct block_device_operations;
1612 
1613 /* These macros are for out of kernel modules to test that
1614  * the kernel supports the unlocked_ioctl and compat_ioctl
1615  * fields in struct file_operations. */
1616 #define HAVE_COMPAT_IOCTL 1
1617 #define HAVE_UNLOCKED_IOCTL 1
1618 
1619 /*
1620  * These flags let !MMU mmap() govern direct device mapping vs immediate
1621  * copying more easily for MAP_PRIVATE, especially for ROM filesystems.
1622  *
1623  * NOMMU_MAP_COPY:      Copy can be mapped (MAP_PRIVATE)
1624  * NOMMU_MAP_DIRECT:    Can be mapped directly (MAP_SHARED)
1625  * NOMMU_MAP_READ:      Can be mapped for reading
1626  * NOMMU_MAP_WRITE:     Can be mapped for writing
1627  * NOMMU_MAP_EXEC:      Can be mapped for execution
1628  */
1629 #define NOMMU_MAP_COPY          0x00000001
1630 #define NOMMU_MAP_DIRECT        0x00000008
1631 #define NOMMU_MAP_READ          VM_MAYREAD
1632 #define NOMMU_MAP_WRITE         VM_MAYWRITE
1633 #define NOMMU_MAP_EXEC          VM_MAYEXEC
1634 
1635 #define NOMMU_VMFLAGS \
1636         (NOMMU_MAP_READ | NOMMU_MAP_WRITE | NOMMU_MAP_EXEC)
1637 
1638 
1639 struct iov_iter;
1640 
1641 struct file_operations {
1642         struct module *owner;
1643         loff_t (*llseek) (struct file *, loff_t, int);
1644         ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1645         ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
1646         ssize_t (*read_iter) (struct kiocb *, struct iov_iter *);
1647         ssize_t (*write_iter) (struct kiocb *, struct iov_iter *);
1648         int (*iterate) (struct file *, struct dir_context *);
1649         unsigned int (*poll) (struct file *, struct poll_table_struct *);
1650         long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1651         long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1652         int (*mmap) (struct file *, struct vm_area_struct *);
1653         int (*open) (struct inode *, struct file *);
1654         int (*flush) (struct file *, fl_owner_t id);
1655         int (*release) (struct inode *, struct file *);
1656         int (*fsync) (struct file *, loff_t, loff_t, int datasync);
1657         int (*aio_fsync) (struct kiocb *, int datasync);
1658         int (*fasync) (int, struct file *, int);
1659         int (*lock) (struct file *, int, struct file_lock *);
1660         ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1661         unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1662         int (*check_flags)(int);
1663         int (*flock) (struct file *, int, struct file_lock *);
1664         ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1665         ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
1666         int (*setlease)(struct file *, long, struct file_lock **, void **);
1667         long (*fallocate)(struct file *file, int mode, loff_t offset,
1668                           loff_t len);
1669         void (*show_fdinfo)(struct seq_file *m, struct file *f);
1670 #ifndef CONFIG_MMU
1671         unsigned (*mmap_capabilities)(struct file *);
1672 #endif
1673         ssize_t (*copy_file_range)(struct file *, loff_t, struct file *,
1674                         loff_t, size_t, unsigned int);
1675         int (*clone_file_range)(struct file *, loff_t, struct file *, loff_t,
1676                         u64);
1677         ssize_t (*dedupe_file_range)(struct file *, u64, u64, struct file *,
1678                         u64);
1679 };
1680 
1681 struct inode_operations {
1682         struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int);
1683         const char * (*get_link) (struct dentry *, struct inode *, struct delayed_call *);
1684         int (*permission) (struct inode *, int);
1685         struct posix_acl * (*get_acl)(struct inode *, int);
1686 
1687         int (*readlink) (struct dentry *, char __user *,int);
1688 
1689         int (*create) (struct inode *,struct dentry *, umode_t, bool);
1690         int (*link) (struct dentry *,struct inode *,struct dentry *);
1691         int (*unlink) (struct inode *,struct dentry *);
1692         int (*symlink) (struct inode *,struct dentry *,const char *);
1693         int (*mkdir) (struct inode *,struct dentry *,umode_t);
1694         int (*rmdir) (struct inode *,struct dentry *);
1695         int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t);
1696         int (*rename) (struct inode *, struct dentry *,
1697                         struct inode *, struct dentry *);
1698         int (*rename2) (struct inode *, struct dentry *,
1699                         struct inode *, struct dentry *, unsigned int);
1700         int (*setattr) (struct dentry *, struct iattr *);
1701         int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1702         int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1703         ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1704         ssize_t (*listxattr) (struct dentry *, char *, size_t);
1705         int (*removexattr) (struct dentry *, const char *);
1706         int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
1707                       u64 len);
1708         int (*update_time)(struct inode *, struct timespec *, int);
1709         int (*atomic_open)(struct inode *, struct dentry *,
1710                            struct file *, unsigned open_flag,
1711                            umode_t create_mode, int *opened);
1712         int (*tmpfile) (struct inode *, struct dentry *, umode_t);
1713         int (*set_acl)(struct inode *, struct posix_acl *, int);
1714 } ____cacheline_aligned;
1715 
1716 ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
1717                               unsigned long nr_segs, unsigned long fast_segs,
1718                               struct iovec *fast_pointer,
1719                               struct iovec **ret_pointer);
1720 
1721 extern ssize_t __vfs_read(struct file *, char __user *, size_t, loff_t *);
1722 extern ssize_t __vfs_write(struct file *, const char __user *, size_t, loff_t *);
1723 extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1724 extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1725 extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1726                 unsigned long, loff_t *, int);
1727 extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1728                 unsigned long, loff_t *, int);
1729 extern ssize_t vfs_copy_file_range(struct file *, loff_t , struct file *,
1730                                    loff_t, size_t, unsigned int);
1731 extern int vfs_clone_file_range(struct file *file_in, loff_t pos_in,
1732                 struct file *file_out, loff_t pos_out, u64 len);
1733 extern int vfs_dedupe_file_range(struct file *file,
1734                                  struct file_dedupe_range *same);
1735 
1736 struct super_operations {
1737         struct inode *(*alloc_inode)(struct super_block *sb);
1738         void (*destroy_inode)(struct inode *);
1739 
1740         void (*dirty_inode) (struct inode *, int flags);
1741         int (*write_inode) (struct inode *, struct writeback_control *wbc);
1742         int (*drop_inode) (struct inode *);
1743         void (*evict_inode) (struct inode *);
1744         void (*put_super) (struct super_block *);
1745         int (*sync_fs)(struct super_block *sb, int wait);
1746         int (*freeze_super) (struct super_block *);
1747         int (*freeze_fs) (struct super_block *);
1748         int (*thaw_super) (struct super_block *);
1749         int (*unfreeze_fs) (struct super_block *);
1750         int (*statfs) (struct dentry *, struct kstatfs *);
1751         int (*remount_fs) (struct super_block *, int *, char *);
1752         void (*umount_begin) (struct super_block *);
1753 
1754         int (*show_options)(struct seq_file *, struct dentry *);
1755         int (*show_devname)(struct seq_file *, struct dentry *);
1756         int (*show_path)(struct seq_file *, struct dentry *);
1757         int (*show_stats)(struct seq_file *, struct dentry *);
1758 #ifdef CONFIG_QUOTA
1759         ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1760         ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
1761         struct dquot **(*get_dquots)(struct inode *);
1762 #endif
1763         int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
1764         long (*nr_cached_objects)(struct super_block *,
1765                                   struct shrink_control *);
1766         long (*free_cached_objects)(struct super_block *,
1767                                     struct shrink_control *);
1768 };
1769 
1770 /*
1771  * Inode flags - they have no relation to superblock flags now
1772  */
1773 #define S_SYNC          1       /* Writes are synced at once */
1774 #define S_NOATIME       2       /* Do not update access times */
1775 #define S_APPEND        4       /* Append-only file */
1776 #define S_IMMUTABLE     8       /* Immutable file */
1777 #define S_DEAD          16      /* removed, but still open directory */
1778 #define S_NOQUOTA       32      /* Inode is not counted to quota */
1779 #define S_DIRSYNC       64      /* Directory modifications are synchronous */
1780 #define S_NOCMTIME      128     /* Do not update file c/mtime */
1781 #define S_SWAPFILE      256     /* Do not truncate: swapon got its bmaps */
1782 #define S_PRIVATE       512     /* Inode is fs-internal */
1783 #define S_IMA           1024    /* Inode has an associated IMA struct */
1784 #define S_AUTOMOUNT     2048    /* Automount/referral quasi-directory */
1785 #define S_NOSEC         4096    /* no suid or xattr security attributes */
1786 #ifdef CONFIG_FS_DAX
1787 #define S_DAX           8192    /* Direct Access, avoiding the page cache */
1788 #else
1789 #define S_DAX           0       /* Make all the DAX code disappear */
1790 #endif
1791 
1792 /*
1793  * Note that nosuid etc flags are inode-specific: setting some file-system
1794  * flags just means all the inodes inherit those flags by default. It might be
1795  * possible to override it selectively if you really wanted to with some
1796  * ioctl() that is not currently implemented.
