~ [ source navigation ] ~ [ diff markup ] ~ [ identifier search ] ~

TOMOYO Linux Cross Reference
Linux/include/linux/fs.h

Version: ~ [ linux-5.9-rc6 ] ~ [ linux-5.8.10 ] ~ [ linux-5.7.19 ] ~ [ linux-5.6.19 ] ~ [ linux-5.5.19 ] ~ [ linux-5.4.66 ] ~ [ linux-5.3.18 ] ~ [ linux-5.2.21 ] ~ [ linux-5.1.21 ] ~ [ linux-5.0.21 ] ~ [ linux-4.20.17 ] ~ [ linux-4.19.146 ] ~ [ linux-4.18.20 ] ~ [ linux-4.17.19 ] ~ [ linux-4.16.18 ] ~ [ linux-4.15.18 ] ~ [ linux-4.14.198 ] ~ [ linux-4.13.16 ] ~ [ linux-4.12.14 ] ~ [ linux-4.11.12 ] ~ [ linux-4.10.17 ] ~ [ linux-4.9.236 ] ~ [ linux-4.8.17 ] ~ [ linux-4.7.10 ] ~ [ linux-4.6.7 ] ~ [ linux-4.5.7 ] ~ [ linux-4.4.236 ] ~ [ linux-4.3.6 ] ~ [ linux-4.2.8 ] ~ [ linux-4.1.52 ] ~ [ linux-4.0.9 ] ~ [ linux-3.19.8 ] ~ [ linux-3.18.140 ] ~ [ linux-3.17.8 ] ~ [ linux-3.16.85 ] ~ [ linux-3.15.10 ] ~ [ linux-3.14.79 ] ~ [ linux-3.13.11 ] ~ [ linux-3.12.74 ] ~ [ linux-3.11.10 ] ~ [ linux-3.10.108 ] ~ [ linux-2.6.32.71 ] ~ [ linux-2.6.0 ] ~ [ linux-2.4.37.11 ] ~ [ unix-v6-master ] ~ [ ccs-tools-1.8.5 ] ~ [ policy-sample ] ~
Architecture: ~ [ i386 ] ~ [ alpha ] ~ [ m68k ] ~ [ mips ] ~ [ ppc ] ~ [ sparc ] ~ [ sparc64 ] ~

  1 #ifndef _LINUX_FS_H
  2 #define _LINUX_FS_H
  3 
  4 
  5 #include <linux/linkage.h>
  6 #include <linux/wait.h>
  7 #include <linux/kdev_t.h>
  8 #include <linux/dcache.h>
  9 #include <linux/path.h>
 10 #include <linux/stat.h>
 11 #include <linux/cache.h>
 12 #include <linux/list.h>
 13 #include <linux/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/capability.h>
 21 #include <linux/semaphore.h>
 22 #include <linux/fiemap.h>
 23 #include <linux/rculist_bl.h>
 24 #include <linux/atomic.h>
 25 #include <linux/shrinker.h>
 26 #include <linux/migrate_mode.h>
 27 #include <linux/uidgid.h>
 28 #include <linux/lockdep.h>
 29 #include <linux/percpu-rwsem.h>
 30 #include <linux/blk_types.h>
 31 
 32 #include <asm/byteorder.h>
 33 #include <uapi/linux/fs.h>
 34 
 35 struct export_operations;
 36 struct hd_geometry;
 37 struct iovec;
 38 struct nameidata;
 39 struct kiocb;
 40 struct kobject;
 41 struct pipe_inode_info;
 42 struct poll_table_struct;
 43 struct kstatfs;
 44 struct vm_area_struct;
 45 struct vfsmount;
 46 struct cred;
 47 struct swap_info_struct;
 48 struct seq_file;
 49 
 50 extern void __init inode_init(void);
 51 extern void __init inode_init_early(void);
 52 extern void __init files_init(unsigned long);
 53 
 54 extern struct files_stat_struct files_stat;
 55 extern unsigned long get_max_files(void);
 56 extern int sysctl_nr_open;
 57 extern struct inodes_stat_t inodes_stat;
 58 extern int leases_enable, lease_break_time;
 59 extern int sysctl_protected_symlinks;
 60 extern int sysctl_protected_hardlinks;
 61 
 62 struct buffer_head;
 63 typedef int (get_block_t)(struct inode *inode, sector_t iblock,
 64                         struct buffer_head *bh_result, int create);
 65 typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
 66                         ssize_t bytes, void *private, int ret,
 67                         bool is_async);
 68 
 69 #define MAY_EXEC                0x00000001
 70 #define MAY_WRITE               0x00000002
 71 #define MAY_READ                0x00000004
 72 #define MAY_APPEND              0x00000008
 73 #define MAY_ACCESS              0x00000010
 74 #define MAY_OPEN                0x00000020
 75 #define MAY_CHDIR               0x00000040
 76 /* called from RCU mode, don't block */
 77 #define MAY_NOT_BLOCK           0x00000080
 78 
 79 /*
 80  * flags in file.f_mode.  Note that FMODE_READ and FMODE_WRITE must correspond
 81  * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
 82  */
 83 
 84 /* file is open for reading */
 85 #define FMODE_READ              ((__force fmode_t)0x1)
 86 /* file is open for writing */
 87 #define FMODE_WRITE             ((__force fmode_t)0x2)
 88 /* file is seekable */
 89 #define FMODE_LSEEK             ((__force fmode_t)0x4)
 90 /* file can be accessed using pread */
 91 #define FMODE_PREAD             ((__force fmode_t)0x8)
 92 /* file can be accessed using pwrite */
 93 #define FMODE_PWRITE            ((__force fmode_t)0x10)
 94 /* File is opened for execution with sys_execve / sys_uselib */
 95 #define FMODE_EXEC              ((__force fmode_t)0x20)
 96 /* File is opened with O_NDELAY (only set for block devices) */
 97 #define FMODE_NDELAY            ((__force fmode_t)0x40)
 98 /* File is opened with O_EXCL (only set for block devices) */
 99 #define FMODE_EXCL              ((__force fmode_t)0x80)
100 /* File is opened using open(.., 3, ..) and is writeable only for ioctls
101    (specialy hack for floppy.c) */
102 #define FMODE_WRITE_IOCTL       ((__force fmode_t)0x100)
103 /* 32bit hashes as llseek() offset (for directories) */
104 #define FMODE_32BITHASH         ((__force fmode_t)0x200)
105 /* 64bit hashes as llseek() offset (for directories) */
106 #define FMODE_64BITHASH         ((__force fmode_t)0x400)
107 
108 /*
109  * Don't update ctime and mtime.
110  *
111  * Currently a special hack for the XFS open_by_handle ioctl, but we'll
112  * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
113  */
114 #define FMODE_NOCMTIME          ((__force fmode_t)0x800)
115 
116 /* Expect random access pattern */
117 #define FMODE_RANDOM            ((__force fmode_t)0x1000)
118 
119 /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
120 #define FMODE_UNSIGNED_OFFSET   ((__force fmode_t)0x2000)
121 
122 /* File is opened with O_PATH; almost nothing can be done with it */
123 #define FMODE_PATH              ((__force fmode_t)0x4000)
124 
125 /* File was opened by fanotify and shouldn't generate fanotify events */
126 #define FMODE_NONOTIFY          ((__force fmode_t)0x1000000)
127 
128 /*
129  * Flag for rw_copy_check_uvector and compat_rw_copy_check_uvector
130  * that indicates that they should check the contents of the iovec are
131  * valid, but not check the memory that the iovec elements
132  * points too.
133  */
134 #define CHECK_IOVEC_ONLY -1
135 
136 /*
137  * The below are the various read and write types that we support. Some of
138  * them include behavioral modifiers that send information down to the
139  * block layer and IO scheduler. Terminology:
140  *
141  *      The block layer uses device plugging to defer IO a little bit, in
142  *      the hope that we will see more IO very shortly. This increases
143  *      coalescing of adjacent IO and thus reduces the number of IOs we
144  *      have to send to the device. It also allows for better queuing,
145  *      if the IO isn't mergeable. If the caller is going to be waiting
146  *      for the IO, then he must ensure that the device is unplugged so
147  *      that the IO is dispatched to the driver.
148  *
149  *      All IO is handled async in Linux. This is fine for background
150  *      writes, but for reads or writes that someone waits for completion
151  *      on, we want to notify the block layer and IO scheduler so that they
152  *      know about it. That allows them to make better scheduling
153  *      decisions. So when the below references 'sync' and 'async', it
154  *      is referencing this priority hint.
155  *
156  * With that in mind, the available types are:
157  *
158  * READ                 A normal read operation. Device will be plugged.
159  * READ_SYNC            A synchronous read. Device is not plugged, caller can
160  *                      immediately wait on this read without caring about
161  *                      unplugging.
162  * READA                Used for read-ahead operations. Lower priority, and the
163  *                      block layer could (in theory) choose to ignore this
164  *                      request if it runs into resource problems.
165  * WRITE                A normal async write. Device will be plugged.
166  * WRITE_SYNC           Synchronous write. Identical to WRITE, but passes down
167  *                      the hint that someone will be waiting on this IO
168  *                      shortly. The write equivalent of READ_SYNC.
169  * WRITE_ODIRECT        Special case write for O_DIRECT only.
170  * WRITE_FLUSH          Like WRITE_SYNC but with preceding cache flush.
171  * WRITE_FUA            Like WRITE_SYNC but data is guaranteed to be on
172  *                      non-volatile media on completion.
173  * WRITE_FLUSH_FUA      Combination of WRITE_FLUSH and FUA. The IO is preceded
174  *                      by a cache flush and data is guaranteed to be on
175  *                      non-volatile media on completion.
176  *
177  */
178 #define RW_MASK                 REQ_WRITE
179 #define RWA_MASK                REQ_RAHEAD
180 
181 #define READ                    0
182 #define WRITE                   RW_MASK
183 #define READA                   RWA_MASK
184 #define KERNEL_READ             (READ|REQ_KERNEL)
185 #define KERNEL_WRITE            (WRITE|REQ_KERNEL)
186 
187 #define READ_SYNC               (READ | REQ_SYNC)
188 #define WRITE_SYNC              (WRITE | REQ_SYNC | REQ_NOIDLE)
189 #define WRITE_ODIRECT           (WRITE | REQ_SYNC)
190 #define WRITE_FLUSH             (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
191 #define WRITE_FUA               (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
192 #define WRITE_FLUSH_FUA         (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
193 
194 /*
195  * Attribute flags.  These should be or-ed together to figure out what
196  * has been changed!
197  */
198 #define ATTR_MODE       (1 << 0)
199 #define ATTR_UID        (1 << 1)
200 #define ATTR_GID        (1 << 2)
201 #define ATTR_SIZE       (1 << 3)
202 #define ATTR_ATIME      (1 << 4)
203 #define ATTR_MTIME      (1 << 5)
204 #define ATTR_CTIME      (1 << 6)
205 #define ATTR_ATIME_SET  (1 << 7)
206 #define ATTR_MTIME_SET  (1 << 8)
207 #define ATTR_FORCE      (1 << 9) /* Not a change, but a change it */
208 #define ATTR_ATTR_FLAG  (1 << 10)
209 #define ATTR_KILL_SUID  (1 << 11)
210 #define ATTR_KILL_SGID  (1 << 12)
211 #define ATTR_FILE       (1 << 13)
212 #define ATTR_KILL_PRIV  (1 << 14)
213 #define ATTR_OPEN       (1 << 15) /* Truncating from open(O_TRUNC) */
214 #define ATTR_TIMES_SET  (1 << 16)
215 
216 /*
217  * This is the Inode Attributes structure, used for notify_change().  It
218  * uses the above definitions as flags, to know which values have changed.
219  * Also, in this manner, a Filesystem can look at only the values it cares
220  * about.  Basically, these are the attributes that the VFS layer can
221  * request to change from the FS layer.
222  *
223  * Derek Atkins <warlord@MIT.EDU> 94-10-20
224  */
225 struct iattr {
226         unsigned int    ia_valid;
227         umode_t         ia_mode;
228         kuid_t          ia_uid;
229         kgid_t          ia_gid;
230         loff_t          ia_size;
231         struct timespec ia_atime;
232         struct timespec ia_mtime;
233         struct timespec ia_ctime;
234 
235         /*
236          * Not an attribute, but an auxiliary info for filesystems wanting to
237          * implement an ftruncate() like method.  NOTE: filesystem should
238          * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
239          */
240         struct file     *ia_file;
241 };
242 
243 /*
244  * Includes for diskquotas.
245  */
246 #include <linux/quota.h>
247 
248 /** 
249  * enum positive_aop_returns - aop return codes with specific semantics
250  *
251  * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
252  *                          completed, that the page is still locked, and
253  *                          should be considered active.  The VM uses this hint
254  *                          to return the page to the active list -- it won't
255  *                          be a candidate for writeback again in the near
256  *                          future.  Other callers must be careful to unlock
257  *                          the page if they get this return.  Returned by
258  *                          writepage(); 
259  *
260  * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
261  *                      unlocked it and the page might have been truncated.
262  *                      The caller should back up to acquiring a new page and
263  *                      trying again.  The aop will be taking reasonable
264  *                      precautions not to livelock.  If the caller held a page
265  *                      reference, it should drop it before retrying.  Returned
266  *                      by readpage().
267  *
268  * address_space_operation functions return these large constants to indicate
269  * special semantics to the caller.  These are much larger than the bytes in a
270  * page to allow for functions that return the number of bytes operated on in a
271  * given page.
