1 #ifndef _FS_CEPH_SUPER_H
2 #define _FS_CEPH_SUPER_H
4 #include <linux/ceph/ceph_debug.h>
6 #include <asm/unaligned.h>
7 #include <linux/backing-dev.h>
8 #include <linux/completion.h>
9 #include <linux/exportfs.h>
11 #include <linux/mempool.h>
12 #include <linux/pagemap.h>
13 #include <linux/wait.h>
14 #include <linux/writeback.h>
15 #include <linux/slab.h>
16 #include <linux/posix_acl.h>
17 #include <linux/refcount.h>
19 #include <linux/ceph/libceph.h>
21 #ifdef CONFIG_CEPH_FSCACHE
22 #include <linux/fscache.h>
25 /* f_type in struct statfs */
26 #define CEPH_SUPER_MAGIC 0x00c36400
28 /* large granularity for statfs utilization stats to facilitate
29 * large volume sizes on 32-bit machines. */
30 #define CEPH_BLOCK_SHIFT 22 /* 4 MB */
31 #define CEPH_BLOCK (1 << CEPH_BLOCK_SHIFT)
33 #define CEPH_MOUNT_OPT_DIRSTAT (1<<4) /* `cat dirname` for stats */
34 #define CEPH_MOUNT_OPT_RBYTES (1<<5) /* dir st_bytes = rbytes */
35 #define CEPH_MOUNT_OPT_NOASYNCREADDIR (1<<7) /* no dcache readdir */
36 #define CEPH_MOUNT_OPT_INO32 (1<<8) /* 32 bit inos */
37 #define CEPH_MOUNT_OPT_DCACHE (1<<9) /* use dcache for readdir etc */
38 #define CEPH_MOUNT_OPT_FSCACHE (1<<10) /* use fscache */
39 #define CEPH_MOUNT_OPT_NOPOOLPERM (1<<11) /* no pool permission check */
40 #define CEPH_MOUNT_OPT_MOUNTWAIT (1<<12) /* mount waits if no mds is up */
42 #define CEPH_MOUNT_OPT_DEFAULT CEPH_MOUNT_OPT_DCACHE
44 #define ceph_set_mount_opt(fsc, opt) \
45 (fsc)->mount_options->flags |= CEPH_MOUNT_OPT_##opt;
46 #define ceph_test_mount_opt(fsc, opt) \
47 (!!((fsc)->mount_options->flags & CEPH_MOUNT_OPT_##opt))
49 #define CEPH_RSIZE_DEFAULT (64*1024*1024) /* max read size */
50 #define CEPH_RASIZE_DEFAULT (8192*1024) /* max readahead */
51 #define CEPH_MAX_READDIR_DEFAULT 1024
52 #define CEPH_MAX_READDIR_BYTES_DEFAULT (512*1024)
53 #define CEPH_SNAPDIRNAME_DEFAULT ".snap"
55 struct ceph_mount_options
{
59 int wsize
; /* max write size */
60 int rsize
; /* max read size */
61 int rasize
; /* max readahead */
62 int congestion_kb
; /* max writeback in flight */
63 int caps_wanted_delay_min
, caps_wanted_delay_max
;
64 int cap_release_safety
;
65 int max_readdir
; /* max readdir result (entires) */
66 int max_readdir_bytes
; /* max readdir result (bytes) */
69 * everything above this point can be memcmp'd; everything below
70 * is handled in compare_mount_options()
73 char *snapdir_name
; /* default ".snap" */
74 char *mds_namespace
; /* default NULL */
75 char *server_path
; /* default "/" */
78 struct ceph_fs_client
{
79 struct super_block
*sb
;
81 struct ceph_mount_options
*mount_options
;
82 struct ceph_client
*client
;
84 unsigned long mount_state
;
85 int min_caps
; /* min caps i added */
87 struct ceph_mds_client
*mdsc
;
90 mempool_t
*wb_pagevec_pool
;
91 struct workqueue_struct
*wb_wq
;
92 struct workqueue_struct
*pg_inv_wq
;
93 struct workqueue_struct
*trunc_wq
;
94 atomic_long_t writeback_count
;
96 #ifdef CONFIG_DEBUG_FS
97 struct dentry
*debugfs_dentry_lru
, *debugfs_caps
;
98 struct dentry
*debugfs_congestion_kb
;
99 struct dentry
*debugfs_bdi
;
100 struct dentry
*debugfs_mdsc
, *debugfs_mdsmap
;
101 struct dentry
*debugfs_mds_sessions
;
104 #ifdef CONFIG_CEPH_FSCACHE
105 struct fscache_cookie
*fscache
;
111 * File i/o capability. This tracks shared state with the metadata
112 * server that allows us to cache or writeback attributes or to read
113 * and write data. For any given inode, we should have one or more
114 * capabilities, one issued by each metadata server, and our
115 * cumulative access is the OR of all issued capabilities.