1797  *
1798  * Exception: MS_RDONLY is always applied to the entire file system.
1799  *
1800  * Unfortunately, it is possible to change a filesystems flags with it mounted
1801  * with files in use.  This means that all of the inodes will not have their
1802  * i_flags updated.  Hence, i_flags no longer inherit the superblock mount
1803  * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
1804  */
1805 #define __IS_FLG(inode, flg)    ((inode)->i_sb->s_flags & (flg))
1806 
1807 #define IS_RDONLY(inode)        ((inode)->i_sb->s_flags & MS_RDONLY)
1808 #define IS_SYNC(inode)          (__IS_FLG(inode, MS_SYNCHRONOUS) || \
1809                                         ((inode)->i_flags & S_SYNC))
1810 #define IS_DIRSYNC(inode)       (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
1811                                         ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
1812 #define IS_MANDLOCK(inode)      __IS_FLG(inode, MS_MANDLOCK)
1813 #define IS_NOATIME(inode)       __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
1814 #define IS_I_VERSION(inode)     __IS_FLG(inode, MS_I_VERSION)
1815 
1816 #define IS_NOQUOTA(inode)       ((inode)->i_flags & S_NOQUOTA)
1817 #define IS_APPEND(inode)        ((inode)->i_flags & S_APPEND)
1818 #define IS_IMMUTABLE(inode)     ((inode)->i_flags & S_IMMUTABLE)
1819 #define IS_POSIXACL(inode)      __IS_FLG(inode, MS_POSIXACL)
1820 
1821 #define IS_DEADDIR(inode)       ((inode)->i_flags & S_DEAD)
1822 #define IS_NOCMTIME(inode)      ((inode)->i_flags & S_NOCMTIME)
1823 #define IS_SWAPFILE(inode)      ((inode)->i_flags & S_SWAPFILE)
1824 #define IS_PRIVATE(inode)       ((inode)->i_flags & S_PRIVATE)
1825 #define IS_IMA(inode)           ((inode)->i_flags & S_IMA)
1826 #define IS_AUTOMOUNT(inode)     ((inode)->i_flags & S_AUTOMOUNT)
1827 #define IS_NOSEC(inode)         ((inode)->i_flags & S_NOSEC)
1828 #define IS_DAX(inode)           ((inode)->i_flags & S_DAX)
1829 
1830 #define IS_WHITEOUT(inode)      (S_ISCHR(inode->i_mode) && \
1831                                  (inode)->i_rdev == WHITEOUT_DEV)
1832 
1833 /*
1834  * Inode state bits.  Protected by inode->i_lock
1835  *
1836  * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1837  * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1838  *
1839  * Four bits define the lifetime of an inode.  Initially, inodes are I_NEW,
1840  * until that flag is cleared.  I_WILL_FREE, I_FREEING and I_CLEAR are set at
1841  * various stages of removing an inode.
1842  *
1843  * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1844  *
1845  * I_DIRTY_SYNC         Inode is dirty, but doesn't have to be written on
1846  *                      fdatasync().  i_atime is the usual cause.
1847  * I_DIRTY_DATASYNC     Data-related inode changes pending. We keep track of
1848  *                      these changes separately from I_DIRTY_SYNC so that we
1849  *                      don't have to write inode on fdatasync() when only
1850  *                      mtime has changed in it.
1851  * I_DIRTY_PAGES        Inode has dirty pages.  Inode itself may be clean.
1852  * I_NEW                Serves as both a mutex and completion notification.
1853  *                      New inodes set I_NEW.  If two processes both create
1854  *                      the same inode, one of them will release its inode and
1855  *                      wait for I_NEW to be released before returning.
1856  *                      Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1857  *                      also cause waiting on I_NEW, without I_NEW actually
1858  *                      being set.  find_inode() uses this to prevent returning
1859  *                      nearly-dead inodes.
1860  * I_WILL_FREE          Must be set when calling write_inode_now() if i_count
1861  *                      is zero.  I_FREEING must be set when I_WILL_FREE is
1862  *                      cleared.
1863  * I_FREEING            Set when inode is about to be freed but still has dirty
1864  *                      pages or buffers attached or the inode itself is still
1865  *                      dirty.
1866  * I_CLEAR              Added by clear_inode().  In this state the inode is
1867  *                      clean and can be destroyed.  Inode keeps I_FREEING.
1868  *
1869  *                      Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1870  *                      prohibited for many purposes.  iget() must wait for
1871  *                      the inode to be completely released, then create it
1872  *                      anew.  Other functions will just ignore such inodes,
1873  *                      if appropriate.  I_NEW is used for waiting.
1874  *
1875  * I_SYNC               Writeback of inode is running. The bit is set during
1876  *                      data writeback, and cleared with a wakeup on the bit
1877  *                      address once it is done. The bit is also used to pin
1878  *                      the inode in memory for flusher thread.
1879  *
1880  * I_REFERENCED         Marks the inode as recently references on the LRU list.
1881  *
1882  * I_DIO_WAKEUP         Never set.  Only used as a key for wait_on_bit().
1883  *
1884  * I_WB_SWITCH          Cgroup bdi_writeback switching in progress.  Used to
1885  *                      synchronize competing switching instances and to tell
1886  *                      wb stat updates to grab mapping->tree_lock.  See
1887  *                      inode_switch_wb_work_fn() for details.
1888  *
1889  * Q: What is the difference between I_WILL_FREE and I_FREEING?
1890  */
1891 #define I_DIRTY_SYNC            (1 << 0)
1892 #define I_DIRTY_DATASYNC        (1 << 1)
1893 #define I_DIRTY_PAGES           (1 << 2)
1894 #define __I_NEW                 3
1895 #define I_NEW                   (1 << __I_NEW)
1896 #define I_WILL_FREE             (1 << 4)
1897 #define I_FREEING               (1 << 5)
1898 #define I_CLEAR                 (1 << 6)
1899 #define __I_SYNC                7
1900 #define I_SYNC                  (1 << __I_SYNC)
1901 #define I_REFERENCED            (1 << 8)
1902 #define __I_DIO_WAKEUP          9
1903 #define I_DIO_WAKEUP            (1 << __I_DIO_WAKEUP)
1904 #define I_LINKABLE              (1 << 10)
1905 #define I_DIRTY_TIME            (1 << 11)
1906 #define __I_DIRTY_TIME_EXPIRED  12
1907 #define I_DIRTY_TIME_EXPIRED    (1 << __I_DIRTY_TIME_EXPIRED)
1908 #define I_WB_SWITCH             (1 << 13)
1909 
1910 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1911 #define I_DIRTY_ALL (I_DIRTY | I_DIRTY_TIME)
1912 
1913 extern void __mark_inode_dirty(struct inode *, int);
1914 static inline void mark_inode_dirty(struct inode *inode)
1915 {
1916         __mark_inode_dirty(inode, I_DIRTY);
1917 }
1918 
1919 static inline void mark_inode_dirty_sync(struct inode *inode)
1920 {
1921         __mark_inode_dirty(inode, I_DIRTY_SYNC);
1922 }
1923 
1924 extern void inc_nlink(struct inode *inode);
1925 extern void drop_nlink(struct inode *inode);
1926 extern void clear_nlink(struct inode *inode);
1927 extern void set_nlink(struct inode *inode, unsigned int nlink);
1928 
1929 static inline void inode_inc_link_count(struct inode *inode)
1930 {
1931         inc_nlink(inode);
1932         mark_inode_dirty(inode);
1933 }
1934 
1935 static inline void inode_dec_link_count(struct inode *inode)
1936 {
1937         drop_nlink(inode);
1938         mark_inode_dirty(inode);
1939 }
1940 
1941 /**
1942  * inode_inc_iversion - increments i_version
1943  * @inode: inode that need to be updated
1944  *
1945  * Every time the inode is modified, the i_version field will be incremented.