272  */
273 
274 enum positive_aop_returns {
275         AOP_WRITEPAGE_ACTIVATE  = 0x80000,
276         AOP_TRUNCATED_PAGE      = 0x80001,
277 };
278 
279 #define AOP_FLAG_UNINTERRUPTIBLE        0x0001 /* will not do a short write */
280 #define AOP_FLAG_CONT_EXPAND            0x0002 /* called from cont_expand */
281 #define AOP_FLAG_NOFS                   0x0004 /* used by filesystem to direct
282                                                 * helper code (eg buffer layer)
283                                                 * to clear GFP_FS from alloc */
284 
285 /*
286  * oh the beauties of C type declarations.
287  */
288 struct page;
289 struct address_space;
290 struct writeback_control;
291 
292 struct iov_iter {
293         const struct iovec *iov;
294         unsigned long nr_segs;
295         size_t iov_offset;
296         size_t count;
297 };
298 
299 size_t iov_iter_copy_from_user_atomic(struct page *page,
300                 struct iov_iter *i, unsigned long offset, size_t bytes);
301 size_t iov_iter_copy_from_user(struct page *page,
302                 struct iov_iter *i, unsigned long offset, size_t bytes);
303 void iov_iter_advance(struct iov_iter *i, size_t bytes);
304 int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes);
305 size_t iov_iter_single_seg_count(const struct iov_iter *i);
306 
307 static inline void iov_iter_init(struct iov_iter *i,
308                         const struct iovec *iov, unsigned long nr_segs,
309                         size_t count, size_t written)
310 {
311         i->iov = iov;
312         i->nr_segs = nr_segs;
313         i->iov_offset = 0;
314         i->count = count + written;
315 
316         iov_iter_advance(i, written);
317 }
318 
319 static inline size_t iov_iter_count(struct iov_iter *i)
320 {
321         return i->count;
322 }
323 
324 /*
325  * "descriptor" for what we're up to with a read.
326  * This allows us to use the same read code yet
327  * have multiple different users of the data that
328  * we read from a file.
329  *
330  * The simplest case just copies the data to user
331  * mode.
332  */
333 typedef struct {
334         size_t written;
335         size_t count;
336         union {
337                 char __user *buf;
338                 void *data;
339         } arg;
340         int error;
341 } read_descriptor_t;
342 
343 typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
344                 unsigned long, unsigned long);
345 
346 struct address_space_operations {
347         int (*writepage)(struct page *page, struct writeback_control *wbc);
348         int (*readpage)(struct file *, struct page *);
349 
350         /* Write back some dirty pages from this mapping. */
351         int (*writepages)(struct address_space *, struct writeback_control *);
352 
353         /* Set a page dirty.  Return true if this dirtied it */
354         int (*set_page_dirty)(struct page *page);
355 
356         int (*readpages)(struct file *filp, struct address_space *mapping,
357                         struct list_head *pages, unsigned nr_pages);
358 
359         int (*write_begin)(struct file *, struct address_space *mapping,
360                                 loff_t pos, unsigned len, unsigned flags,
361                                 struct page **pagep, void **fsdata);
362         int (*write_end)(struct file *, struct address_space *mapping,
363                                 loff_t pos, unsigned len, unsigned copied,
364                                 struct page *page, void *fsdata);
365 
366         /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
367         sector_t (*bmap)(struct address_space *, sector_t);
368         void (*invalidatepage) (struct page *, unsigned long);
369         int (*releasepage) (struct page *, gfp_t);
370         void (*freepage)(struct page *);
371         ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
372                         loff_t offset, unsigned long nr_segs);
373         int (*get_xip_mem)(struct address_space *, pgoff_t, int,
374                                                 void **, unsigned long *);
375         /*
376          * migrate the contents of a page to the specified target. If sync
377          * is false, it must not block.
378          */
379         int (*migratepage) (struct address_space *,
380                         struct page *, struct page *, enum migrate_mode);
381         int (*launder_page) (struct page *);
382         int (*is_partially_uptodate) (struct page *, read_descriptor_t *,
383                                         unsigned long);
384         int (*error_remove_page)(struct address_space *, struct page *);
385 
386         /* swapfile support */
387         int (*swap_activate)(struct swap_info_struct *sis, struct file *file,
388                                 sector_t *span);
389         void (*swap_deactivate)(struct file *file);
390 };
391 
392 extern const struct address_space_operations empty_aops;
393 
394 /*
395  * pagecache_write_begin/pagecache_write_end must be used by general code
396  * to write into the pagecache.
397  */
398 int pagecache_write_begin(struct file *, struct address_space *mapping,
399                                 loff_t pos, unsigned len, unsigned flags,
400                                 struct page **pagep, void **fsdata);
401 
402 int pagecache_write_end(struct file *, struct address_space *mapping,
403                                 loff_t pos, unsigned len, unsigned copied,
404                                 struct page *page, void *fsdata);
405 
406 struct backing_dev_info;
407 struct address_space {
408         struct inode            *host;          /* owner: inode, block_device */
409         struct radix_tree_root  page_tree;      /* radix tree of all pages */
410         spinlock_t              tree_lock;      /* and lock protecting it */
411         unsigned int            i_mmap_writable;/* count VM_SHARED mappings */
412         struct rb_root          i_mmap;         /* tree of private and shared mappings */
413         struct list_head        i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
414         struct mutex            i_mmap_mutex;   /* protect tree, count, list */
415         /* Protected by tree_lock together with the radix tree */
416         unsigned long           nrpages;        /* number of total pages */
417         pgoff_t                 writeback_index;/* writeback starts here */
418         const struct address_space_operations *a_ops;   /* methods */
419         unsigned long           flags;          /* error bits/gfp mask */
420         struct backing_dev_info *backing_dev_info; /* device readahead, etc */
421         spinlock_t              private_lock;   /* for use by the address_space */
422         struct list_head        private_list;   /* ditto */
423         void                    *private_data;  /* ditto */
424 } __attribute__((aligned(sizeof(long))));
425         /*
426          * On most architectures that alignment is already the case; but
427          * must be enforced here for CRIS, to let the least significant bit
428          * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
429          */
430 struct request_queue;
431 
432 struct block_device {
433         dev_t                   bd_dev;  /* not a kdev_t - it's a search key */
434         int                     bd_openers;
435         struct inode *          bd_inode;       /* will die */
436         struct super_block *    bd_super;
437         struct mutex            bd_mutex;       /* open/close mutex */
438         struct list_head        bd_inodes;
439         void *                  bd_claiming;
440         void *                  bd_holder;
441         int                     bd_holders;
442         bool                    bd_write_holder;
443 #ifdef CONFIG_SYSFS
444         struct list_head        bd_holder_disks;
445 #endif
446         struct block_device *   bd_contains;
447         unsigned                bd_block_size;
448         struct hd_struct *      bd_part;
449         /* number of times partitions within this device have been opened. */
450         unsigned                bd_part_count;
451         int                     bd_invalidated;
452         struct gendisk *        bd_disk;
453         struct request_queue *  bd_queue;
454         struct list_head        bd_list;
455         /*
456          * Private data.  You must have bd_claim'ed the block_device
457          * to use this.  NOTE:  bd_claim allows an owner to claim
458          * the same device multiple times, the owner must take special
459          * care to not mess up bd_private for that case.
460          */
461         unsigned long           bd_private;
462 
463         /* The counter of freeze processes */
464         int                     bd_fsfreeze_count;
465         /* Mutex for freeze */
466         struct mutex            bd_fsfreeze_mutex;
467 };
468 
469 /*
470  * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
471  * radix trees
472  */
473 #define PAGECACHE_TAG_DIRTY     0
474 #define PAGECACHE_TAG_WRITEBACK 1
475 #define PAGECACHE_TAG_TOWRITE   2
476 
477 int mapping_tagged(struct address_space *mapping, int tag);
478 
479 /*
480  * Might pages of this file be mapped into userspace?
481  */
482 static inline int mapping_mapped(struct address_space *mapping)
483 {
484         return  !RB_EMPTY_ROOT(&mapping->i_mmap) ||
485                 !list_empty(&mapping->i_mmap_nonlinear);
486 }
487 
488 /*
489  * Might pages of this file have been modified in userspace?
490  * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
491  * marks vma as VM_SHARED if it is shared, and the file was opened for
492  * writing i.e. vma may be mprotected writable even if now readonly.
493  */
494 static inline int mapping_writably_mapped(struct address_space *mapping)
495 {
496         return mapping->i_mmap_writable != 0;
497 }
498 
499 /*
500  * Use sequence counter to get consistent i_size on 32-bit processors.
501  */
502 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
503 #include <linux/seqlock.h>
504 #define __NEED_I_SIZE_ORDERED
505 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
506 #else
507 #define i_size_ordered_init(inode) do { } while (0)
508 #endif
509 
510 struct posix_acl;
511 #define ACL_NOT_CACHED ((void *)(-1))
512 
513 #define IOP_FASTPERM    0x0001
514 #define IOP_LOOKUP      0x0002
515 #define IOP_NOFOLLOW    0x0004
516 
517 /*
518  * Keep mostly read-only and often accessed (especially for
519  * the RCU path lookup and 'stat' data) fields at the beginning
520  * of the 'struct inode'
521  */
522 struct inode {
523         umode_t                 i_mode;
524         unsigned short          i_opflags;
525         kuid_t                  i_uid;
526         kgid_t                  i_gid;
527         unsigned int            i_flags;
528 
529 #ifdef CONFIG_FS_POSIX_ACL
530         struct posix_acl        *i_acl;
531         struct posix_acl        *i_default_acl;
532 #endif
533 
534         const struct inode_operations   *i_op;
535         struct super_block      *i_sb;
536         struct address_space    *i_mapping;
537 
538 #ifdef CONFIG_SECURITY
539         void                    *i_security;
540 #endif
541 
542         /* Stat data, not accessed from path walking */
543         unsigned long           i_ino;
544         /*
545          * Filesystems may only read i_nlink directly.  They shall use the
546          * following functions for modification:
547          *
548          *    (set|clear|inc|drop)_nlink
549          *    inode_(inc|dec)_link_count
550          */
551         union {
552                 const unsigned int i_nlink;
553                 unsigned int __i_nlink;
554         };
555         dev_t                   i_rdev;
556         loff_t                  i_size;
557         struct timespec         i_atime;
558         struct timespec         i_mtime;
559         struct timespec         i_ctime;
560         spinlock_t              i_lock; /* i_blocks, i_bytes, maybe i_size */
561         unsigned short          i_bytes;
562         unsigned int            i_blkbits;
563         blkcnt_t                i_blocks;
564 
565 #ifdef __NEED_I_SIZE_ORDERED
566         seqcount_t              i_size_seqcount;
567 #endif
568 
569         /* Misc */
570         unsigned long           i_state;
571         struct mutex            i_mutex;
572 
573         unsigned long           dirtied_when;   /* jiffies of first dirtying */
574 
575         struct hlist_node       i_hash;
576         struct list_head        i_wb_list;      /* backing dev IO list */
577         struct list_head        i_lru;          /* inode LRU list */
578         struct list_head        i_sb_list;
579         union {
580                 struct hlist_head       i_dentry;
581                 struct rcu_head         i_rcu;
582         };
583         u64                     i_version;
584         atomic_t                i_count;
585         atomic_t                i_dio_count;
586         atomic_t                i_writecount;
587         const struct file_operations    *i_fop; /* former ->i_op->default_file_ops */
588         struct file_lock        *i_flock;
589         struct address_space    i_data;
590 #ifdef CONFIG_QUOTA
591         struct dquot            *i_dquot[MAXQUOTAS];
592 #endif
593         struct list_head        i_devices;
594         union {
595                 struct pipe_inode_info  *i_pipe;
596                 struct block_device     *i_bdev;
597                 struct cdev             *i_cdev;
598         };
599 
600         __u32                   i_generation;
601 
602 #ifdef CONFIG_FSNOTIFY
603         __u32                   i_fsnotify_mask; /* all events this inode cares about */
604         struct hlist_head       i_fsnotify_marks;
605 #endif
606 
607 #ifdef CONFIG_IMA
608         atomic_t                i_readcount; /* struct files open RO */
609 #endif
610         void                    *i_private; /* fs or device private pointer */
611 };
612 
613 static inline int inode_unhashed(struct inode *inode)
614 {
615         return hlist_unhashed(&inode->i_hash);
616 }
617 
618 /*
619  * inode->i_mutex nesting subclasses for the lock validator:
620  *
621  * 0: the object of the current VFS operation
622  * 1: parent
623  * 2: child/target
624  * 3: quota file
625  *
626  * The locking order between these classes is
627  * parent -> child -> normal -> xattr -> quota
628  */
629 enum inode_i_mutex_lock_class
630 {
631         I_MUTEX_NORMAL,
632         I_MUTEX_PARENT,
633         I_MUTEX_CHILD,
634         I_MUTEX_XATTR,
635         I_MUTEX_QUOTA
636 };
637 
638 /*
639  * NOTE: in a 32bit arch with a preemptable kernel and
640  * an UP compile the i_size_read/write must be atomic
641  * with respect to the local cpu (unlike with preempt disabled),
642  * but they don't need to be atomic with respect to other cpus like in
643  * true SMP (so they need either to either locally disable irq around
644  * the read or for example on x86 they can be still implemented as a
645  * cmpxchg8b without the need of the lock prefix). For SMP compiles
646  * and 64bit archs it makes no difference if preempt is enabled or not.