117 * Each cap is referenced by the inode's i_caps rbtree and by per-mds
118 * session capability lists.
121 struct ceph_inode_info
*ci
;
122 struct rb_node ci_node
; /* per-ci cap tree */
123 struct ceph_mds_session
*session
;
124 struct list_head session_caps
; /* per-session caplist */
125 u64 cap_id
; /* unique cap id (mds provided) */
129 int issued
; /* latest, from the mds */
130 int implemented
; /* implemented superset of
131 issued (for revocation) */
134 /* caps to release */
140 u32 seq
, issue_seq
, mseq
;
141 u32 cap_gen
; /* active/stale cycle */
142 unsigned long last_used
;
143 struct list_head caps_item
;
146 #define CHECK_CAPS_NODELAY 1 /* do not delay any further */
147 #define CHECK_CAPS_AUTHONLY 2 /* only check auth cap */
148 #define CHECK_CAPS_FLUSH 4 /* flush any dirty caps */
150 struct ceph_cap_flush
{
152 int caps
; /* 0 means capsnap */
153 bool wake
; /* wake up flush waiters when finish ? */
154 struct list_head g_list
; // global
155 struct list_head i_list
; // per inode
159 * Snapped cap state that is pending flush to mds. When a snapshot occurs,
160 * we first complete any in-process sync writes and writeback any dirty
161 * data before flushing the snapped state (tracked here) back to the MDS.
163 struct ceph_cap_snap
{
165 struct list_head ci_item
;
167 struct ceph_cap_flush cap_flush
;
171 struct ceph_snap_context
*context
;
177 struct ceph_buffer
*xattr_blob
;
181 struct timespec mtime
, atime
, ctime
;
185 int writing
; /* a sync write is still in progress */
186 int dirty_pages
; /* dirty pages awaiting writeback */
191 static inline void ceph_put_cap_snap(struct ceph_cap_snap
*capsnap
)
193 if (refcount_dec_and_test(&capsnap
->nref
)) {
194 if (capsnap
->xattr_blob
)
195 ceph_buffer_put(capsnap
->xattr_blob
);
201 * The frag tree describes how a directory is fragmented, potentially across
202 * multiple metadata servers. It is also used to indicate points where
203 * metadata authority is delegated, and whether/where metadata is replicated.
205 * A _leaf_ frag will be present in the i_fragtree IFF there is
206 * delegation info. That is, if mds >= 0 || ndist > 0.
208 #define CEPH_MAX_DIRFRAG_REP 4
210 struct ceph_inode_frag
{
215 int split_by
; /* i.e. 2^(split_by) children */
217 /* delegation and replication info */
218 int mds
; /* -1 if same authority as parent */
219 int ndist
; /* >0 if replicated */
220 int dist
[CEPH_MAX_DIRFRAG_REP
];
224 * We cache inode xattrs as an encoded blob until they are first used,
225 * at which point we parse them into an rbtree.
227 struct ceph_inode_xattr
{
236 int should_free_name
;
243 struct ceph_dentry_info
{
244 struct ceph_mds_session
*lease_session
;
245 u32 lease_gen
, lease_shared_gen
;
247 unsigned long lease_renew_after
, lease_renew_from
;
248 struct list_head lru
;
249 struct dentry
*dentry
;
254 struct ceph_inode_xattrs_info
{
256 * (still encoded) xattr blob. we avoid the overhead of parsing
257 * this until someone actually calls getxattr, etc.
259 * blob->vec.iov_len == 4 implies there are no xattrs; blob ==
260 * NULL means we don't know.