1946  * The filesystem has to be mounted with i_version flag
1947  */
1948 
1949 static inline void inode_inc_iversion(struct inode *inode)
1950 {
1951        spin_lock(&inode->i_lock);
1952        inode->i_version++;
1953        spin_unlock(&inode->i_lock);
1954 }
1955 
1956 enum file_time_flags {
1957         S_ATIME = 1,
1958         S_MTIME = 2,
1959         S_CTIME = 4,
1960         S_VERSION = 8,
1961 };
1962 
1963 extern bool atime_needs_update(const struct path *, struct inode *);
1964 extern void touch_atime(const struct path *);
1965 static inline void file_accessed(struct file *file)
1966 {
1967         if (!(file->f_flags & O_NOATIME))
1968                 touch_atime(&file->f_path);
1969 }
1970 
1971 int sync_inode(struct inode *inode, struct writeback_control *wbc);
1972 int sync_inode_metadata(struct inode *inode, int wait);
1973 
1974 struct file_system_type {
1975         const char *name;
1976         int fs_flags;
1977 #define FS_REQUIRES_DEV         1 
1978 #define FS_BINARY_MOUNTDATA     2
1979 #define FS_HAS_SUBTYPE          4
1980 #define FS_USERNS_MOUNT         8       /* Can be mounted by userns root */
1981 #define FS_USERNS_DEV_MOUNT     16 /* A userns mount does not imply MNT_NODEV */
1982 #define FS_USERNS_VISIBLE       32      /* FS must already be visible */
1983 #define FS_RENAME_DOES_D_MOVE   32768   /* FS will handle d_move() during rename() internally. */
1984         struct dentry *(*mount) (struct file_system_type *, int,
1985                        const char *, void *);
1986         void (*kill_sb) (struct super_block *);
1987         struct module *owner;
1988         struct file_system_type * next;
1989         struct hlist_head fs_supers;
1990 
1991         struct lock_class_key s_lock_key;
1992         struct lock_class_key s_umount_key;
1993         struct lock_class_key s_vfs_rename_key;
1994         struct lock_class_key s_writers_key[SB_FREEZE_LEVELS];
1995 
1996         struct lock_class_key i_lock_key;
1997         struct lock_class_key i_mutex_key;
1998         struct lock_class_key i_mutex_dir_key;
1999 };
2000 
2001 #define MODULE_ALIAS_FS(NAME) MODULE_ALIAS("fs-" NAME)
2002 
2003 extern struct dentry *mount_ns(struct file_system_type *fs_type, int flags,
2004         void *data, int (*fill_super)(struct super_block *, void *, int));
2005 extern struct dentry *mount_bdev(struct file_system_type *fs_type,
2006         int flags, const char *dev_name, void *data,
2007         int (*fill_super)(struct super_block *, void *, int));
2008 extern struct dentry *mount_single(struct file_system_type *fs_type,
2009         int flags, void *data,
2010         int (*fill_super)(struct super_block *, void *, int));
2011 extern struct dentry *mount_nodev(struct file_system_type *fs_type,
2012         int flags, void *data,
2013         int (*fill_super)(struct super_block *, void *, int));
2014 extern struct dentry *mount_subtree(struct vfsmount *mnt, const char *path);
2015 void generic_shutdown_super(struct super_block *sb);
2016 void kill_block_super(struct super_block *sb);
2017 void kill_anon_super(struct super_block *sb);
2018 void kill_litter_super(struct super_block *sb);
2019 void deactivate_super(struct super_block *sb);
2020 void deactivate_locked_super(struct super_block *sb);
2021 int set_anon_super(struct super_block *s, void *data);
2022 int get_anon_bdev(dev_t *);
2023 void free_anon_bdev(dev_t);
2024 struct super_block *sget(struct file_system_type *type,
2025                         int (*test)(struct super_block *,void *),
2026                         int (*set)(struct super_block *,void *),
2027                         int flags, void *data);
2028 extern struct dentry *mount_pseudo(struct file_system_type *, char *,
2029         const struct super_operations *ops,
2030         const struct dentry_operations *dops,
2031         unsigned long);
2032 
2033 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
2034 #define fops_get(fops) \
2035         (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
2036 #define fops_put(fops) \
2037         do { if (fops) module_put((fops)->owner); } while(0)
2038 /*
2039  * This one is to be used *ONLY* from ->open() instances.
2040  * fops must be non-NULL, pinned down *and* module dependencies
2041  * should be sufficient to pin the caller down as well.
2042  */
2043 #define replace_fops(f, fops) \
2044         do {    \
2045                 struct file *__file = (f); \
2046                 fops_put(__file->f_op); \
2047                 BUG_ON(!(__file->f_op = (fops))); \
2048         } while(0)
2049 
2050 extern int register_filesystem(struct file_system_type *);
2051 extern int unregister_filesystem(struct file_system_type *);
2052 extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
2053 #define kern_mount(type) kern_mount_data(type, NULL)
2054 extern void kern_unmount(struct vfsmount *mnt);
2055 extern int may_umount_tree(struct vfsmount *);
2056 extern int may_umount(struct vfsmount *);
2057 extern long do_mount(const char *, const char __user *,
2058                      const char *, unsigned long, void *);
2059 extern struct vfsmount *collect_mounts(struct path *);
2060 extern void drop_collected_mounts(struct vfsmount *);
2061 extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
2062                           struct vfsmount *);
2063 extern int vfs_statfs(struct path *, struct kstatfs *);
2064 extern int user_statfs(const char __user *, struct kstatfs *);
2065 extern int fd_statfs(int, struct kstatfs *);
2066 extern int vfs_ustat(dev_t, struct kstatfs *);
2067 extern int freeze_super(struct super_block *super);
2068 extern int thaw_super(struct super_block *super);
2069 extern bool our_mnt(struct vfsmount *mnt);
2070 
2071 extern int current_umask(void);
2072 
2073 extern void ihold(struct inode * inode);
2074 extern void iput(struct inode *);
2075 extern int generic_update_time(struct inode *, struct timespec *, int);
2076 
2077 /* /sys/fs */
2078 extern struct kobject *fs_kobj;
2079 
2080 #define MAX_RW_COUNT (INT_MAX & PAGE_MASK)
2081 
2082 #ifdef CONFIG_MANDATORY_FILE_LOCKING
2083 extern int locks_mandatory_locked(struct file *);
2084 extern int locks_mandatory_area(struct inode *, struct file *, loff_t, loff_t, unsigned char);
2085 
2086 /*
2087  * Candidates for mandatory locking have the setgid bit set
2088  * but no group execute bit -  an otherwise meaningless combination.