647  */
648 static inline loff_t i_size_read(const struct inode *inode)
649 {
650 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
651         loff_t i_size;
652         unsigned int seq;
653 
654         do {
655                 seq = read_seqcount_begin(&inode->i_size_seqcount);
656                 i_size = inode->i_size;
657         } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
658         return i_size;
659 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
660         loff_t i_size;
661 
662         preempt_disable();
663         i_size = inode->i_size;
664         preempt_enable();
665         return i_size;
666 #else
667         return inode->i_size;
668 #endif
669 }
670 
671 /*
672  * NOTE: unlike i_size_read(), i_size_write() does need locking around it
673  * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
674  * can be lost, resulting in subsequent i_size_read() calls spinning forever.
675  */
676 static inline void i_size_write(struct inode *inode, loff_t i_size)
677 {
678 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
679         preempt_disable();
680         write_seqcount_begin(&inode->i_size_seqcount);
681         inode->i_size = i_size;
682         write_seqcount_end(&inode->i_size_seqcount);
683         preempt_enable();
684 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
685         preempt_disable();
686         inode->i_size = i_size;
687         preempt_enable();
688 #else
689         inode->i_size = i_size;
690 #endif
691 }
692 
693 /* Helper functions so that in most cases filesystems will
694  * not need to deal directly with kuid_t and kgid_t and can
695  * instead deal with the raw numeric values that are stored
696  * in the filesystem.
697  */
698 static inline uid_t i_uid_read(const struct inode *inode)
699 {
700         return from_kuid(&init_user_ns, inode->i_uid);
701 }
702 
703 static inline gid_t i_gid_read(const struct inode *inode)
704 {
705         return from_kgid(&init_user_ns, inode->i_gid);
706 }
707 
708 static inline void i_uid_write(struct inode *inode, uid_t uid)
709 {
710         inode->i_uid = make_kuid(&init_user_ns, uid);
711 }
712 
713 static inline void i_gid_write(struct inode *inode, gid_t gid)
714 {
715         inode->i_gid = make_kgid(&init_user_ns, gid);
716 }
717 
718 static inline unsigned iminor(const struct inode *inode)
719 {
720         return MINOR(inode->i_rdev);
721 }
722 
723 static inline unsigned imajor(const struct inode *inode)
724 {
725         return MAJOR(inode->i_rdev);
726 }
727 
728 extern struct block_device *I_BDEV(struct inode *inode);
729 
730 struct fown_struct {
731         rwlock_t lock;          /* protects pid, uid, euid fields */
732         struct pid *pid;        /* pid or -pgrp where SIGIO should be sent */
733         enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
734         kuid_t uid, euid;       /* uid/euid of process setting the owner */
735         int signum;             /* posix.1b rt signal to be delivered on IO */
736 };
737 
738 /*
739  * Track a single file's readahead state
740  */
741 struct file_ra_state {
742         pgoff_t start;                  /* where readahead started */
743         unsigned int size;              /* # of readahead pages */
744         unsigned int async_size;        /* do asynchronous readahead when
745                                            there are only # of pages ahead */
746 
747         unsigned int ra_pages;          /* Maximum readahead window */
748         unsigned int mmap_miss;         /* Cache miss stat for mmap accesses */
749         loff_t prev_pos;                /* Cache last read() position */
750 };
751 
752 /*
753  * Check if @index falls in the readahead windows.
754  */
755 static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
756 {
757         return (index >= ra->start &&
758                 index <  ra->start + ra->size);
759 }
760 
761 #define FILE_MNT_WRITE_TAKEN    1
762 #define FILE_MNT_WRITE_RELEASED 2
763 
764 struct file {
765         union {
766                 struct llist_node       fu_llist;
767                 struct rcu_head         fu_rcuhead;
768         } f_u;
769         struct path             f_path;
770 #define f_dentry        f_path.dentry
771         struct inode            *f_inode;       /* cached value */
772         const struct file_operations    *f_op;
773 
774         /*
775          * Protects f_ep_links, f_flags, f_pos vs i_size in lseek SEEK_CUR.
776          * Must not be taken from IRQ context.
777          */
778         spinlock_t              f_lock;
779         atomic_long_t           f_count;
780         unsigned int            f_flags;
781         fmode_t                 f_mode;
782         loff_t                  f_pos;
783         struct fown_struct      f_owner;
784         const struct cred       *f_cred;
785         struct file_ra_state    f_ra;
786 
787         u64                     f_version;
788 #ifdef CONFIG_SECURITY
789         void                    *f_security;
790 #endif
791         /* needed for tty driver, and maybe others */
792         void                    *private_data;
793 
794 #ifdef CONFIG_EPOLL
795         /* Used by fs/eventpoll.c to link all the hooks to this file */
796         struct list_head        f_ep_links;
797         struct list_head        f_tfile_llink;
798 #endif /* #ifdef CONFIG_EPOLL */
799         struct address_space    *f_mapping;
800 #ifdef CONFIG_DEBUG_WRITECOUNT
801         unsigned long f_mnt_write_state;
802 #endif
803 };
804 
805 struct file_handle {
806         __u32 handle_bytes;
807         int handle_type;
808         /* file identifier */
809         unsigned char f_handle[0];
810 };
811 
812 static inline struct file *get_file(struct file *f)
813 {
814         atomic_long_inc(&f->f_count);
815         return f;
816 }
817 #define fput_atomic(x)  atomic_long_add_unless(&(x)->f_count, -1, 1)
818 #define file_count(x)   atomic_long_read(&(x)->f_count)
819 
820 #ifdef CONFIG_DEBUG_WRITECOUNT
821 static inline void file_take_write(struct file *f)
822 {
823         WARN_ON(f->f_mnt_write_state != 0);
824         f->f_mnt_write_state = FILE_MNT_WRITE_TAKEN;
825 }
826 static inline void file_release_write(struct file *f)
827 {
828         f->f_mnt_write_state |= FILE_MNT_WRITE_RELEASED;
829 }
830 static inline void file_reset_write(struct file *f)
831 {
832         f->f_mnt_write_state = 0;
833 }
834 static inline void file_check_state(struct file *f)
835 {
836         /*
837          * At this point, either both or neither of these bits
838          * should be set.
839          */
840         WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN);
841         WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_RELEASED);
842 }
843 static inline int file_check_writeable(struct file *f)
844 {
845         if (f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN)
846                 return 0;
847         printk(KERN_WARNING "writeable file with no "
848                             "mnt_want_write()\n");
849         WARN_ON(1);
850         return -EINVAL;
851 }
852 #else /* !CONFIG_DEBUG_WRITECOUNT */
853 static inline void file_take_write(struct file *filp) {}
854 static inline void file_release_write(struct file *filp) {}
855 static inline void file_reset_write(struct file *filp) {}
856 static inline void file_check_state(struct file *filp) {}
857 static inline int file_check_writeable(struct file *filp)
858 {
859         return 0;
860 }
861 #endif /* CONFIG_DEBUG_WRITECOUNT */
862 
863 #define MAX_NON_LFS     ((1UL<<31) - 1)
864 
865 /* Page cache limit. The filesystems should put that into their s_maxbytes 
866    limits, otherwise bad things can happen in VM. */ 
867 #if BITS_PER_LONG==32
868 #define MAX_LFS_FILESIZE        (((loff_t)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1) 
869 #elif BITS_PER_LONG==64
870 #define MAX_LFS_FILESIZE        ((loff_t)0x7fffffffffffffffLL)
871 #endif
872 
873 #define FL_POSIX        1
874 #define FL_FLOCK        2
875 #define FL_ACCESS       8       /* not trying to lock, just looking */
876 #define FL_EXISTS       16      /* when unlocking, test for existence */
877 #define FL_LEASE        32      /* lease held on this file */
878 #define FL_CLOSE        64      /* unlock on close */
879 #define FL_SLEEP        128     /* A blocking lock */
880 #define FL_DOWNGRADE_PENDING    256 /* Lease is being downgraded */
881 #define FL_UNLOCK_PENDING       512 /* Lease is being broken */
882 
883 /*
884  * Special return value from posix_lock_file() and vfs_lock_file() for
885  * asynchronous locking.
886  */
887 #define FILE_LOCK_DEFERRED 1
888 
889 /*
890  * The POSIX file lock owner is determined by
891  * the "struct files_struct" in the thread group
892  * (or NULL for no owner - BSD locks).
893  *
894  * Lockd stuffs a "host" pointer into this.
895  */
896 typedef struct files_struct *fl_owner_t;
897 
898 struct file_lock_operations {
899         void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
900         void (*fl_release_private)(struct file_lock *);
901 };
902 
903 struct lock_manager_operations {
904         int (*lm_compare_owner)(struct file_lock *, struct file_lock *);
905         void (*lm_notify)(struct file_lock *);  /* unblock callback */
906         int (*lm_grant)(struct file_lock *, struct file_lock *, int);
907         void (*lm_break)(struct file_lock *);
908         int (*lm_change)(struct file_lock **, int);
909 };
910 
911 struct lock_manager {
912         struct list_head list;
913 };
914 
915 struct net;
916 void locks_start_grace(struct net *, struct lock_manager *);
917 void locks_end_grace(struct lock_manager *);
918 int locks_in_grace(struct net *);
919 
920 /* that will die - we need it for nfs_lock_info */
921 #include <linux/nfs_fs_i.h>
922 
923 struct file_lock {
924         struct file_lock *fl_next;      /* singly linked list for this inode  */
925         struct list_head fl_link;       /* doubly linked list of all locks */
926         struct list_head fl_block;      /* circular list of blocked processes */
927         fl_owner_t fl_owner;
928         unsigned int fl_flags;
929         unsigned char fl_type;
930         unsigned int fl_pid;
931         struct pid *fl_nspid;
932         wait_queue_head_t fl_wait;
933         struct file *fl_file;
934         loff_t fl_start;
935         loff_t fl_end;
936 
937         struct fasync_struct *  fl_fasync; /* for lease break notifications */
938         /* for lease breaks: */
939         unsigned long fl_break_time;
940         unsigned long fl_downgrade_time;
941 
942         const struct file_lock_operations *fl_ops;      /* Callbacks for filesystems */
943         const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
944         union {
945                 struct nfs_lock_info    nfs_fl;
946                 struct nfs4_lock_info   nfs4_fl;
947                 struct {
948                         struct list_head link;  /* link in AFS vnode's pending_locks list */
949                         int state;              /* state of grant or error if -ve */
950                 } afs;
951         } fl_u;
952 };
953 
954 /* The following constant reflects the upper bound of the file/locking space */
955 #ifndef OFFSET_MAX
956 #define INT_LIMIT(x)    (~((x)1 << (sizeof(x)*8 - 1)))
957 #define OFFSET_MAX      INT_LIMIT(loff_t)
958 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
959 #endif
960 
961 #include <linux/fcntl.h>
962 
963 extern void send_sigio(struct fown_struct *fown, int fd, int band);
964 
965 #ifdef CONFIG_FILE_LOCKING
966 extern int fcntl_getlk(struct file *, struct flock __user *);
967 extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
968                         struct flock __user *);
969 
970 #if BITS_PER_LONG == 32
971 extern int fcntl_getlk64(struct file *, struct flock64 __user *);
972 extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
973                         struct flock64 __user *);
974 #endif
975 
976 extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
977 extern int fcntl_getlease(struct file *filp);
978 
979 /* fs/locks.c */
980 void locks_free_lock(struct file_lock *fl);
981 extern void locks_init_lock(struct file_lock *);
982 extern struct file_lock * locks_alloc_lock(void);
983 extern void locks_copy_lock(struct file_lock *, struct file_lock *);
984 extern void __locks_copy_lock(struct file_lock *, const struct file_lock *);
985 extern void locks_remove_posix(struct file *, fl_owner_t);
986 extern void locks_remove_flock(struct file *);
987 extern void locks_release_private(struct file_lock *);
988 extern void posix_test_lock(struct file *, struct file_lock *);
989 extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
990 extern int posix_lock_file_wait(struct file *, struct file_lock *);
991 extern int posix_unblock_lock(struct file *, struct file_lock *);
992 extern int vfs_test_lock(struct file *, struct file_lock *);
993 extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
994 extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
995 extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
996 extern int __break_lease(struct inode *inode, unsigned int flags);
997 extern void lease_get_mtime(struct inode *, struct timespec *time);
998 extern int generic_setlease(struct file *, long, struct file_lock **);
999 extern int vfs_setlease(struct file *, long, struct file_lock **);
1000 extern int lease_modify(struct file_lock **, int);
1001 extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
1002 extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
1003 extern void locks_delete_block(struct file_lock *waiter);
1004 extern void lock_flocks(void);
1005 extern void unlock_flocks(void);
1006 #else /* !CONFIG_FILE_LOCKING */
1007 static inline int fcntl_getlk(struct file *file, struct flock __user *user)
1008 {
1009         return -EINVAL;
1010 }
1011 
1012 static inline int fcntl_setlk(unsigned int fd, struct file *file,
1013                               unsigned int cmd, struct flock __user *user)
1014 {
1015         return -EACCES;
1016 }
1017 
1018 #if BITS_PER_LONG == 32
1019 static inline int fcntl_getlk64(struct file *file, struct flock64 __user *user)
1020 {
1021         return -EINVAL;
1022 }
1023 
1024 static inline int fcntl_setlk64(unsigned int fd, struct file *file,
1025                                 unsigned int cmd, struct flock64 __user *user)
1026 {
1027         return -EACCES;
1028 }
1029 #endif
1030 static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
1031 {
1032         return 0;
1033 }
1034 
1035 static inline int fcntl_getlease(struct file *filp)
1036 {
1037         return 0;
1038 }
1039 
1040 static inline void locks_init_lock(struct file_lock *fl)
1041 {
1042         return;
1043 }
1044 
1045 static inline void __locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1046 {
1047         return;
1048 }
1049 
1050 static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1051 {
1052         return;
1053 }
1054 
1055 static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
1056 {
1057         return;
1058 }
1059 
1060 static inline void locks_remove_flock(struct file *filp)
1061 {
1062         return;
1063 }
1064 
1065 static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
1066 {
1067         return;
1068 }
1069 
1070 static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
1071                                   struct file_lock *conflock)
1072 {
1073         return -ENOLCK;
1074 }
1075 
1076 static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
1077 {
1078         return -ENOLCK;
1079 }
1080 
1081 static inline int posix_unblock_lock(struct file *filp,
1082                                      struct file_lock *waiter)
1083 {
1084         return -ENOENT;
1085 }
1086 
1087 static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
1088 {
1089         return 0;
1090 }
1091 
1092 static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
1093                                 struct file_lock *fl, struct file_lock *conf)
1094 {
1095         return -ENOLCK;
1096 }
1097 
1098 static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
1099 {
1100         return 0;
1101 }
1102 
1103 static inline int flock_lock_file_wait(struct file *filp,
1104                                        struct file_lock *request)
1105 {
1106         return -ENOLCK;
1107 }
1108 
1109 static inline int __break_lease(struct inode *inode, unsigned int mode)
1110 {
1111         return 0;
1112 }
1113 
1114 static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
1115 {
1116         return;
1117 }
1118 
1119 static inline int generic_setlease(struct file *filp, long arg,
1120                                     struct file_lock **flp)
1121 {
1122         return -EINVAL;
1123 }
1124 
1125 static inline int vfs_setlease(struct file *filp, long arg,
1126                                struct file_lock **lease)
1127 {
1128         return -EINVAL;
1129 }
1130 
1131 static inline int lease_modify(struct file_lock **before, int arg)
1132 {
1133         return -EINVAL;
1134 }
1135 
1136 static inline int lock_may_read(struct inode *inode, loff_t start,
1137                                 unsigned long len)
1138 {
1139         return 1;
1140 }
1141 
1142 static inline int lock_may_write(struct inode *inode, loff_t start,
1143                                  unsigned long len)
1144 {
1145         return 1;
1146 }
1147 
1148 static inline void locks_delete_block(struct file_lock *waiter)
1149 {
1150 }
1151 
1152 static inline void lock_flocks(void)
1153 {
1154 }
1155 
1156 static inline void unlock_flocks(void)
1157 {
1158 }
1159 
1160 #endif /* !CONFIG_FILE_LOCKING */
1161 
1162 
1163 struct fasync_struct {
1164         spinlock_t              fa_lock;
1165         int                     magic;
1166         int                     fa_fd;
1167         struct fasync_struct    *fa_next; /* singly linked list */
1168         struct file             *fa_file;
1169         struct rcu_head         fa_rcu;
1170 };
1171 
1172 #define FASYNC_MAGIC 0x4601
1173 
1174 /* SMP safe fasync helpers: */
1175 extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
1176 extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
1177 extern int fasync_remove_entry(struct file *, struct fasync_struct **);
1178 extern struct fasync_struct *fasync_alloc(void);
1179 extern void fasync_free(struct fasync_struct *);
1180 
1181 /* can be called from interrupts */
1182 extern void kill_fasync(struct fasync_struct **, int, int);
1183 
1184 extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
1185 extern int f_setown(struct file *filp, unsigned long arg, int force);
1186 extern void f_delown(struct file *filp);
1187 extern pid_t f_getown(struct file *filp);
1188 extern int send_sigurg(struct fown_struct *fown);
1189 
1190 struct mm_struct;
1191 
1192 /*
1193  *      Umount options
1194  */
1195 
1196 #define MNT_FORCE       0x00000001      /* Attempt to forcibily umount */
1197 #define MNT_DETACH      0x00000002      /* Just detach from the tree */
1198 #define MNT_EXPIRE      0x00000004      /* Mark for expiry */
1199 #define UMOUNT_NOFOLLOW 0x00000008      /* Don't follow symlink on umount */
1200 #define UMOUNT_UNUSED   0x80000000      /* Flag guaranteed to be unused */
1201 
1202 extern struct list_head super_blocks;
1203 extern spinlock_t sb_lock;
1204 
1205 /* Possible states of 'frozen' field */
1206 enum {
1207         SB_UNFROZEN = 0,                /* FS is unfrozen */
1208         SB_FREEZE_WRITE = 1,            /* Writes, dir ops, ioctls frozen */
1209         SB_FREEZE_PAGEFAULT = 2,        /* Page faults stopped as well */
1210         SB_FREEZE_FS = 3,               /* For internal FS use (e.g. to stop
1211                                          * internal threads if needed) */
1212         SB_FREEZE_COMPLETE = 4,         /* ->freeze_fs finished successfully */
1213 };
1214 
1215 #define SB_FREEZE_LEVELS (SB_FREEZE_COMPLETE - 1)
1216 
1217 struct sb_writers {
1218         /* Counters for counting writers at each level */
1219         struct percpu_counter   counter[SB_FREEZE_LEVELS];
1220         wait_queue_head_t       wait;           /* queue for waiting for
1221                                                    writers / faults to finish */
1222         int                     frozen;         /* Is sb frozen? */
1223         wait_queue_head_t       wait_unfrozen;  /* queue for waiting for
1224                                                    sb to be thawed */
1225 #ifdef CONFIG_DEBUG_LOCK_ALLOC
1226         struct lockdep_map      lock_map[SB_FREEZE_LEVELS];
1227 #endif
1228 };
1229 
1230 struct super_block {
1231         struct list_head        s_list;         /* Keep this first */
1232         dev_t                   s_dev;          /* search index; _not_ kdev_t */
1233         unsigned char           s_blocksize_bits;
1234         unsigned long           s_blocksize;
1235         loff_t                  s_maxbytes;     /* Max file size */
1236         struct file_system_type *s_type;
1237         const struct super_operations   *s_op;
1238         const struct dquot_operations   *dq_op;
1239         const struct quotactl_ops       *s_qcop;
1240         const struct export_operations *s_export_op;
1241         unsigned long           s_flags;
1242         unsigned long           s_magic;
1243         struct dentry           *s_root;
1244         struct rw_semaphore     s_umount;
1245         int                     s_count;
1246         atomic_t                s_active;
1247 #ifdef CONFIG_SECURITY
1248         void                    *s_security;
1249 #endif
1250         const struct xattr_handler **s_xattr;
1251 
1252         struct list_head        s_inodes;       /* all inodes */
1253         struct hlist_bl_head    s_anon;         /* anonymous dentries for (nfs) exporting */
1254         struct list_head        s_mounts;       /* list of mounts; _not_ for fs use */
1255         /* s_dentry_lru, s_nr_dentry_unused protected by dcache.c lru locks */
1256         struct list_head        s_dentry_lru;   /* unused dentry lru */
1257         int                     s_nr_dentry_unused;     /* # of dentry on lru */
1258 
1259         /* s_inode_lru_lock protects s_inode_lru and s_nr_inodes_unused */
1260         spinlock_t              s_inode_lru_lock ____cacheline_aligned_in_smp;
1261         struct list_head        s_inode_lru;            /* unused inode lru */
1262         int                     s_nr_inodes_unused;     /* # of inodes on lru */
1263 
1264         struct block_device     *s_bdev;
1265         struct backing_dev_info *s_bdi;
1266         struct mtd_info         *s_mtd;
1267         struct hlist_node       s_instances;
1268         struct quota_info       s_dquot;        /* Diskquota specific options */
1269 
1270         struct sb_writers       s_writers;
1271 
1272         char s_id[32];                          /* Informational name */
1273         u8 s_uuid[16];                          /* UUID */
1274 
1275         void                    *s_fs_info;     /* Filesystem private info */
1276         unsigned int            s_max_links;
1277         fmode_t                 s_mode;
1278 
1279         /* Granularity of c/m/atime in ns.
1280            Cannot be worse than a second */
1281         u32                s_time_gran;
1282 
1283         /*
1284          * The next field is for VFS *only*. No filesystems have any business
1285          * even looking at it. You had been warned.
1286          */
1287         struct mutex s_vfs_rename_mutex;        /* Kludge */
1288 
1289         /*
1290          * Filesystem subtype.  If non-empty the filesystem type field
1291          * in /proc/mounts will be "type.subtype"
1292          */
1293         char *s_subtype;
1294 
1295         /*
1296          * Saved mount options for lazy filesystems using
1297          * generic_show_options()
1298          */
1299         char __rcu *s_options;
1300         const struct dentry_operations *s_d_op; /* default d_op for dentries */
1301 
1302         /*
1303          * Saved pool identifier for cleancache (-1 means none)
1304          */
1305         int cleancache_poolid;
1306 
1307         struct shrinker s_shrink;       /* per-sb shrinker handle */
1308 
1309         /* Number of inodes with nlink == 0 but still referenced */
1310         atomic_long_t s_remove_count;
1311 
1312         /* Being remounted read-only */
1313         int s_readonly_remount;
1314 };
1315 
1316 /* superblock cache pruning functions */
1317 extern void prune_icache_sb(struct super_block *sb, int nr_to_scan);
1318 extern void prune_dcache_sb(struct super_block *sb, int nr_to_scan);
1319 
1320 extern struct timespec current_fs_time(struct super_block *sb);
1321 
1322 /*
1323  * Snapshotting support.
1324  */
1325 
1326 void __sb_end_write(struct super_block *sb, int level);
1327 int __sb_start_write(struct super_block *sb, int level, bool wait);
1328 
1329 /**
1330  * sb_end_write - drop write access to a superblock
1331  * @sb: the super we wrote to
1332  *
1333  * Decrement number of writers to the filesystem. Wake up possible waiters
1334  * wanting to freeze the filesystem.
1335  */
1336 static inline void sb_end_write(struct super_block *sb)
1337 {
1338         __sb_end_write(sb, SB_FREEZE_WRITE);
1339 }
1340 
1341 /**
1342  * sb_end_pagefault - drop write access to a superblock from a page fault
1343  * @sb: the super we wrote to
1344  *
1345  * Decrement number of processes handling write page fault to the filesystem.
1346  * Wake up possible waiters wanting to freeze the filesystem.
1347  */
1348 static inline void sb_end_pagefault(struct super_block *sb)
1349 {
1350         __sb_end_write(sb, SB_FREEZE_PAGEFAULT);
1351 }
1352 
1353 /**
1354  * sb_end_intwrite - drop write access to a superblock for internal fs purposes
1355  * @sb: the super we wrote to
1356  *
1357  * Decrement fs-internal number of writers to the filesystem.  Wake up possible
1358  * waiters wanting to freeze the filesystem.
1359  */
1360 static inline void sb_end_intwrite(struct super_block *sb)
1361 {
1362         __sb_end_write(sb, SB_FREEZE_FS);
1363 }
1364 
1365 /**
1366  * sb_start_write - get write access to a superblock
1367  * @sb: the super we write to
1368  *
1369  * When a process wants to write data or metadata to a file system (i.e. dirty
1370  * a page or an inode), it should embed the operation in a sb_start_write() -
1371  * sb_end_write() pair to get exclusion against file system freezing. This
1372  * function increments number of writers preventing freezing. If the file
1373  * system is already frozen, the function waits until the file system is
1374  * thawed.
1375  *
1376  * Since freeze protection behaves as a lock, users have to preserve
1377  * ordering of freeze protection and other filesystem locks. Generally,
1378  * freeze protection should be the outermost lock. In particular, we have:
1379  *
1380  * sb_start_write
1381  *   -> i_mutex                 (write path, truncate, directory ops, ...)
1382  *   -> s_umount                (freeze_super, thaw_super)
1383  */
1384 static inline void sb_start_write(struct super_block *sb)
1385 {
1386         __sb_start_write(sb, SB_FREEZE_WRITE, true);
1387 }
1388 
1389 static inline int sb_start_write_trylock(struct super_block *sb)
1390 {
1391         return __sb_start_write(sb, SB_FREEZE_WRITE, false);
1392 }
1393 
1394 /**
1395  * sb_start_pagefault - get write access to a superblock from a page fault
1396  * @sb: the super we write to
1397  *
1398  * When a process starts handling write page fault, it should embed the
1399  * operation into sb_start_pagefault() - sb_end_pagefault() pair to get
1400  * exclusion against file system freezing. This is needed since the page fault
1401  * is going to dirty a page. This function increments number of running page
1402  * faults preventing freezing. If the file system is already frozen, the
1403  * function waits until the file system is thawed.
1404  *
1405  * Since page fault freeze protection behaves as a lock, users have to preserve
1406  * ordering of freeze protection and other filesystem locks. It is advised to
1407  * put sb_start_pagefault() close to mmap_sem in lock ordering. Page fault
1408  * handling code implies lock dependency:
1409  *
1410  * mmap_sem
1411  *   -> sb_start_pagefault
1412  */
1413 static inline void sb_start_pagefault(struct super_block *sb)
1414 {
1415         __sb_start_write(sb, SB_FREEZE_PAGEFAULT, true);
1416 }
1417 
1418 /*
1419  * sb_start_intwrite - get write access to a superblock for internal fs purposes
1420  * @sb: the super we write to
1421  *
1422  * This is the third level of protection against filesystem freezing. It is
1423  * free for use by a filesystem. The only requirement is that it must rank
1424  * below sb_start_pagefault.