2089  */
2090 
2091 static inline int __mandatory_lock(struct inode *ino)
2092 {
2093         return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
2094 }
2095 
2096 /*
2097  * ... and these candidates should be on MS_MANDLOCK mounted fs,
2098  * otherwise these will be advisory locks
2099  */
2100 
2101 static inline int mandatory_lock(struct inode *ino)
2102 {
2103         return IS_MANDLOCK(ino) && __mandatory_lock(ino);
2104 }
2105 
2106 static inline int locks_verify_locked(struct file *file)
2107 {
2108         if (mandatory_lock(file_inode(file)))
2109                 return locks_mandatory_locked(file);
2110         return 0;
2111 }
2112 
2113 static inline int locks_verify_truncate(struct inode *inode,
2114                                     struct file *f,
2115                                     loff_t size)
2116 {
2117         if (!inode->i_flctx || !mandatory_lock(inode))
2118                 return 0;
2119 
2120         if (size < inode->i_size) {
2121                 return locks_mandatory_area(inode, f, size, inode->i_size - 1,
2122                                 F_WRLCK);
2123         } else {
2124                 return locks_mandatory_area(inode, f, inode->i_size, size - 1,
2125                                 F_WRLCK);
2126         }
2127 }
2128 
2129 #else /* !CONFIG_MANDATORY_FILE_LOCKING */
2130 
2131 static inline int locks_mandatory_locked(struct file *file)
2132 {
2133         return 0;
2134 }
2135 
2136 static inline int locks_mandatory_area(struct inode *inode, struct file *filp,
2137                                        loff_t start, loff_t end, unsigned char type)
2138 {
2139         return 0;
2140 }
2141 
2142 static inline int __mandatory_lock(struct inode *inode)
2143 {
2144         return 0;
2145 }
2146 
2147 static inline int mandatory_lock(struct inode *inode)
2148 {
2149         return 0;
2150 }
2151 
2152 static inline int locks_verify_locked(struct file *file)
2153 {
2154         return 0;
2155 }
2156 
2157 static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
2158                                         size_t size)
2159 {
2160         return 0;
2161 }
2162 
2163 #endif /* CONFIG_MANDATORY_FILE_LOCKING */
2164 
2165 
2166 #ifdef CONFIG_FILE_LOCKING
2167 static inline int break_lease(struct inode *inode, unsigned int mode)
2168 {
2169         /*
2170          * Since this check is lockless, we must ensure that any refcounts
2171          * taken are done before checking i_flctx->flc_lease. Otherwise, we
2172          * could end up racing with tasks trying to set a new lease on this
2173          * file.
2174          */
2175         smp_mb();
2176         if (inode->i_flctx && !list_empty_careful(&inode->i_flctx->flc_lease))
2177                 return __break_lease(inode, mode, FL_LEASE);
2178         return 0;
2179 }
2180 
2181 static inline int break_deleg(struct inode *inode, unsigned int mode)
2182 {
2183         /*
2184          * Since this check is lockless, we must ensure that any refcounts
2185          * taken are done before checking i_flctx->flc_lease. Otherwise, we
2186          * could end up racing with tasks trying to set a new lease on this
2187          * file.
2188          */
2189         smp_mb();
2190         if (inode->i_flctx && !list_empty_careful(&inode->i_flctx->flc_lease))
2191                 return __break_lease(inode, mode, FL_DELEG);
2192         return 0;
2193 }
2194 
2195 static inline int try_break_deleg(struct inode *inode, struct inode **delegated_inode)
2196 {
2197         int ret;
2198 
2199         ret = break_deleg(inode, O_WRONLY|O_NONBLOCK);
2200         if (ret == -EWOULDBLOCK && delegated_inode) {
2201                 *delegated_inode = inode;
2202                 ihold(inode);
2203         }
2204         return ret;
2205 }
2206 
2207 static inline int break_deleg_wait(struct inode **delegated_inode)
2208 {
2209         int ret;
2210 
2211         ret = break_deleg(*delegated_inode, O_WRONLY);
2212         iput(*delegated_inode);
2213         *delegated_inode = NULL;
2214         return ret;
2215 }
2216 
2217 static inline int break_layout(struct inode *inode, bool wait)
2218 {
2219         smp_mb();
2220         if (inode->i_flctx && !list_empty_careful(&inode->i_flctx->flc_lease))
2221                 return __break_lease(inode,
2222                                 wait ? O_WRONLY : O_WRONLY | O_NONBLOCK,
2223                                 FL_LAYOUT);
2224         return 0;
2225 }
2226 
2227 #else /* !CONFIG_FILE_LOCKING */
2228 static inline int break_lease(struct inode *inode, unsigned int mode)
2229 {
2230         return 0;
2231 }
2232 
2233 static inline int break_deleg(struct inode *inode, unsigned int mode)
2234 {
2235         return 0;
2236 }
2237 
2238 static inline int try_break_deleg(struct inode *inode, struct inode **delegated_inode)
2239 {
2240         return 0;
2241 }
2242 
2243 static inline int break_deleg_wait(struct inode **delegated_inode)
2244 {
2245         BUG();
2246         return 0;
2247 }
2248 
2249 static inline int break_layout(struct inode *inode, bool wait)
2250 {
2251         return 0;
2252 }
2253 
2254 #endif /* CONFIG_FILE_LOCKING */
2255 
2256 /* fs/open.c */
2257 struct audit_names;
2258 struct filename {
2259         const char              *name;  /* pointer to actual string */
2260         const __user char       *uptr;  /* original userland pointer */
2261         struct audit_names      *aname;
2262         int                     refcnt;
2263         const char              iname[];
2264 };
2265 
2266 extern long vfs_truncate(struct path *, loff_t);
2267 extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
2268                        struct file *filp);
2269 extern int vfs_fallocate(struct file *file, int mode, loff_t offset,
2270                         loff_t len);
2271 extern long do_sys_open(int dfd, const char __user *filename, int flags,
2272                         umode_t mode);
2273 extern struct file *file_open_name(struct filename *, int, umode_t);
2274 extern struct file *filp_open(const char *, int, umode_t);
2275 extern struct file *file_open_root(struct dentry *, struct vfsmount *,
2276                                    const char *, int, umode_t);
2277 extern struct file * dentry_open(const struct path *, int, const struct cred *);
2278 extern int filp_close(struct file *, fl_owner_t id);
2279 
2280 extern struct filename *getname_flags(const char __user *, int, int *);
2281 extern struct filename *getname(const char __user *);
2282 extern struct filename *getname_kernel(const char *);
2283 extern void putname(struct filename *name);
2284 
2285 enum {
2286         FILE_CREATED = 1,
2287         FILE_OPENED = 2
2288 };
2289 extern int finish_open(struct file *file, struct dentry *dentry,
2290                         int (*open)(struct inode *, struct file *),
2291                         int *opened);
2292 extern int finish_no_open(struct file *file, struct dentry *dentry);
2293 
2294 /* fs/ioctl.c */
2295 
2296 extern int ioctl_preallocate(struct file *filp, void __user *argp);
2297 
2298 /* fs/dcache.c */
2299 extern void __init vfs_caches_init_early(void);
2300 extern void __init vfs_caches_init(void);
2301 
2302 extern struct kmem_cache *names_cachep;
2303 
2304 #define __getname()             kmem_cache_alloc(names_cachep, GFP_KERNEL)
2305 #define __putname(name)         kmem_cache_free(names_cachep, (void *)(name))
2306 
2307 #ifdef CONFIG_BLOCK
2308 extern int register_blkdev(unsigned int, const char *);
2309 extern void unregister_blkdev(unsigned int, const char *);
2310 extern struct block_device *bdget(dev_t);
2311 extern struct block_device *bdgrab(struct block_device *bdev);
2312 extern void bd_set_size(struct block_device *, loff_t size);
2313 extern void bd_forget(struct inode *inode);
2314 extern void bdput(struct block_device *);
2315 extern void invalidate_bdev(struct block_device *);
2316 extern void iterate_bdevs(void (*)(struct block_device *, void *), void *);
2317 extern int sync_blockdev(struct block_device *bdev);
2318 extern void kill_bdev(struct block_device *);
2319 extern struct super_block *freeze_bdev(struct block_device *);
2320 extern void emergency_thaw_all(void);
2321 extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
2322 extern int fsync_bdev(struct block_device *);
2323 #ifdef CONFIG_FS_DAX
2324 extern bool blkdev_dax_capable(struct block_device *bdev);
2325 #else
2326 static inline bool blkdev_dax_capable(struct block_device *bdev)
2327 {
2328         return false;
2329 }
2330 #endif
2331 
2332 extern struct super_block *blockdev_superblock;
2333 
2334 static inline bool sb_is_blkdev_sb(struct super_block *sb)
2335 {
2336         return sb == blockdev_superblock;
2337 }
2338 #else
2339 static inline void bd_forget(struct inode *inode) {}
2340 static inline int sync_blockdev(struct block_device *bdev) { return 0; }