1425  *
1426  * For example filesystem can call sb_start_intwrite() when starting a
1427  * transaction which somewhat eases handling of freezing for internal sources
1428  * of filesystem changes (internal fs threads, discarding preallocation on file
1429  * close, etc.).
1430  */
1431 static inline void sb_start_intwrite(struct super_block *sb)
1432 {
1433         __sb_start_write(sb, SB_FREEZE_FS, true);
1434 }
1435 
1436 
1437 extern bool inode_owner_or_capable(const struct inode *inode);
1438 
1439 /*
1440  * VFS helper functions..
1441  */
1442 extern int vfs_create(struct inode *, struct dentry *, umode_t, bool);
1443 extern int vfs_mkdir(struct inode *, struct dentry *, umode_t);
1444 extern int vfs_mknod(struct inode *, struct dentry *, umode_t, dev_t);
1445 extern int vfs_symlink(struct inode *, struct dentry *, const char *);
1446 extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
1447 extern int vfs_rmdir(struct inode *, struct dentry *);
1448 extern int vfs_unlink(struct inode *, struct dentry *);
1449 extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
1450 
1451 /*
1452  * VFS dentry helper functions.
1453  */
1454 extern void dentry_unhash(struct dentry *dentry);
1455 
1456 /*
1457  * VFS file helper functions.
1458  */
1459 extern void inode_init_owner(struct inode *inode, const struct inode *dir,
1460                         umode_t mode);
1461 /*
1462  * VFS FS_IOC_FIEMAP helper definitions.
1463  */
1464 struct fiemap_extent_info {
1465         unsigned int fi_flags;          /* Flags as passed from user */
1466         unsigned int fi_extents_mapped; /* Number of mapped extents */
1467         unsigned int fi_extents_max;    /* Size of fiemap_extent array */
1468         struct fiemap_extent __user *fi_extents_start; /* Start of
1469                                                         fiemap_extent array */
1470 };
1471 int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
1472                             u64 phys, u64 len, u32 flags);
1473 int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
1474 
1475 /*
1476  * File types
1477  *
1478  * NOTE! These match bits 12..15 of stat.st_mode
1479  * (ie "(i_mode >> 12) & 15").
1480  */
1481 #define DT_UNKNOWN      0
1482 #define DT_FIFO         1
1483 #define DT_CHR          2
1484 #define DT_DIR          4
1485 #define DT_BLK          6
1486 #define DT_REG          8
1487 #define DT_LNK          10
1488 #define DT_SOCK         12
1489 #define DT_WHT          14
1490 
1491 /*
1492  * This is the "filldir" function type, used by readdir() to let
1493  * the kernel specify what kind of dirent layout it wants to have.
1494  * This allows the kernel to read directories into kernel space or
1495  * to have different dirent layouts depending on the binary type.
1496  */
1497 typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned);
1498 struct block_device_operations;
1499 
1500 /* These macros are for out of kernel modules to test that
1501  * the kernel supports the unlocked_ioctl and compat_ioctl
1502  * fields in struct file_operations. */
1503 #define HAVE_COMPAT_IOCTL 1
1504 #define HAVE_UNLOCKED_IOCTL 1
1505 
1506 struct file_operations {
1507         struct module *owner;
1508         loff_t (*llseek) (struct file *, loff_t, int);
1509         ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1510         ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
1511         ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1512         ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1513         int (*readdir) (struct file *, void *, filldir_t);
1514         unsigned int (*poll) (struct file *, struct poll_table_struct *);
1515         long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1516         long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1517         int (*mmap) (struct file *, struct vm_area_struct *);
1518         int (*open) (struct inode *, struct file *);
1519         int (*flush) (struct file *, fl_owner_t id);
1520         int (*release) (struct inode *, struct file *);
1521         int (*fsync) (struct file *, loff_t, loff_t, int datasync);
1522         int (*aio_fsync) (struct kiocb *, int datasync);
1523         int (*fasync) (int, struct file *, int);
1524         int (*lock) (struct file *, int, struct file_lock *);
1525         ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1526         unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1527         int (*check_flags)(int);
1528         int (*flock) (struct file *, int, struct file_lock *);
1529         ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1530         ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
1531         int (*setlease)(struct file *, long, struct file_lock **);
1532         long (*fallocate)(struct file *file, int mode, loff_t offset,
1533                           loff_t len);
1534         int (*show_fdinfo)(struct seq_file *m, struct file *f);
1535 };
1536 
1537 struct inode_operations {
1538         struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int);
1539         void * (*follow_link) (struct dentry *, struct nameidata *);
1540         int (*permission) (struct inode *, int);
1541         struct posix_acl * (*get_acl)(struct inode *, int);
1542 
1543         int (*readlink) (struct dentry *, char __user *,int);
1544         void (*put_link) (struct dentry *, struct nameidata *, void *);
1545 
1546         int (*create) (struct inode *,struct dentry *, umode_t, bool);
1547         int (*link) (struct dentry *,struct inode *,struct dentry *);
1548         int (*unlink) (struct inode *,struct dentry *);
1549         int (*symlink) (struct inode *,struct dentry *,const char *);
1550         int (*mkdir) (struct inode *,struct dentry *,umode_t);
1551         int (*rmdir) (struct inode *,struct dentry *);
1552         int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t);
1553         int (*rename) (struct inode *, struct dentry *,
1554                         struct inode *, struct dentry *);
1555         int (*setattr) (struct dentry *, struct iattr *);
1556         int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1557         int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1558         ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1559         ssize_t (*listxattr) (struct dentry *, char *, size_t);
1560         int (*removexattr) (struct dentry *, const char *);
1561         int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
1562                       u64 len);
1563         int (*update_time)(struct inode *, struct timespec *, int);
1564         int (*atomic_open)(struct inode *, struct dentry *,
1565                            struct file *, unsigned open_flag,
1566                            umode_t create_mode, int *opened);
1567 } ____cacheline_aligned;
1568 
1569 ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
1570                               unsigned long nr_segs, unsigned long fast_segs,
1571                               struct iovec *fast_pointer,
1572                               struct iovec **ret_pointer);
1573 
1574 extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1575 extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1576 extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1577                 unsigned long, loff_t *);
1578 extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1579                 unsigned long, loff_t *);
1580 
1581 struct super_operations {
1582         struct inode *(*alloc_inode)(struct super_block *sb);
1583         void (*destroy_inode)(struct inode *);
1584 
1585         void (*dirty_inode) (struct inode *, int flags);
1586         int (*write_inode) (struct inode *, struct writeback_control *wbc);
1587         int (*drop_inode) (struct inode *);
1588         void (*evict_inode) (struct inode *);
1589         void (*put_super) (struct super_block *);
1590         int (*sync_fs)(struct super_block *sb, int wait);
1591         int (*freeze_fs) (struct super_block *);
1592         int (*unfreeze_fs) (struct super_block *);
1593         int (*statfs) (struct dentry *, struct kstatfs *);
1594         int (*remount_fs) (struct super_block *, int *, char *);
1595         void (*umount_begin) (struct super_block *);
1596 
1597         int (*show_options)(struct seq_file *, struct dentry *);
1598         int (*show_devname)(struct seq_file *, struct dentry *);
1599         int (*show_path)(struct seq_file *, struct dentry *);
1600         int (*show_stats)(struct seq_file *, struct dentry *);
1601 #ifdef CONFIG_QUOTA
1602         ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1603         ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
1604 #endif
1605         int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
1606         int (*nr_cached_objects)(struct super_block *);
1607         void (*free_cached_objects)(struct super_block *, int);
1608 };
1609 
1610 /*
1611  * Inode flags - they have no relation to superblock flags now
1612  */
1613 #define S_SYNC          1       /* Writes are synced at once */
1614 #define S_NOATIME       2       /* Do not update access times */
1615 #define S_APPEND        4       /* Append-only file */
1616 #define S_IMMUTABLE     8       /* Immutable file */
1617 #define S_DEAD          16      /* removed, but still open directory */
1618 #define S_NOQUOTA       32      /* Inode is not counted to quota */
1619 #define S_DIRSYNC       64      /* Directory modifications are synchronous */
1620 #define S_NOCMTIME      128     /* Do not update file c/mtime */
1621 #define S_SWAPFILE      256     /* Do not truncate: swapon got its bmaps */
1622 #define S_PRIVATE       512     /* Inode is fs-internal */
1623 #define S_IMA           1024    /* Inode has an associated IMA struct */
1624 #define S_AUTOMOUNT     2048    /* Automount/referral quasi-directory */
1625 #define S_NOSEC         4096    /* no suid or xattr security attributes */
1626 
1627 /*
1628  * Note that nosuid etc flags are inode-specific: setting some file-system
1629  * flags just means all the inodes inherit those flags by default. It might be
1630  * possible to override it selectively if you really wanted to with some
1631  * ioctl() that is not currently implemented.
1632  *
1633  * Exception: MS_RDONLY is always applied to the entire file system.
1634  *
1635  * Unfortunately, it is possible to change a filesystems flags with it mounted
1636  * with files in use.  This means that all of the inodes will not have their
1637  * i_flags updated.  Hence, i_flags no longer inherit the superblock mount
1638  * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
1639  */
1640 #define __IS_FLG(inode, flg)    ((inode)->i_sb->s_flags & (flg))
1641 
1642 #define IS_RDONLY(inode)        ((inode)->i_sb->s_flags & MS_RDONLY)
1643 #define IS_SYNC(inode)          (__IS_FLG(inode, MS_SYNCHRONOUS) || \
1644                                         ((inode)->i_flags & S_SYNC))
1645 #define IS_DIRSYNC(inode)       (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
1646                                         ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
1647 #define IS_MANDLOCK(inode)      __IS_FLG(inode, MS_MANDLOCK)
1648 #define IS_NOATIME(inode)       __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
1649 #define IS_I_VERSION(inode)     __IS_FLG(inode, MS_I_VERSION)
1650 
1651 #define IS_NOQUOTA(inode)       ((inode)->i_flags & S_NOQUOTA)
1652 #define IS_APPEND(inode)        ((inode)->i_flags & S_APPEND)
1653 #define IS_IMMUTABLE(inode)     ((inode)->i_flags & S_IMMUTABLE)
1654 #define IS_POSIXACL(inode)      __IS_FLG(inode, MS_POSIXACL)
1655 
1656 #define IS_DEADDIR(inode)       ((inode)->i_flags & S_DEAD)
1657 #define IS_NOCMTIME(inode)      ((inode)->i_flags & S_NOCMTIME)
1658 #define IS_SWAPFILE(inode)      ((inode)->i_flags & S_SWAPFILE)
1659 #define IS_PRIVATE(inode)       ((inode)->i_flags & S_PRIVATE)
1660 #define IS_IMA(inode)           ((inode)->i_flags & S_IMA)
1661 #define IS_AUTOMOUNT(inode)     ((inode)->i_flags & S_AUTOMOUNT)
1662 #define IS_NOSEC(inode)         ((inode)->i_flags & S_NOSEC)
1663 
1664 /*
1665  * Inode state bits.  Protected by inode->i_lock
1666  *
1667  * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1668  * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1669  *
1670  * Four bits define the lifetime of an inode.  Initially, inodes are I_NEW,
1671  * until that flag is cleared.  I_WILL_FREE, I_FREEING and I_CLEAR are set at
1672  * various stages of removing an inode.
1673  *
1674  * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1675  *
1676  * I_DIRTY_SYNC         Inode is dirty, but doesn't have to be written on
1677  *                      fdatasync().  i_atime is the usual cause.
1678  * I_DIRTY_DATASYNC     Data-related inode changes pending. We keep track of
1679  *                      these changes separately from I_DIRTY_SYNC so that we
1680  *                      don't have to write inode on fdatasync() when only
1681  *                      mtime has changed in it.
1682  * I_DIRTY_PAGES        Inode has dirty pages.  Inode itself may be clean.
1683  * I_NEW                Serves as both a mutex and completion notification.
1684  *                      New inodes set I_NEW.  If two processes both create
1685  *                      the same inode, one of them will release its inode and
1686  *                      wait for I_NEW to be released before returning.
1687  *                      Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1688  *                      also cause waiting on I_NEW, without I_NEW actually
1689  *                      being set.  find_inode() uses this to prevent returning
1690  *                      nearly-dead inodes.
1691  * I_WILL_FREE          Must be set when calling write_inode_now() if i_count
1692  *                      is zero.  I_FREEING must be set when I_WILL_FREE is
1693  *                      cleared.
1694  * I_FREEING            Set when inode is about to be freed but still has dirty
1695  *                      pages or buffers attached or the inode itself is still
1696  *                      dirty.
1697  * I_CLEAR              Added by clear_inode().  In this state the inode is
1698  *                      clean and can be destroyed.  Inode keeps I_FREEING.
1699  *
1700  *                      Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1701  *                      prohibited for many purposes.  iget() must wait for
1702  *                      the inode to be completely released, then create it
1703  *                      anew.  Other functions will just ignore such inodes,
1704  *                      if appropriate.  I_NEW is used for waiting.
1705  *
1706  * I_SYNC               Writeback of inode is running. The bit is set during
1707  *                      data writeback, and cleared with a wakeup on the bit
1708  *                      address once it is done. The bit is also used to pin
1709  *                      the inode in memory for flusher thread.
1710  *
1711  * I_REFERENCED         Marks the inode as recently references on the LRU list.
1712  *
1713  * I_DIO_WAKEUP         Never set.  Only used as a key for wait_on_bit().
1714  *
1715  * Q: What is the difference between I_WILL_FREE and I_FREEING?
1716  */
1717 #define I_DIRTY_SYNC            (1 << 0)
1718 #define I_DIRTY_DATASYNC        (1 << 1)
1719 #define I_DIRTY_PAGES           (1 << 2)
1720 #define __I_NEW                 3
1721 #define I_NEW                   (1 << __I_NEW)
1722 #define I_WILL_FREE             (1 << 4)
1723 #define I_FREEING               (1 << 5)
1724 #define I_CLEAR                 (1 << 6)
1725 #define __I_SYNC                7
1726 #define I_SYNC                  (1 << __I_SYNC)
1727 #define I_REFERENCED            (1 << 8)
1728 #define __I_DIO_WAKEUP          9
1729 #define I_DIO_WAKEUP            (1 << I_DIO_WAKEUP)
1730 
1731 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1732 
1733 extern void __mark_inode_dirty(struct inode *, int);
1734 static inline void mark_inode_dirty(struct inode *inode)
1735 {
1736         __mark_inode_dirty(inode, I_DIRTY);
1737 }
1738 
1739 static inline void mark_inode_dirty_sync(struct inode *inode)
1740 {
1741         __mark_inode_dirty(inode, I_DIRTY_SYNC);
1742 }
1743 
1744 extern void inc_nlink(struct inode *inode);
1745 extern void drop_nlink(struct inode *inode);
1746 extern void clear_nlink(struct inode *inode);
1747 extern void set_nlink(struct inode *inode, unsigned int nlink);
1748 
1749 static inline void inode_inc_link_count(struct inode *inode)
1750 {
1751         inc_nlink(inode);
1752         mark_inode_dirty(inode);
1753 }
1754 
1755 static inline void inode_dec_link_count(struct inode *inode)
1756 {
1757         drop_nlink(inode);
1758         mark_inode_dirty(inode);
1759 }
1760 
1761 /**
1762  * inode_inc_iversion - increments i_version
1763  * @inode: inode that need to be updated
1764  *
1765  * Every time the inode is modified, the i_version field will be incremented.
1766  * The filesystem has to be mounted with i_version flag
1767  */
1768 
1769 static inline void inode_inc_iversion(struct inode *inode)
1770 {
1771        spin_lock(&inode->i_lock);
1772        inode->i_version++;
1773        spin_unlock(&inode->i_lock);
1774 }
1775 
1776 enum file_time_flags {
1777         S_ATIME = 1,
1778         S_MTIME = 2,
1779         S_CTIME = 4,
1780         S_VERSION = 8,
1781 };
1782 
1783 extern void touch_atime(struct path *);
1784 static inline void file_accessed(struct file *file)
1785 {
1786         if (!(file->f_flags & O_NOATIME))
1787                 touch_atime(&file->f_path);
1788 }
1789 
1790 int sync_inode(struct inode *inode, struct writeback_control *wbc);
1791 int sync_inode_metadata(struct inode *inode, int wait);
1792 
1793 struct file_system_type {
1794         const char *name;
1795         int fs_flags;
1796 #define FS_REQUIRES_DEV         1 
1797 #define FS_BINARY_MOUNTDATA     2
1798 #define FS_HAS_SUBTYPE          4
1799 #define FS_USERNS_MOUNT         8       /* Can be mounted by userns root */
1800 #define FS_USERNS_DEV_MOUNT     16 /* A userns mount does not imply MNT_NODEV */
1801 #define FS_RENAME_DOES_D_MOVE   32768   /* FS will handle d_move() during rename() internally. */
1802         struct dentry *(*mount) (struct file_system_type *, int,
1803                        const char *, void *);
1804         void (*kill_sb) (struct super_block *);
1805         struct module *owner;
1806         struct file_system_type * next;
1807         struct hlist_head fs_supers;
1808 
1809         struct lock_class_key s_lock_key;
1810         struct lock_class_key s_umount_key;
1811         struct lock_class_key s_vfs_rename_key;
1812         struct lock_class_key s_writers_key[SB_FREEZE_LEVELS];
1813 
1814         struct lock_class_key i_lock_key;
1815         struct lock_class_key i_mutex_key;
1816         struct lock_class_key i_mutex_dir_key;
1817 };
1818 
1819 #define MODULE_ALIAS_FS(NAME) MODULE_ALIAS("fs-" NAME)
1820 
1821 extern struct dentry *mount_ns(struct file_system_type *fs_type, int flags,
1822         void *data, int (*fill_super)(struct super_block *, void *, int));
1823 extern struct dentry *mount_bdev(struct file_system_type *fs_type,
1824         int flags, const char *dev_name, void *data,
1825         int (*fill_super)(struct super_block *, void *, int));
1826 extern struct dentry *mount_single(struct file_system_type *fs_type,
1827         int flags, void *data,
1828         int (*fill_super)(struct super_block *, void *, int));
1829 extern struct dentry *mount_nodev(struct file_system_type *fs_type,
1830         int flags, void *data,
1831         int (*fill_super)(struct super_block *, void *, int));
1832 extern struct dentry *mount_subtree(struct vfsmount *mnt, const char *path);
1833 void generic_shutdown_super(struct super_block *sb);
1834 void kill_block_super(struct super_block *sb);
1835 void kill_anon_super(struct super_block *sb);
1836 void kill_litter_super(struct super_block *sb);
1837 void deactivate_super(struct super_block *sb);
1838 void deactivate_locked_super(struct super_block *sb);
1839 int set_anon_super(struct super_block *s, void *data);
1840 int get_anon_bdev(dev_t *);
1841 void free_anon_bdev(dev_t);
1842 struct super_block *sget(struct file_system_type *type,
1843                         int (*test)(struct super_block *,void *),
1844                         int (*set)(struct super_block *,void *),
1845                         int flags, void *data);
1846 extern struct dentry *mount_pseudo(struct file_system_type *, char *,
1847         const struct super_operations *ops,
1848         const struct dentry_operations *dops,
1849         unsigned long);
1850 
1851 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1852 #define fops_get(fops) \
1853         (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1854 #define fops_put(fops) \
1855         do { if (fops) module_put((fops)->owner); } while(0)
1856 
1857 extern int register_filesystem(struct file_system_type *);
1858 extern int unregister_filesystem(struct file_system_type *);
1859 extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
1860 #define kern_mount(type) kern_mount_data(type, NULL)
1861 extern void kern_unmount(struct vfsmount *mnt);
1862 extern int may_umount_tree(struct vfsmount *);
1863 extern int may_umount(struct vfsmount *);
1864 extern long do_mount(const char *, const char *, const char *, unsigned long, void *);
1865 extern struct vfsmount *collect_mounts(struct path *);
1866 extern void drop_collected_mounts(struct vfsmount *);
1867 extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
1868                           struct vfsmount *);
1869 extern int vfs_statfs(struct path *, struct kstatfs *);
1870 extern int user_statfs(const char __user *, struct kstatfs *);
1871 extern int fd_statfs(int, struct kstatfs *);
1872 extern int vfs_ustat(dev_t, struct kstatfs *);
1873 extern int freeze_super(struct super_block *super);
1874 extern int thaw_super(struct super_block *super);
1875 extern bool our_mnt(struct vfsmount *mnt);
1876 
1877 extern int current_umask(void);
1878 
1879 /* /sys/fs */
1880 extern struct kobject *fs_kobj;
1881 
1882 #define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
1883 extern int rw_verify_area(int, struct file *, loff_t *, size_t);
1884 
1885 #define FLOCK_VERIFY_READ  1
1886 #define FLOCK_VERIFY_WRITE 2
1887 
1888 #ifdef CONFIG_FILE_LOCKING
1889 extern int locks_mandatory_locked(struct inode *);
1890 extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
1891 
1892 /*
1893  * Candidates for mandatory locking have the setgid bit set
1894  * but no group execute bit -  an otherwise meaningless combination.
1895  */
1896 
1897 static inline int __mandatory_lock(struct inode *ino)
1898 {
1899         return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
1900 }
1901 
1902 /*
1903  * ... and these candidates should be on MS_MANDLOCK mounted fs,
1904  * otherwise these will be advisory locks
1905  */
1906 
1907 static inline int mandatory_lock(struct inode *ino)
1908 {
1909         return IS_MANDLOCK(ino) && __mandatory_lock(ino);
1910 }
1911 
1912 static inline int locks_verify_locked(struct inode *inode)
1913 {
1914         if (mandatory_lock(inode))
1915                 return locks_mandatory_locked(inode);
1916         return 0;
1917 }
1918 
1919 static inline int locks_verify_truncate(struct inode *inode,
1920                                     struct file *filp,
1921                                     loff_t size)
1922 {
1923         if (inode->i_flock && mandatory_lock(inode))
1924                 return locks_mandatory_area(
1925                         FLOCK_VERIFY_WRITE, inode, filp,
1926                         size < inode->i_size ? size : inode->i_size,
1927                         (size < inode->i_size ? inode->i_size - size
1928                          : size - inode->i_size)
1929                 );
1930         return 0;
1931 }
1932 
1933 static inline int break_lease(struct inode *inode, unsigned int mode)
1934 {
1935         if (inode->i_flock)
1936                 return __break_lease(inode, mode);
1937         return 0;
1938 }
1939 #else /* !CONFIG_FILE_LOCKING */
1940 static inline int locks_mandatory_locked(struct inode *inode)
1941 {
1942         return 0;
1943 }
1944 
1945 static inline int locks_mandatory_area(int rw, struct inode *inode,
1946                                        struct file *filp, loff_t offset,
1947                                        size_t count)
1948 {
1949         return 0;
1950 }
1951 
1952 static inline int __mandatory_lock(struct inode *inode)
1953 {
1954         return 0;
1955 }
1956 
1957 static inline int mandatory_lock(struct inode *inode)
1958 {
1959         return 0;
1960 }
1961 
1962 static inline int locks_verify_locked(struct inode *inode)
1963 {
1964         return 0;
1965 }
1966 
1967 static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
1968                                         size_t size)
1969 {
1970         return 0;
1971 }
1972 
1973 static inline int break_lease(struct inode *inode, unsigned int mode)
1974 {
1975         return 0;
1976 }
1977 
1978 #endif /* CONFIG_FILE_LOCKING */
1979 
1980 /* fs/open.c */
1981 struct audit_names;
1982 struct filename {
1983         const char              *name;  /* pointer to actual string */
1984         const __user char       *uptr;  /* original userland pointer */
1985         struct audit_names      *aname;
1986         bool                    separate; /* should "name" be freed? */
1987 };
1988 
1989 extern long vfs_truncate(struct path *, loff_t);
1990 extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
1991                        struct file *filp);
1992 extern int do_fallocate(struct file *file, int mode, loff_t offset,
1993                         loff_t len);
1994 extern long do_sys_open(int dfd, const char __user *filename, int flags,
1995                         umode_t mode);
1996 extern struct file *file_open_name(struct filename *, int, umode_t);
1997 extern struct file *filp_open(const char *, int, umode_t);
1998 extern struct file *file_open_root(struct dentry *, struct vfsmount *,
1999                                    const char *, int);
2000 extern struct file * dentry_open(const struct path *, int, const struct cred *);
2001 extern int filp_close(struct file *, fl_owner_t id);
2002 
2003 extern struct filename *getname(const char __user *);
2004 
2005 enum {
2006         FILE_CREATED = 1,
2007         FILE_OPENED = 2
2008 };
2009 extern int finish_open(struct file *file, struct dentry *dentry,
2010                         int (*open)(struct inode *, struct file *),
2011                         int *opened);
2012 extern int finish_no_open(struct file *file, struct dentry *dentry);
2013 
2014 /* fs/ioctl.c */
2015 
2016 extern int ioctl_preallocate(struct file *filp, void __user *argp);
2017 
2018 /* fs/dcache.c */
2019 extern void __init vfs_caches_init_early(void);
2020 extern void __init vfs_caches_init(unsigned long);
2021 
2022 extern struct kmem_cache *names_cachep;
2023 
2024 extern void final_putname(struct filename *name);
2025 
2026 #define __getname()             kmem_cache_alloc(names_cachep, GFP_KERNEL)
2027 #define __putname(name)         kmem_cache_free(names_cachep, (void *)(name))
2028 #ifndef CONFIG_AUDITSYSCALL
2029 #define putname(name)           final_putname(name)
2030 #else
2031 extern void putname(struct filename *name);
2032 #endif
2033 
2034 #ifdef CONFIG_BLOCK
2035 extern int register_blkdev(unsigned int, const char *);
2036 extern void unregister_blkdev(unsigned int, const char *);
2037 extern struct block_device *bdget(dev_t);
2038 extern struct block_device *bdgrab(struct block_device *bdev);
2039 extern void bd_set_size(struct block_device *, loff_t size);
2040 extern void bd_forget(struct inode *inode);
2041 extern void bdput(struct block_device *);
2042 extern void invalidate_bdev(struct block_device *);
2043 extern void iterate_bdevs(void (*)(struct block_device *, void *), void *);
2044 extern int sync_blockdev(struct block_device *bdev);
2045 extern void kill_bdev(struct block_device *);
2046 extern struct super_block *freeze_bdev(struct block_device *);
2047 extern void emergency_thaw_all(void);
2048 extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
2049 extern int fsync_bdev(struct block_device *);
2050 #else
2051 static inline void bd_forget(struct inode *inode) {}
2052 static inline int sync_blockdev(struct block_device *bdev) { return 0; }
2053 static inline void kill_bdev(struct block_device *bdev) {}
2054 static inline void invalidate_bdev(struct block_device *bdev) {}
2055 
2056 static inline struct super_block *freeze_bdev(struct block_device *sb)
2057 {
2058         return NULL;
2059 }
2060 
2061 static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
2062 {
2063         return 0;
2064 }
2065 
2066 static inline void iterate_bdevs(void (*f)(struct block_device *, void *), void *arg)