2341 static inline void kill_bdev(struct block_device *bdev) {}
2342 static inline void invalidate_bdev(struct block_device *bdev) {}
2343 
2344 static inline struct super_block *freeze_bdev(struct block_device *sb)
2345 {
2346         return NULL;
2347 }
2348 
2349 static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
2350 {
2351         return 0;
2352 }
2353 
2354 static inline void iterate_bdevs(void (*f)(struct block_device *, void *), void *arg)
2355 {
2356 }
2357 
2358 static inline bool sb_is_blkdev_sb(struct super_block *sb)
2359 {
2360         return false;
2361 }
2362 #endif
2363 extern int sync_filesystem(struct super_block *);
2364 extern const struct file_operations def_blk_fops;
2365 extern const struct file_operations def_chr_fops;
2366 #ifdef CONFIG_BLOCK
2367 extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
2368 extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
2369 extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
2370 extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
2371 extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
2372                                                void *holder);
2373 extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
2374                                               void *holder);
2375 extern void blkdev_put(struct block_device *bdev, fmode_t mode);
2376 extern int __blkdev_reread_part(struct block_device *bdev);
2377 extern int blkdev_reread_part(struct block_device *bdev);
2378 
2379 #ifdef CONFIG_SYSFS
2380 extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
2381 extern void bd_unlink_disk_holder(struct block_device *bdev,
2382                                   struct gendisk *disk);
2383 #else
2384 static inline int bd_link_disk_holder(struct block_device *bdev,
2385                                       struct gendisk *disk)
2386 {
2387         return 0;
2388 }
2389 static inline void bd_unlink_disk_holder(struct block_device *bdev,
2390                                          struct gendisk *disk)
2391 {
2392 }
2393 #endif
2394 #endif
2395 
2396 /* fs/char_dev.c */
2397 #define CHRDEV_MAJOR_HASH_SIZE  255
2398 extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
2399 extern int register_chrdev_region(dev_t, unsigned, const char *);
2400 extern int __register_chrdev(unsigned int major, unsigned int baseminor,
2401                              unsigned int count, const char *name,
2402                              const struct file_operations *fops);
2403 extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
2404                                 unsigned int count, const char *name);
2405 extern void unregister_chrdev_region(dev_t, unsigned);
2406 extern void chrdev_show(struct seq_file *,off_t);
2407 
2408 static inline int register_chrdev(unsigned int major, const char *name,
2409                                   const struct file_operations *fops)
2410 {
2411         return __register_chrdev(major, 0, 256, name, fops);
2412 }
2413 
2414 static inline void unregister_chrdev(unsigned int major, const char *name)
2415 {
2416         __unregister_chrdev(major, 0, 256, name);
2417 }
2418 
2419 /* fs/block_dev.c */
2420 #define BDEVNAME_SIZE   32      /* Largest string for a blockdev identifier */
2421 #define BDEVT_SIZE      10      /* Largest string for MAJ:MIN for blkdev */
2422 
2423 #ifdef CONFIG_BLOCK
2424 #define BLKDEV_MAJOR_HASH_SIZE  255
2425 extern const char *__bdevname(dev_t, char *buffer);
2426 extern const char *bdevname(struct block_device *bdev, char *buffer);
2427 extern struct block_device *lookup_bdev(const char *);
2428 extern void blkdev_show(struct seq_file *,off_t);
2429 
2430 #else
2431 #define BLKDEV_MAJOR_HASH_SIZE  0
2432 #endif
2433 
2434 extern void init_special_inode(struct inode *, umode_t, dev_t);
2435 
2436 /* Invalid inode operations -- fs/bad_inode.c */
2437 extern void make_bad_inode(struct inode *);
2438 extern bool is_bad_inode(struct inode *);
2439 
2440 #ifdef CONFIG_BLOCK
2441 /*
2442  * return READ, READA, or WRITE
2443  */
2444 #define bio_rw(bio)             ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2445 
2446 /*
2447  * return data direction, READ or WRITE
2448  */
2449 #define bio_data_dir(bio)       ((bio)->bi_rw & 1)
2450 
2451 extern void check_disk_size_change(struct gendisk *disk,
2452                                    struct block_device *bdev);
2453 extern int revalidate_disk(struct gendisk *);
2454 extern int check_disk_change(struct block_device *);
2455 extern int __invalidate_device(struct block_device *, bool);
2456 extern int invalidate_partition(struct gendisk *, int);
2457 #endif
2458 unsigned long invalidate_mapping_pages(struct address_space *mapping,
2459                                         pgoff_t start, pgoff_t end);
2460 
2461 static inline void invalidate_remote_inode(struct inode *inode)
2462 {
2463         if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
2464             S_ISLNK(inode->i_mode))
2465                 invalidate_mapping_pages(inode->i_mapping, 0, -1);
2466 }
2467 extern int invalidate_inode_pages2(struct address_space *mapping);
2468 extern int invalidate_inode_pages2_range(struct address_space *mapping,
2469                                          pgoff_t start, pgoff_t end);
2470 extern int write_inode_now(struct inode *, int);
2471 extern int filemap_fdatawrite(struct address_space *);
2472 extern int filemap_flush(struct address_space *);
2473 extern int filemap_fdatawait(struct address_space *);
2474 extern void filemap_fdatawait_keep_errors(struct address_space *);
2475 extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
2476                                    loff_t lend);
2477 extern int filemap_write_and_wait(struct address_space *mapping);
2478 extern int filemap_write_and_wait_range(struct address_space *mapping,
2479                                         loff_t lstart, loff_t lend);
2480 extern int __filemap_fdatawrite_range(struct address_space *mapping,
2481                                 loff_t start, loff_t end, int sync_mode);
2482 extern int filemap_fdatawrite_range(struct address_space *mapping,
2483                                 loff_t start, loff_t end);
2484 
2485 extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
2486                            int datasync);
2487 extern int vfs_fsync(struct file *file, int datasync);
2488 static inline int generic_write_sync(struct file *file, loff_t pos, loff_t count)
2489 {
2490         if (!(file->f_flags & O_DSYNC) && !IS_SYNC(file->f_mapping->host))
2491                 return 0;
2492         return vfs_fsync_range(file, pos, pos + count - 1,
2493                                (file->f_flags & __O_SYNC) ? 0 : 1);
2494 }
2495 extern void emergency_sync(void);
2496 extern void emergency_remount(void);
2497 #ifdef CONFIG_BLOCK
2498 extern sector_t bmap(struct inode *, sector_t);
2499 #endif
2500 extern int notify_change(struct dentry *, struct iattr *, struct inode **);
2501 extern int inode_permission(struct inode *, int);
2502 extern int __inode_permission(struct inode *, int);
2503 extern int generic_permission(struct inode *, int);
2504 extern int __check_sticky(struct inode *dir, struct inode *inode);
2505 
2506 static inline bool execute_ok(struct inode *inode)
2507 {
2508         return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
2509 }
2510 
2511 static inline void file_start_write(struct file *file)
2512 {
2513         if (!S_ISREG(file_inode(file)->i_mode))
2514                 return;
2515         __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, true);
2516 }
2517 
2518 static inline bool file_start_write_trylock(struct file *file)
2519 {
2520         if (!S_ISREG(file_inode(file)->i_mode))
2521                 return true;
2522         return __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, false);
2523 }
2524 
2525 static inline void file_end_write(struct file *file)
2526 {
2527         if (!S_ISREG(file_inode(file)->i_mode))
2528                 return;
2529         __sb_end_write(file_inode(file)->i_sb, SB_FREEZE_WRITE);
2530 }
2531 
2532 /*
2533  * get_write_access() gets write permission for a file.
2534  * put_write_access() releases this write permission.
2535  * This is used for regular files.
2536  * We cannot support write (and maybe mmap read-write shared) accesses and
2537  * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2538  * can have the following values:
2539  * 0: no writers, no VM_DENYWRITE mappings
2540  * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2541  * > 0: (i_writecount) users are writing to the file.
2542  *
2543  * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2544  * except for the cases where we don't hold i_writecount yet. Then we need to
2545  * use {get,deny}_write_access() - these functions check the sign and refuse
2546  * to do the change if sign is wrong.
2547  */
2548 static inline int get_write_access(struct inode *inode)
2549 {
2550         return atomic_inc_unless_negative(&inode->i_writecount) ? 0 : -ETXTBSY;
2551 }
2552 static inline int deny_write_access(struct file *file)
2553 {
2554         struct inode *inode = file_inode(file);
2555         return atomic_dec_unless_positive(&inode->i_writecount) ? 0 : -ETXTBSY;
2556 }
2557 static inline void put_write_access(struct inode * inode)
2558 {
2559         atomic_dec(&inode->i_writecount);
2560 }
2561 static inline void allow_write_access(struct file *file)
2562 {
2563         if (file)
2564                 atomic_inc(&file_inode(file)->i_writecount);
2565 }
2566 static inline bool inode_is_open_for_write(const struct inode *inode)
2567 {
2568         return atomic_read(&inode->i_writecount) > 0;
2569 }
2570 
2571 #ifdef CONFIG_IMA
2572 static inline void i_readcount_dec(struct inode *inode)
2573 {
2574         BUG_ON(!atomic_read(&inode->i_readcount));
2575         atomic_dec(&inode->i_readcount);
2576 }
2577 static inline void i_readcount_inc(struct inode *inode)
2578 {
2579         atomic_inc(&inode->i_readcount);
2580 }
2581 #else
2582 static inline void i_readcount_dec(struct inode *inode)
2583 {
2584         return;
2585 }
2586 static inline void i_readcount_inc(struct inode *inode)
2587 {
2588         return;
2589 }
2590 #endif
2591 extern int do_pipe_flags(int *, int);
2592 
2593 enum kernel_read_file_id {
2594         READING_FIRMWARE = 1,
2595         READING_MODULE,
2596         READING_KEXEC_IMAGE,
2597         READING_KEXEC_INITRAMFS,
2598         READING_POLICY,
2599         READING_MAX_ID
2600 };
2601 
2602 extern int kernel_read(struct file *, loff_t, char *, unsigned long);
2603 extern int kernel_read_file(struct file *, void **, loff_t *, loff_t,
2604                             enum kernel_read_file_id);
2605 extern int kernel_read_file_from_path(char *, void **, loff_t *, loff_t,
2606                                       enum kernel_read_file_id);
2607 extern int kernel_read_file_from_fd(int, void **, loff_t *, loff_t,
2608                                     enum kernel_read_file_id);
2609 extern ssize_t kernel_write(struct file *, const char *, size_t, loff_t);
2610 extern ssize_t __kernel_write(struct file *, const char *, size_t, loff_t *);
2611 extern struct file * open_exec(const char *);
2612  
2613 /* fs/dcache.c -- generic fs support functions */
2614 extern bool is_subdir(struct dentry *, struct dentry *);
2615 extern bool path_is_under(struct path *, struct path *);
2616 
2617 extern char *file_path(struct file *, char *, int);
2618 
2619 #include <linux/err.h>
2620 
2621 /* needed for stackable file system support */
2622 extern loff_t default_llseek(struct file *file, loff_t offset, int whence);
2623 
2624 extern loff_t vfs_llseek(struct file *file, loff_t offset, int whence);
2625 
2626 extern int inode_init_always(struct super_block *, struct inode *);
2627 extern void inode_init_once(struct inode *);
2628 extern void address_space_init_once(struct address_space *mapping);
2629 extern struct inode * igrab(struct inode *);
2630 extern ino_t iunique(struct super_block *, ino_t);
2631 extern int inode_needs_sync(struct inode *inode);
2632 extern int generic_delete_inode(struct inode *inode);
2633 static inline int generic_drop_inode(struct inode *inode)
2634 {
2635         return !inode->i_nlink || inode_unhashed(inode);
2636 }
2637 
2638 extern struct inode *ilookup5_nowait(struct super_block *sb,
2639                 unsigned long hashval, int (*test)(struct inode *, void *),
2640                 void *data);
2641 extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
2642                 int (*test)(struct inode *, void *), void *data);
2643 extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
2644 
2645 extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
2646 extern struct inode * iget_locked(struct super_block *, unsigned long);
2647 extern struct inode *find_inode_nowait(struct super_block *,
2648                                        unsigned long,
2649                                        int (*match)(struct inode *,
2650                                                     unsigned long, void *),
2651                                        void *data);
2652 extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
2653 extern int insert_inode_locked(struct inode *);
2654 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2655 extern void lockdep_annotate_inode_mutex_key(struct inode *inode);
2656 #else
2657 static inline void lockdep_annotate_inode_mutex_key(struct inode *inode) { };
2658 #endif
2659 extern void unlock_new_inode(struct inode *);
2660 extern unsigned int get_next_ino(void);
2661 
2662 extern void __iget(struct inode * inode);
2663 extern void iget_failed(struct inode *);
2664 extern void clear_inode(struct inode *);
2665 extern void __destroy_inode(struct inode *);
2666 extern struct inode *new_inode_pseudo(struct super_block *sb);
2667 extern struct inode *new_inode(struct super_block *sb);
2668 extern void free_inode_nonrcu(struct inode *inode);
2669 extern int should_remove_suid(struct dentry *);
2670 extern int file_remove_privs(struct file *);
2671 extern int dentry_needs_remove_privs(struct dentry *dentry);
2672 static inline int file_needs_remove_privs(struct file *file)
2673 {
2674         return dentry_needs_remove_privs(file->f_path.dentry);
2675 }
2676 
2677 extern void __insert_inode_hash(struct inode *, unsigned long hashval);
2678 static inline void insert_inode_hash(struct inode *inode)
2679 {
2680         __insert_inode_hash(inode, inode->i_ino);
2681 }
2682 
2683 extern void __remove_inode_hash(struct inode *);
2684 static inline void remove_inode_hash(struct inode *inode)
2685 {
2686         if (!inode_unhashed(inode) && !hlist_fake(&inode->i_hash))
2687                 __remove_inode_hash(inode);
2688 }
2689 
2690 extern void inode_sb_list_add(struct inode *inode);
2691 
2692 #ifdef CONFIG_BLOCK
2693 extern blk_qc_t submit_bio(int, struct bio *);
2694 extern int bdev_read_only(struct block_device *);
2695 #endif
2696 extern int set_blocksize(struct block_device *, int);
2697 extern int sb_set_blocksize(struct super_block *, int);
2698 extern int sb_min_blocksize(struct super_block *, int);
2699 
2700 extern int generic_file_mmap(struct file *, struct vm_area_struct *);
2701 extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
2702 extern ssize_t generic_write_checks(struct kiocb *, struct iov_iter *);
2703 extern ssize_t generic_file_read_iter(struct kiocb *, struct iov_iter *);
2704 extern ssize_t __generic_file_write_iter(struct kiocb *, struct iov_iter *);
2705 extern ssize_t generic_file_write_iter(struct kiocb *, struct iov_iter *);
2706 extern ssize_t generic_file_direct_write(struct kiocb *, struct iov_iter *, loff_t);
2707 extern ssize_t generic_perform_write(struct file *, struct iov_iter *, loff_t);
2708 
2709 ssize_t vfs_iter_read(struct file *file, struct iov_iter *iter, loff_t *ppos);
2710 ssize_t vfs_iter_write(struct file *file, struct iov_iter *iter, loff_t *ppos);
2711 
2712 /* fs/block_dev.c */
2713 extern ssize_t blkdev_read_iter(struct kiocb *iocb, struct iov_iter *to);
2714 extern ssize_t blkdev_write_iter(struct kiocb *iocb, struct iov_iter *from);
2715 extern int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
2716                         int datasync);
2717 extern void block_sync_page(struct page *page);
2718 
2719 /* fs/splice.c */
2720 extern ssize_t generic_file_splice_read(struct file *, loff_t *,
2721                 struct pipe_inode_info *, size_t, unsigned int);
2722 extern ssize_t default_file_splice_read(struct file *, loff_t *,
2723                 struct pipe_inode_info *, size_t, unsigned int);
2724 extern ssize_t iter_file_splice_write(struct pipe_inode_info *,
2725                 struct file *, loff_t *, size_t, unsigned int);
2726 extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
2727                 struct file *out, loff_t *, size_t len, unsigned int flags);
2728 extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
2729                 loff_t *opos, size_t len, unsigned int flags);
2730 
2731 
2732 extern void
2733 file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
2734 extern loff_t noop_llseek(struct file *file, loff_t offset, int whence);
2735 extern loff_t no_llseek(struct file *file, loff_t offset, int whence);
2736 extern loff_t vfs_setpos(struct file *file, loff_t offset, loff_t maxsize);
2737 extern loff_t generic_file_llseek(struct file *file, loff_t offset, int whence);
2738 extern loff_t generic_file_llseek_size(struct file *file, loff_t offset,
2739                 int whence, loff_t maxsize, loff_t eof);
2740 extern loff_t fixed_size_llseek(struct file *file, loff_t offset,
2741                 int whence, loff_t size);
2742 extern loff_t no_seek_end_llseek_size(struct file *, loff_t, int, loff_t);
2743 extern loff_t no_seek_end_llseek(struct file *, loff_t, int);
2744 extern int generic_file_open(struct inode * inode, struct file * filp);
2745 extern int nonseekable_open(struct inode * inode, struct file * filp);
2746 
2747 #ifdef CONFIG_BLOCK
2748 typedef void (dio_submit_t)(int rw, struct bio *bio, struct inode *inode,
2749                             loff_t file_offset);
2750 
2751 enum {
2752         /* need locking between buffered and direct access */
2753         DIO_LOCKING     = 0x01,
2754 
2755         /* filesystem does not support filling holes */
2756         DIO_SKIP_HOLES  = 0x02,
2757 
2758         /* filesystem can handle aio writes beyond i_size */
2759         DIO_ASYNC_EXTEND = 0x04,
2760 
2761         /* inode/fs/bdev does not need truncate protection */
2762         DIO_SKIP_DIO_COUNT = 0x08,
2763 };
2764 
2765 void dio_end_io(struct bio *bio, int error);
2766 
2767 ssize_t __blockdev_direct_IO(struct kiocb *iocb, struct inode *inode,
2768                              struct block_device *bdev, struct iov_iter *iter,
2769                              loff_t offset, get_block_t get_block,
2770                              dio_iodone_t end_io, dio_submit_t submit_io,
2771                              int flags);
2772 
2773 static inline ssize_t blockdev_direct_IO(struct kiocb *iocb,
2774                                          struct inode *inode,
2775                                          struct iov_iter *iter, loff_t offset,
2776                                          get_block_t get_block)
2777 {
2778         return __blockdev_direct_IO(iocb, inode, inode->i_sb->s_bdev, iter,
2779                                     offset, get_block, NULL, NULL,
2780                                     DIO_LOCKING | DIO_SKIP_HOLES);
2781 }
2782 #endif
2783 
2784 void inode_dio_wait(struct inode *inode);
2785 
2786 /*
2787  * inode_dio_begin - signal start of a direct I/O requests
2788  * @inode: inode the direct I/O happens on
2789  *
2790  * This is called once we've finished processing a direct I/O request,
2791  * and is used to wake up callers waiting for direct I/O to be quiesced.