2067 {
2068 }
2069 #endif
2070 extern int sync_filesystem(struct super_block *);
2071 extern const struct file_operations def_blk_fops;
2072 extern const struct file_operations def_chr_fops;
2073 extern const struct file_operations bad_sock_fops;
2074 #ifdef CONFIG_BLOCK
2075 extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
2076 extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
2077 extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
2078 extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
2079 extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
2080                                                void *holder);
2081 extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
2082                                               void *holder);
2083 extern void blkdev_put(struct block_device *bdev, fmode_t mode);
2084 #ifdef CONFIG_SYSFS
2085 extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
2086 extern void bd_unlink_disk_holder(struct block_device *bdev,
2087                                   struct gendisk *disk);
2088 #else
2089 static inline int bd_link_disk_holder(struct block_device *bdev,
2090                                       struct gendisk *disk)
2091 {
2092         return 0;
2093 }
2094 static inline void bd_unlink_disk_holder(struct block_device *bdev,
2095                                          struct gendisk *disk)
2096 {
2097 }
2098 #endif
2099 #endif
2100 
2101 /* fs/char_dev.c */
2102 #define CHRDEV_MAJOR_HASH_SIZE  255
2103 extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
2104 extern int register_chrdev_region(dev_t, unsigned, const char *);
2105 extern int __register_chrdev(unsigned int major, unsigned int baseminor,
2106                              unsigned int count, const char *name,
2107                              const struct file_operations *fops);
2108 extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
2109                                 unsigned int count, const char *name);
2110 extern void unregister_chrdev_region(dev_t, unsigned);
2111 extern void chrdev_show(struct seq_file *,off_t);
2112 
2113 static inline int register_chrdev(unsigned int major, const char *name,
2114                                   const struct file_operations *fops)
2115 {
2116         return __register_chrdev(major, 0, 256, name, fops);
2117 }
2118 
2119 static inline void unregister_chrdev(unsigned int major, const char *name)
2120 {
2121         __unregister_chrdev(major, 0, 256, name);
2122 }
2123 
2124 /* fs/block_dev.c */
2125 #define BDEVNAME_SIZE   32      /* Largest string for a blockdev identifier */
2126 #define BDEVT_SIZE      10      /* Largest string for MAJ:MIN for blkdev */
2127 
2128 #ifdef CONFIG_BLOCK
2129 #define BLKDEV_MAJOR_HASH_SIZE  255
2130 extern const char *__bdevname(dev_t, char *buffer);
2131 extern const char *bdevname(struct block_device *bdev, char *buffer);
2132 extern struct block_device *lookup_bdev(const char *);
2133 extern void blkdev_show(struct seq_file *,off_t);
2134 
2135 #else
2136 #define BLKDEV_MAJOR_HASH_SIZE  0
2137 #endif
2138 
2139 extern void init_special_inode(struct inode *, umode_t, dev_t);
2140 
2141 /* Invalid inode operations -- fs/bad_inode.c */
2142 extern void make_bad_inode(struct inode *);
2143 extern int is_bad_inode(struct inode *);
2144 
2145 #ifdef CONFIG_BLOCK
2146 /*
2147  * return READ, READA, or WRITE
2148  */
2149 #define bio_rw(bio)             ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2150 
2151 /*
2152  * return data direction, READ or WRITE
2153  */
2154 #define bio_data_dir(bio)       ((bio)->bi_rw & 1)
2155 
2156 extern void check_disk_size_change(struct gendisk *disk,
2157                                    struct block_device *bdev);
2158 extern int revalidate_disk(struct gendisk *);
2159 extern int check_disk_change(struct block_device *);
2160 extern int __invalidate_device(struct block_device *, bool);
2161 extern int invalidate_partition(struct gendisk *, int);
2162 #endif
2163 unsigned long invalidate_mapping_pages(struct address_space *mapping,
2164                                         pgoff_t start, pgoff_t end);
2165 
2166 static inline void invalidate_remote_inode(struct inode *inode)
2167 {
2168         if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
2169             S_ISLNK(inode->i_mode))
2170                 invalidate_mapping_pages(inode->i_mapping, 0, -1);
2171 }
2172 extern int invalidate_inode_pages2(struct address_space *mapping);
2173 extern int invalidate_inode_pages2_range(struct address_space *mapping,
2174                                          pgoff_t start, pgoff_t end);
2175 extern int write_inode_now(struct inode *, int);
2176 extern int filemap_fdatawrite(struct address_space *);
2177 extern int filemap_flush(struct address_space *);
2178 extern int filemap_fdatawait(struct address_space *);
2179 extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
2180                                    loff_t lend);
2181 extern int filemap_write_and_wait(struct address_space *mapping);
2182 extern int filemap_write_and_wait_range(struct address_space *mapping,
2183                                         loff_t lstart, loff_t lend);
2184 extern int __filemap_fdatawrite_range(struct address_space *mapping,
2185                                 loff_t start, loff_t end, int sync_mode);
2186 extern int filemap_fdatawrite_range(struct address_space *mapping,
2187                                 loff_t start, loff_t end);
2188 
2189 extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
2190                            int datasync);
2191 extern int vfs_fsync(struct file *file, int datasync);
2192 extern int generic_write_sync(struct file *file, loff_t pos, loff_t count);
2193 extern void emergency_sync(void);
2194 extern void emergency_remount(void);
2195 #ifdef CONFIG_BLOCK
2196 extern sector_t bmap(struct inode *, sector_t);
2197 #endif
2198 extern int notify_change(struct dentry *, struct iattr *);
2199 extern int inode_permission(struct inode *, int);
2200 extern int generic_permission(struct inode *, int);
2201 
2202 static inline bool execute_ok(struct inode *inode)
2203 {
2204         return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
2205 }
2206 
2207 static inline struct inode *file_inode(struct file *f)
2208 {
2209         return f->f_inode;
2210 }
2211 
2212 static inline void file_start_write(struct file *file)
2213 {
2214         if (!S_ISREG(file_inode(file)->i_mode))
2215                 return;
2216         __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, true);
2217 }
2218 
2219 static inline bool file_start_write_trylock(struct file *file)
2220 {
2221         if (!S_ISREG(file_inode(file)->i_mode))
2222                 return true;
2223         return __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, false);
2224 }
2225 
2226 static inline void file_end_write(struct file *file)
2227 {
2228         if (!S_ISREG(file_inode(file)->i_mode))
2229                 return;
2230         __sb_end_write(file_inode(file)->i_sb, SB_FREEZE_WRITE);
2231 }
2232 
2233 /*
2234  * get_write_access() gets write permission for a file.
2235  * put_write_access() releases this write permission.
2236  * This is used for regular files.
2237  * We cannot support write (and maybe mmap read-write shared) accesses and
2238  * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2239  * can have the following values:
2240  * 0: no writers, no VM_DENYWRITE mappings
2241  * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2242  * > 0: (i_writecount) users are writing to the file.
2243  *
2244  * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2245  * except for the cases where we don't hold i_writecount yet. Then we need to
2246  * use {get,deny}_write_access() - these functions check the sign and refuse
2247  * to do the change if sign is wrong.
2248  */
2249 static inline int get_write_access(struct inode *inode)
2250 {
2251         return atomic_inc_unless_negative(&inode->i_writecount) ? 0 : -ETXTBSY;
2252 }
2253 static inline int deny_write_access(struct file *file)
2254 {
2255         struct inode *inode = file_inode(file);
2256         return atomic_dec_unless_positive(&inode->i_writecount) ? 0 : -ETXTBSY;
2257 }
2258 static inline void put_write_access(struct inode * inode)
2259 {
2260         atomic_dec(&inode->i_writecount);
2261 }
2262 static inline void allow_write_access(struct file *file)
2263 {
2264         if (file)
2265                 atomic_inc(&file_inode(file)->i_writecount);
2266 }
2267 #ifdef CONFIG_IMA
2268 static inline void i_readcount_dec(struct inode *inode)
2269 {
2270         BUG_ON(!atomic_read(&inode->i_readcount));
2271         atomic_dec(&inode->i_readcount);
2272 }
2273 static inline void i_readcount_inc(struct inode *inode)
2274 {
2275         atomic_inc(&inode->i_readcount);
2276 }
2277 #else
2278 static inline void i_readcount_dec(struct inode *inode)
2279 {
2280         return;
2281 }
2282 static inline void i_readcount_inc(struct inode *inode)
2283 {
2284         return;
2285 }
2286 #endif
2287 extern int do_pipe_flags(int *, int);
2288 
2289 extern int kernel_read(struct file *, loff_t, char *, unsigned long);
2290 extern ssize_t kernel_write(struct file *, const char *, size_t, loff_t);
2291 extern struct file * open_exec(const char *);
2292  
2293 /* fs/dcache.c -- generic fs support functions */
2294 extern int is_subdir(struct dentry *, struct dentry *);
2295 extern int path_is_under(struct path *, struct path *);
2296 extern ino_t find_inode_number(struct dentry *, struct qstr *);
2297 
2298 #include <linux/err.h>
2299 
2300 /* needed for stackable file system support */
2301 extern loff_t default_llseek(struct file *file, loff_t offset, int whence);
2302 
2303 extern loff_t vfs_llseek(struct file *file, loff_t offset, int whence);
2304 
2305 extern int inode_init_always(struct super_block *, struct inode *);
2306 extern void inode_init_once(struct inode *);
2307 extern void address_space_init_once(struct address_space *mapping);
2308 extern void ihold(struct inode * inode);
2309 extern void iput(struct inode *);
2310 extern struct inode * igrab(struct inode *);
2311 extern ino_t iunique(struct super_block *, ino_t);
2312 extern int inode_needs_sync(struct inode *inode);
2313 extern int generic_delete_inode(struct inode *inode);
2314 static inline int generic_drop_inode(struct inode *inode)
2315 {
2316         return !inode->i_nlink || inode_unhashed(inode);
2317 }
2318 
2319 extern struct inode *ilookup5_nowait(struct super_block *sb,
2320                 unsigned long hashval, int (*test)(struct inode *, void *),
2321                 void *data);
2322 extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
2323                 int (*test)(struct inode *, void *), void *data);
2324 extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
2325 
2326 extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
2327 extern struct inode * iget_locked(struct super_block *, unsigned long);
2328 extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
2329 extern int insert_inode_locked(struct inode *);
2330 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2331 extern void lockdep_annotate_inode_mutex_key(struct inode *inode);
2332 #else
2333 static inline void lockdep_annotate_inode_mutex_key(struct inode *inode) { };
2334 #endif
2335 extern void unlock_new_inode(struct inode *);
2336 extern unsigned int get_next_ino(void);
2337 
2338 extern void __iget(struct inode * inode);
2339 extern void iget_failed(struct inode *);
2340 extern void clear_inode(struct inode *);
2341 extern void __destroy_inode(struct inode *);
2342 extern struct inode *new_inode_pseudo(struct super_block *sb);
2343 extern struct inode *new_inode(struct super_block *sb);
2344 extern void free_inode_nonrcu(struct inode *inode);
2345 extern int should_remove_suid(struct dentry *);
2346 extern int file_remove_suid(struct file *);
2347 
2348 extern void __insert_inode_hash(struct inode *, unsigned long hashval);
2349 static inline void insert_inode_hash(struct inode *inode)
2350 {
2351         __insert_inode_hash(inode, inode->i_ino);
2352 }
2353 
2354 extern void __remove_inode_hash(struct inode *);
2355 static inline void remove_inode_hash(struct inode *inode)
2356 {
2357         if (!inode_unhashed(inode))
2358                 __remove_inode_hash(inode);
2359 }
2360 
2361 extern void inode_sb_list_add(struct inode *inode);
2362 
2363 #ifdef CONFIG_BLOCK
2364 extern void submit_bio(int, struct bio *);
2365 extern int bdev_read_only(struct block_device *);
2366 #endif
2367 extern int set_blocksize(struct block_device *, int);
2368 extern int sb_set_blocksize(struct super_block *, int);
2369 extern int sb_min_blocksize(struct super_block *, int);
2370 
2371 extern int generic_file_mmap(struct file *, struct vm_area_struct *);
2372 extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
2373 extern int generic_file_remap_pages(struct vm_area_struct *, unsigned long addr,
2374                 unsigned long size, pgoff_t pgoff);
2375 extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
2376 int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
2377 extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2378 extern ssize_t __generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long,
2379                 loff_t *);
2380 extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2381 extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
2382                 unsigned long *, loff_t, loff_t *, size_t, size_t);
2383 extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
2384                 unsigned long, loff_t, loff_t *, size_t, ssize_t);
2385 extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
2386 extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
2387 extern int generic_segment_checks(const struct iovec *iov,
2388                 unsigned long *nr_segs, size_t *count, int access_flags);
2389 
2390 /* fs/block_dev.c */
2391 extern ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
2392                                 unsigned long nr_segs, loff_t pos);
2393 extern int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
2394                         int datasync);
2395 extern void block_sync_page(struct page *page);
2396 
2397 /* fs/splice.c */
2398 extern ssize_t generic_file_splice_read(struct file *, loff_t *,
2399                 struct pipe_inode_info *, size_t, unsigned int);
2400 extern ssize_t default_file_splice_read(struct file *, loff_t *,
2401                 struct pipe_inode_info *, size_t, unsigned int);
2402 extern ssize_t generic_file_splice_write(struct pipe_inode_info *,
2403                 struct file *, loff_t *, size_t, unsigned int);
2404 extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
2405                 struct file *out, loff_t *, size_t len, unsigned int flags);
2406 
2407 extern void
2408 file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
2409 extern loff_t noop_llseek(struct file *file, loff_t offset, int whence);
2410 extern loff_t no_llseek(struct file *file, loff_t offset, int whence);
2411 extern loff_t generic_file_llseek(struct file *file, loff_t offset, int whence);
2412 extern loff_t generic_file_llseek_size(struct file *file, loff_t offset,
2413                 int whence, loff_t maxsize, loff_t eof);
2414 extern int generic_file_open(struct inode * inode, struct file * filp);
2415 extern int nonseekable_open(struct inode * inode, struct file * filp);
2416 
2417 #ifdef CONFIG_FS_XIP
2418 extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
2419                              loff_t *ppos);
2420 extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
2421 extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
2422                               size_t len, loff_t *ppos);
2423 extern int xip_truncate_page(struct address_space *mapping, loff_t from);
2424 #else
2425 static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
2426 {
2427         return 0;
2428 }
2429 #endif
2430 
2431 #ifdef CONFIG_BLOCK
2432 typedef void (dio_submit_t)(int rw, struct bio *bio, struct inode *inode,
2433                             loff_t file_offset);
2434 
2435 enum {
2436         /* need locking between buffered and direct access */
2437         DIO_LOCKING     = 0x01,
2438 
2439         /* filesystem does not support filling holes */
2440         DIO_SKIP_HOLES  = 0x02,
2441 };
2442 
2443 void dio_end_io(struct bio *bio, int error);
2444 
2445 ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
2446         struct block_device *bdev, const struct iovec *iov, loff_t offset,
2447         unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
2448         dio_submit_t submit_io, int flags);
2449 
2450 static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
2451                 struct inode *inode, const struct iovec *iov, loff_t offset,
2452                 unsigned long nr_segs, get_block_t get_block)
2453 {
2454         return __blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
2455                                     offset, nr_segs, get_block, NULL, NULL,
2456                                     DIO_LOCKING | DIO_SKIP_HOLES);
2457 }
2458 #endif
2459 
2460 void inode_dio_wait(struct inode *inode);
2461 void inode_dio_done(struct inode *inode);
2462 
2463 extern const struct file_operations generic_ro_fops;
2464 
2465 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2466 
2467 extern int vfs_readlink(struct dentry *, char __user *, int, const char *);
2468 extern int vfs_follow_link(struct nameidata *, const char *);
2469 extern int page_readlink(struct dentry *, char __user *, int);
2470 extern void *page_follow_link_light(struct dentry *, struct nameidata *);
2471 extern void page_put_link(struct dentry *, struct nameidata *, void *);
2472 extern int __page_symlink(struct inode *inode, const char *symname, int len,
2473                 int nofs);
2474 extern int page_symlink(struct inode *inode, const char *symname, int len);
2475 extern const struct inode_operations page_symlink_inode_operations;
2476 extern int generic_readlink(struct dentry *, char __user *, int);
2477 extern void generic_fillattr(struct inode *, struct kstat *);
2478 extern int vfs_getattr(struct path *, struct kstat *);
2479 void __inode_add_bytes(struct inode *inode, loff_t bytes);
2480 void inode_add_bytes(struct inode *inode, loff_t bytes);
2481 void __inode_sub_bytes(struct inode *inode, loff_t bytes);
2482 void inode_sub_bytes(struct inode *inode, loff_t bytes);
2483 loff_t inode_get_bytes(struct inode *inode);
2484 void inode_set_bytes(struct inode *inode, loff_t bytes);
2485 
2486 extern int vfs_readdir(struct file *, filldir_t, void *);
2487 
2488 extern int vfs_stat(const char __user *, struct kstat *);
2489 extern int vfs_lstat(const char __user *, struct kstat *);
2490 extern int vfs_fstat(unsigned int, struct kstat *);
2491 extern int vfs_fstatat(int , const char __user *, struct kstat *, int);
2492 
2493 extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
2494                     unsigned long arg);
2495 extern int __generic_block_fiemap(struct inode *inode,
2496                                   struct fiemap_extent_info *fieinfo,
2497                                   loff_t start, loff_t len,
2498                                   get_block_t *get_block);
2499 extern int generic_block_fiemap(struct inode *inode,
2500                                 struct fiemap_extent_info *fieinfo, u64 start,
2501                                 u64 len, get_block_t *get_block);
2502 
2503 extern void get_filesystem(struct file_system_type *fs);
2504 extern void put_filesystem(struct file_system_type *fs);
2505 extern struct file_system_type *get_fs_type(const char *name);
2506 extern struct super_block *get_super(struct block_device *);
2507 extern struct super_block *get_super_thawed(struct block_device *);
2508 extern struct super_block *get_active_super(struct block_device *bdev);
2509 extern void drop_super(struct super_block *sb);
2510 extern void iterate_supers(void (*)(struct super_block *, void *), void *);
2511 extern void iterate_supers_type(struct file_system_type *,
2512                                 void (*)(struct super_block *, void *), void *);
2513 
2514 extern int dcache_dir_open(struct inode *, struct file *);
2515 extern int dcache_dir_close(struct inode *, struct file *);
2516 extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
2517 extern int dcache_readdir(struct file *, void *, filldir_t);
2518 extern int simple_setattr(struct dentry *, struct iattr *);
2519 extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2520 extern int simple_statfs(struct dentry *, struct kstatfs *);
2521 extern int simple_open(struct inode *inode, struct file *file);
2522 extern int simple_link(struct dentry *, struct inode *, struct dentry *);
2523 extern int simple_unlink(struct inode *, struct dentry *);
2524 extern int simple_rmdir(struct inode *, struct dentry *);
2525 extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
2526 extern int noop_fsync(struct file *, loff_t, loff_t, int);
2527 extern int simple_empty(struct dentry *);
2528 extern int simple_readpage(struct file *file, struct page *page);
2529 extern int simple_write_begin(struct file *file, struct address_space *mapping,
2530                         loff_t pos, unsigned len, unsigned flags,
2531                         struct page **pagep, void **fsdata);
2532 extern int simple_write_end(struct file *file, struct address_space *mapping,
2533                         loff_t pos, unsigned len, unsigned copied,
2534                         struct page *page, void *fsdata);
2535 
2536 extern struct dentry *simple_lookup(struct inode *, struct dentry *, unsigned int flags);
2537 extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
2538 extern const struct file_operations simple_dir_operations;
2539 extern const struct inode_operations simple_dir_inode_operations;
2540 struct tree_descr { char *name; const struct file_operations *ops; int mode; };
2541 struct dentry *d_alloc_name(struct dentry *, const char *);
2542 extern int simple_fill_super(struct super_block *, unsigned long, struct tree_descr *);
2543 extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
2544 extern void simple_release_fs(struct vfsmount **mount, int *count);
2545 
2546 extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
2547                         loff_t *ppos, const void *from, size_t available);
2548 extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
2549                 const void __user *from, size_t count);
2550 
2551 extern int generic_file_fsync(struct file *, loff_t, loff_t, int);
2552 
2553 extern int generic_check_addressable(unsigned, u64);
2554 
2555 #ifdef CONFIG_MIGRATION
2556 extern int buffer_migrate_page(struct address_space *,
2557                                 struct page *, struct page *,
2558                                 enum migrate_mode);
2559 #else
2560 #define buffer_migrate_page NULL
2561 #endif
2562 
2563 extern int inode_change_ok(const struct inode *, struct iattr *);
2564 extern int inode_newsize_ok(const struct inode *, loff_t offset);
2565 extern void setattr_copy(struct inode *inode, const struct iattr *attr);
2566 
2567 extern int file_update_time(struct file *file);
2568 
2569 extern int generic_show_options(struct seq_file *m, struct dentry *root);
2570 extern void save_mount_options(struct super_block *sb, char *options);
2571 extern void replace_mount_options(struct super_block *sb, char *options);
2572 
2573 static inline ino_t parent_ino(struct dentry *dentry)
2574 {
2575         ino_t res;
2576 
2577         /*
2578          * Don't strictly need d_lock here? If the parent ino could change
2579          * then surely we'd have a deeper race in the caller?
2580          */
2581         spin_lock(&dentry->d_lock);
2582         res = dentry->d_parent->d_inode->i_ino;
2583         spin_unlock(&dentry->d_lock);
2584         return res;
2585 }
2586 
2587 /* Transaction based IO helpers */
2588 
2589 /*
2590  * An argresp is stored in an allocated page and holds the
2591  * size of the argument or response, along with its content
2592  */
2593 struct simple_transaction_argresp {
2594         ssize_t size;
2595         char data[0];
2596 };
2597 
2598 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2599 
2600 char *simple_transaction_get(struct file *file, const char __user *buf,
2601                                 size_t size);
2602 ssize_t simple_transaction_read(struct file *file, char __user *buf,
2603                                 size_t size, loff_t *pos);
2604 int simple_transaction_release(struct inode *inode, struct file *file);
2605 
2606 void simple_transaction_set(struct file *file, size_t n);
2607 
2608 /*
2609  * simple attribute files
2610  *
2611  * These attributes behave similar to those in sysfs:
2612  *
2613  * Writing to an attribute immediately sets a value, an open file can be
2614  * written to multiple times.
2615  *
2616  * Reading from an attribute creates a buffer from the value that might get
2617  * read with multiple read calls. When the attribute has been read
2618  * completely, no further read calls are possible until the file is opened
2619  * again.
2620  *
2621  * All attributes contain a text representation of a numeric value
2622  * that are accessed with the get() and set() functions.
2623  */
2624 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt)            \
2625 static int __fops ## _open(struct inode *inode, struct file *file)      \
2626 {                                                                       \
2627         __simple_attr_check_format(__fmt, 0ull);                        \
2628         return simple_attr_open(inode, file, __get, __set, __fmt);      \
2629 }                                                                       \
2630 static const struct file_operations __fops = {                          \
2631         .owner   = THIS_MODULE,                                         \
2632         .open    = __fops ## _open,                                     \
2633         .release = simple_attr_release,                                 \
2634         .read    = simple_attr_read,                                    \
2635         .write   = simple_attr_write,                                   \
2636         .llseek  = generic_file_llseek,                                 \
2637 };
2638 
2639 static inline __printf(1, 2)
2640 void __simple_attr_check_format(const char *fmt, ...)
2641 {
2642         /* don't do anything, just let the compiler check the arguments; */
2643 }
2644 
2645 int simple_attr_open(struct inode *inode, struct file *file,
2646                      int (*get)(void *, u64 *), int (*set)(void *, u64),
2647                      const char *fmt);
2648 int simple_attr_release(struct inode *inode, struct file *file);
2649 ssize_t simple_attr_read(struct file *file, char __user *buf,
2650                          size_t len, loff_t *ppos);
2651 ssize_t simple_attr_write(struct file *file, const char __user *buf,
2652                           size_t len, loff_t *ppos);
2653 
2654 struct ctl_table;
2655 int proc_nr_files(struct ctl_table *table, int write,
2656                   void __user *buffer, size_t *lenp, loff_t *ppos);
2657 int proc_nr_dentry(struct ctl_table *table, int write,
2658                   void __user *buffer, size_t *lenp, loff_t *ppos);
2659 int proc_nr_inodes(struct ctl_table *table, int write,
2660                    void __user *buffer, size_t *lenp, loff_t *ppos);
2661 int __init get_filesystem_list(char *buf);
2662 
2663 #define __FMODE_EXEC            ((__force int) FMODE_EXEC)
2664 #define __FMODE_NONOTIFY        ((__force int) FMODE_NONOTIFY)
2665 
2666 #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
2667 #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
2668                                             (flag & __FMODE_NONOTIFY)))
2669 
2670 static inline int is_sxid(umode_t mode)
2671 {
2672         return (mode & S_ISUID) || ((mode & S_ISGID) && (mode & S_IXGRP));
2673 }
2674 
2675 static inline void inode_has_no_xattr(struct inode *inode)
2676 {
2677         if (!is_sxid(inode->i_mode) && (inode->i_sb->s_flags & MS_NOSEC))
2678                 inode->i_flags |= S_NOSEC;
2679 }
2680 
2681 #endif /* _LINUX_FS_H */
2682 

~ [ source navigation ] ~ [ diff markup ] ~ [ identifier search ] ~

kernel.org | git.kernel.org | LWN.net | Project Home | Wiki (Japanese) | Wiki (English) | SVN repository | Mail admin

Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.

osdn.jp