2792  */
2793 static inline void inode_dio_begin(struct inode *inode)
2794 {
2795         atomic_inc(&inode->i_dio_count);
2796 }
2797 
2798 /*
2799  * inode_dio_end - signal finish of a direct I/O requests
2800  * @inode: inode the direct I/O happens on
2801  *
2802  * This is called once we've finished processing a direct I/O request,
2803  * and is used to wake up callers waiting for direct I/O to be quiesced.
2804  */
2805 static inline void inode_dio_end(struct inode *inode)
2806 {
2807         if (atomic_dec_and_test(&inode->i_dio_count))
2808                 wake_up_bit(&inode->i_state, __I_DIO_WAKEUP);
2809 }
2810 
2811 extern void inode_set_flags(struct inode *inode, unsigned int flags,
2812                             unsigned int mask);
2813 
2814 extern const struct file_operations generic_ro_fops;
2815 
2816 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2817 
2818 extern int readlink_copy(char __user *, int, const char *);
2819 extern int page_readlink(struct dentry *, char __user *, int);
2820 extern const char *page_get_link(struct dentry *, struct inode *,
2821                                  struct delayed_call *);
2822 extern void page_put_link(void *);
2823 extern int __page_symlink(struct inode *inode, const char *symname, int len,
2824                 int nofs);
2825 extern int page_symlink(struct inode *inode, const char *symname, int len);
2826 extern const struct inode_operations page_symlink_inode_operations;
2827 extern void kfree_link(void *);
2828 extern int generic_readlink(struct dentry *, char __user *, int);
2829 extern void generic_fillattr(struct inode *, struct kstat *);
2830 int vfs_getattr_nosec(struct path *path, struct kstat *stat);
2831 extern int vfs_getattr(struct path *, struct kstat *);
2832 void __inode_add_bytes(struct inode *inode, loff_t bytes);
2833 void inode_add_bytes(struct inode *inode, loff_t bytes);
2834 void __inode_sub_bytes(struct inode *inode, loff_t bytes);
2835 void inode_sub_bytes(struct inode *inode, loff_t bytes);
2836 loff_t inode_get_bytes(struct inode *inode);
2837 void inode_set_bytes(struct inode *inode, loff_t bytes);
2838 const char *simple_get_link(struct dentry *, struct inode *,
2839                             struct delayed_call *);
2840 extern const struct inode_operations simple_symlink_inode_operations;
2841 
2842 extern int iterate_dir(struct file *, struct dir_context *);
2843 
2844 extern int vfs_stat(const char __user *, struct kstat *);
2845 extern int vfs_lstat(const char __user *, struct kstat *);
2846 extern int vfs_fstat(unsigned int, struct kstat *);
2847 extern int vfs_fstatat(int , const char __user *, struct kstat *, int);
2848 
2849 extern int __generic_block_fiemap(struct inode *inode,
2850                                   struct fiemap_extent_info *fieinfo,
2851                                   loff_t start, loff_t len,
2852                                   get_block_t *get_block);
2853 extern int generic_block_fiemap(struct inode *inode,
2854                                 struct fiemap_extent_info *fieinfo, u64 start,
2855                                 u64 len, get_block_t *get_block);
2856 
2857 extern void get_filesystem(struct file_system_type *fs);
2858 extern void put_filesystem(struct file_system_type *fs);
2859 extern struct file_system_type *get_fs_type(const char *name);
2860 extern struct super_block *get_super(struct block_device *);
2861 extern struct super_block *get_super_thawed(struct block_device *);
2862 extern struct super_block *get_active_super(struct block_device *bdev);
2863 extern void drop_super(struct super_block *sb);
2864 extern void iterate_supers(void (*)(struct super_block *, void *), void *);
2865 extern void iterate_supers_type(struct file_system_type *,
2866                                 void (*)(struct super_block *, void *), void *);
2867 
2868 extern int dcache_dir_open(struct inode *, struct file *);
2869 extern int dcache_dir_close(struct inode *, struct file *);
2870 extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
2871 extern int dcache_readdir(struct file *, struct dir_context *);
2872 extern int simple_setattr(struct dentry *, struct iattr *);
2873 extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2874 extern int simple_statfs(struct dentry *, struct kstatfs *);
2875 extern int simple_open(struct inode *inode, struct file *file);
2876 extern int simple_link(struct dentry *, struct inode *, struct dentry *);
2877 extern int simple_unlink(struct inode *, struct dentry *);
2878 extern int simple_rmdir(struct inode *, struct dentry *);
2879 extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
2880 extern int noop_fsync(struct file *, loff_t, loff_t, int);
2881 extern int simple_empty(struct dentry *);
2882 extern int simple_readpage(struct file *file, struct page *page);
2883 extern int simple_write_begin(struct file *file, struct address_space *mapping,
2884                         loff_t pos, unsigned len, unsigned flags,
2885                         struct page **pagep, void **fsdata);
2886 extern int simple_write_end(struct file *file, struct address_space *mapping,
2887                         loff_t pos, unsigned len, unsigned copied,
2888                         struct page *page, void *fsdata);
2889 extern int always_delete_dentry(const struct dentry *);
2890 extern struct inode *alloc_anon_inode(struct super_block *);
2891 extern int simple_nosetlease(struct file *, long, struct file_lock **, void **);
2892 extern const struct dentry_operations simple_dentry_operations;
2893 
2894 extern struct dentry *simple_lookup(struct inode *, struct dentry *, unsigned int flags);
2895 extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
2896 extern const struct file_operations simple_dir_operations;
2897 extern const struct inode_operations simple_dir_inode_operations;
2898 extern void make_empty_dir_inode(struct inode *inode);
2899 extern bool is_empty_dir_inode(struct inode *inode);
2900 struct tree_descr { char *name; const struct file_operations *ops; int mode; };
2901 struct dentry *d_alloc_name(struct dentry *, const char *);
2902 extern int simple_fill_super(struct super_block *, unsigned long, struct tree_descr *);
2903 extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
2904 extern void simple_release_fs(struct vfsmount **mount, int *count);
2905 
2906 extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
2907                         loff_t *ppos, const void *from, size_t available);
2908 extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
2909                 const void __user *from, size_t count);
2910 
2911 extern int __generic_file_fsync(struct file *, loff_t, loff_t, int);
2912 extern int generic_file_fsync(struct file *, loff_t, loff_t, int);
2913 
2914 extern int generic_check_addressable(unsigned, u64);
2915 
2916 #ifdef CONFIG_MIGRATION
2917 extern int buffer_migrate_page(struct address_space *,
2918                                 struct page *, struct page *,
2919                                 enum migrate_mode);
2920 #else
2921 #define buffer_migrate_page NULL
2922 #endif
2923 
2924 extern int inode_change_ok(const struct inode *, struct iattr *);
2925 extern int inode_newsize_ok(const struct inode *, loff_t offset);
2926 extern void setattr_copy(struct inode *inode, const struct iattr *attr);
2927 
2928 extern int file_update_time(struct file *file);
2929 
2930 extern int generic_show_options(struct seq_file *m, struct dentry *root);
2931 extern void save_mount_options(struct super_block *sb, char *options);
2932 extern void replace_mount_options(struct super_block *sb, char *options);
2933 
2934 static inline bool io_is_direct(struct file *filp)
2935 {
2936         return (filp->f_flags & O_DIRECT) || IS_DAX(filp->f_mapping->host);
2937 }
2938 
2939 static inline int iocb_flags(struct file *file)
2940 {
2941         int res = 0;
2942         if (file->f_flags & O_APPEND)
2943                 res |= IOCB_APPEND;
2944         if (io_is_direct(file))
2945                 res |= IOCB_DIRECT;
2946         return res;
2947 }
2948 
2949 static inline ino_t parent_ino(struct dentry *dentry)
2950 {
2951         ino_t res;
2952 
2953         /*
2954          * Don't strictly need d_lock here? If the parent ino could change
2955          * then surely we'd have a deeper race in the caller?
2956          */
2957         spin_lock(&dentry->d_lock);
2958         res = dentry->d_parent->d_inode->i_ino;
2959         spin_unlock(&dentry->d_lock);
2960         return res;
2961 }
2962 
2963 /* Transaction based IO helpers */
2964 
2965 /*
2966  * An argresp is stored in an allocated page and holds the
2967  * size of the argument or response, along with its content
2968  */
2969 struct simple_transaction_argresp {
2970         ssize_t size;
2971         char data[0];
2972 };
2973 
2974 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2975 
2976 char *simple_transaction_get(struct file *file, const char __user *buf,
2977                                 size_t size);
2978 ssize_t simple_transaction_read(struct file *file, char __user *buf,
2979                                 size_t size, loff_t *pos);
2980 int simple_transaction_release(struct inode *inode, struct file *file);
2981 
2982 void simple_transaction_set(struct file *file, size_t n);
2983 
2984 /*
2985  * simple attribute files
2986  *
2987  * These attributes behave similar to those in sysfs:
2988  *
2989  * Writing to an attribute immediately sets a value, an open file can be
2990  * written to multiple times.
2991  *
2992  * Reading from an attribute creates a buffer from the value that might get
2993  * read with multiple read calls. When the attribute has been read
2994  * completely, no further read calls are possible until the file is opened
2995  * again.
2996  *
2997  * All attributes contain a text representation of a numeric value
2998  * that are accessed with the get() and set() functions.
2999  */
3000 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt)            \
3001 static int __fops ## _open(struct inode *inode, struct file *file)      \
3002 {                                                                       \
3003         __simple_attr_check_format(__fmt, 0ull);                        \
3004         return simple_attr_open(inode, file, __get, __set, __fmt);      \
3005 }                                                                       \
3006 static const struct file_operations __fops = {                          \
3007         .owner   = THIS_MODULE,                                         \
3008         .open    = __fops ## _open,                                     \
3009         .release = simple_attr_release,                                 \
3010         .read    = simple_attr_read,                                    \
3011         .write   = simple_attr_write,                                   \
3012         .llseek  = generic_file_llseek,                                 \
3013 }
3014 
3015 static inline __printf(1, 2)
3016 void __simple_attr_check_format(const char *fmt, ...)
3017 {
3018         /* don't do anything, just let the compiler check the arguments; */
3019 }
3020 
3021 int simple_attr_open(struct inode *inode, struct file *file,
3022                      int (*get)(void *, u64 *), int (*set)(void *, u64),
3023                      const char *fmt);
3024 int simple_attr_release(struct inode *inode, struct file *file);
3025 ssize_t simple_attr_read(struct file *file, char __user *buf,
3026                          size_t len, loff_t *ppos);
3027 ssize_t simple_attr_write(struct file *file, const char __user *buf,
3028                           size_t len, loff_t *ppos);
3029 
3030 struct ctl_table;
3031 int proc_nr_files(struct ctl_table *table, int write,
3032                   void __user *buffer, size_t *lenp, loff_t *ppos);
3033 int proc_nr_dentry(struct ctl_table *table, int write,
3034                   void __user *buffer, size_t *lenp, loff_t *ppos);
3035 int proc_nr_inodes(struct ctl_table *table, int write,
3036                    void __user *buffer, size_t *lenp, loff_t *ppos);
3037 int __init get_filesystem_list(char *buf);
3038 
3039 #define __FMODE_EXEC            ((__force int) FMODE_EXEC)
3040 #define __FMODE_NONOTIFY        ((__force int) FMODE_NONOTIFY)
3041 
3042 #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
3043 #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
3044                                             (flag & __FMODE_NONOTIFY)))
3045 
3046 static inline bool is_sxid(umode_t mode)
3047 {
3048         return (mode & S_ISUID) || ((mode & S_ISGID) && (mode & S_IXGRP));
3049 }
3050 
3051 static inline int check_sticky(struct inode *dir, struct inode *inode)
3052 {
3053         if (!(dir->i_mode & S_ISVTX))
3054                 return 0;
3055 
3056         return __check_sticky(dir, inode);
3057 }
3058 
3059 static inline void inode_has_no_xattr(struct inode *inode)
3060 {
3061         if (!is_sxid(inode->i_mode) && (inode->i_sb->s_flags & MS_NOSEC))
3062                 inode->i_flags |= S_NOSEC;
3063 }
3064 
3065 static inline bool is_root_inode(struct inode *inode)
3066 {
3067         return inode == inode->i_sb->s_root->d_inode;
3068 }
3069 
3070 static inline bool dir_emit(struct dir_context *ctx,
3071                             const char *name, int namelen,
3072                             u64 ino, unsigned type)
3073 {
3074         return ctx->actor(ctx, name, namelen, ctx->pos, ino, type) == 0;
3075 }
3076 static inline bool dir_emit_dot(struct file *file, struct dir_context *ctx)
3077 {
3078         return ctx->actor(ctx, ".", 1, ctx->pos,
3079                           file->f_path.dentry->d_inode->i_ino, DT_DIR) == 0;
3080 }
3081 static inline bool dir_emit_dotdot(struct file *file, struct dir_context *ctx)
3082 {
3083         return ctx->actor(ctx, "..", 2, ctx->pos,
3084                           parent_ino(file->f_path.dentry), DT_DIR) == 0;
3085 }
3086 static inline bool dir_emit_dots(struct file *file, struct dir_context *ctx)
3087 {
3088         if (ctx->pos == 0) {
3089                 if (!dir_emit_dot(file, ctx))
3090                         return false;
3091                 ctx->pos = 1;
3092         }
3093         if (ctx->pos == 1) {
3094                 if (!dir_emit_dotdot(file, ctx))
3095                         return false;
3096                 ctx->pos = 2;
3097         }
3098         return true;
3099 }
3100 static inline bool dir_relax(struct inode *inode)
3101 {
3102         inode_unlock(inode);
3103         inode_lock(inode);
3104         return !IS_DEADDIR(inode);
3105 }
3106 
3107 extern bool path_noexec(const struct path *path);
3108 extern void inode_nohighmem(struct inode *inode);
3109 
3110 #endif /* _LINUX_FS_H */
3111 

This page was automatically generated by LXR 0.3.1 (source).  •  Linux is a registered trademark of Linus Torvalds  •  Contact us