262 struct ceph_buffer
*blob
, *prealloc_blob
;
264 struct rb_root index
;
269 u64 version
, index_version
;
275 struct ceph_inode_info
{
276 struct ceph_vino i_vino
; /* ceph ino + snap */
278 spinlock_t i_ceph_lock
;
281 u64 i_inline_version
;
284 unsigned i_ceph_flags
;
285 atomic64_t i_release_count
;
286 atomic64_t i_ordered_count
;
287 atomic64_t i_complete_seq
[2];
289 struct ceph_dir_layout i_dir_layout
;
290 struct ceph_file_layout i_layout
;
294 struct timespec i_rctime
;
295 u64 i_rbytes
, i_rfiles
, i_rsubdirs
;
296 u64 i_files
, i_subdirs
;
298 struct rb_root i_fragtree
;
299 int i_fragtree_nsplits
;
300 struct mutex i_fragtree_mutex
;
302 struct ceph_inode_xattrs_info i_xattrs
;
304 /* capabilities. protected _both_ by i_ceph_lock and cap->session's
306 struct rb_root i_caps
; /* cap list */
307 struct ceph_cap
*i_auth_cap
; /* authoritative cap, if any */
308 unsigned i_dirty_caps
, i_flushing_caps
; /* mask of dirtied fields */
309 struct list_head i_dirty_item
, i_flushing_item
;
310 /* we need to track cap writeback on a per-cap-bit basis, to allow
311 * overlapping, pipelined cap flushes to the mds. we can probably
312 * reduce the tid to 8 bits if we're concerned about inode size. */
313 struct ceph_cap_flush
*i_prealloc_cap_flush
;
314 struct list_head i_cap_flush_list
;
315 wait_queue_head_t i_cap_wq
; /* threads waiting on a capability */
316 unsigned long i_hold_caps_min
; /* jiffies */
317 unsigned long i_hold_caps_max
; /* jiffies */
318 struct list_head i_cap_delay_list
; /* for delayed cap release to mds */
319 struct ceph_cap_reservation i_cap_migration_resv
;
320 struct list_head i_cap_snaps
; /* snapped state pending flush to mds */
321 struct ceph_snap_context
*i_head_snapc
; /* set if wr_buffer_head > 0 or
322 dirty|flushing caps */
323 unsigned i_snap_caps
; /* cap bits for snapped files */
325 int i_nr_by_mode
[CEPH_FILE_MODE_BITS
]; /* open file counts */
327 struct mutex i_truncate_mutex
;
328 u32 i_truncate_seq
; /* last truncate to smaller size */
329 u64 i_truncate_size
; /* and the size we last truncated down to */
330 int i_truncate_pending
; /* still need to call vmtruncate */
332 u64 i_max_size
; /* max file size authorized by mds */
333 u64 i_reported_size
; /* (max_)size reported to or requested of mds */
334 u64 i_wanted_max_size
; /* offset we'd like to write too */
335 u64 i_requested_max_size
; /* max_size we've requested */
337 /* held references to caps */
339 int i_rd_ref
, i_rdcache_ref
, i_wr_ref
, i_wb_ref
;
340 int i_wrbuffer_ref
, i_wrbuffer_ref_head
;
341 u32 i_shared_gen
; /* increment each time we get FILE_SHARED */
342 u32 i_rdcache_gen
; /* incremented each time we get FILE_CACHE. */
343 u32 i_rdcache_revoking
; /* RDCACHE gen to async invalidate, if any */
345 struct list_head i_unsafe_dirops
; /* uncommitted mds dir ops */
346 struct list_head i_unsafe_iops
; /* uncommitted mds inode ops */
347 spinlock_t i_unsafe_lock
;
349 struct ceph_snap_realm
*i_snap_realm
; /* snap realm (if caps) */
350 int i_snap_realm_counter
; /* snap realm (if caps) */
351 struct list_head i_snap_realm_item
;
352 struct list_head i_snap_flush_item
;
354 struct work_struct i_wb_work
; /* writeback work */
355 struct work_struct i_pg_inv_work
; /* page invalidation work */
357 struct work_struct i_vmtruncate_work
;
359 #ifdef CONFIG_CEPH_FSCACHE
360 struct fscache_cookie
*fscache
;
363 struct inode vfs_inode
; /* at end */
366 static inline struct ceph_inode_info
*ceph_inode(struct inode
*inode
)
368 return container_of(inode
, struct ceph_inode_info
, vfs_inode
);
371 static inline struct ceph_fs_client
*ceph_inode_to_client(struct inode
*inode
)
373 return (struct ceph_fs_client
*)inode
->i_sb
->s_fs_info
;
376 static inline struct ceph_fs_client
*ceph_sb_to_client(struct super_block
*sb
)
378 return (struct ceph_fs_client
*)sb
->s_fs_info
;
381 static inline struct ceph_vino
ceph_vino(struct inode
*inode
)
383 return ceph_inode(inode
)->i_vino
;
387 * ino_t is <64 bits on many architectures, blech.
389 * i_ino (kernel inode) st_ino (userspace)
394 static inline u32
ceph_ino_to_ino32(__u64 vino
)
396 u32 ino
= vino
& 0xffffffff;
406 static inline ino_t
ceph_vino_to_ino(struct ceph_vino vino
)
408 #if BITS_PER_LONG == 32
409 return ceph_ino_to_ino32(vino
.ino
);
411 return (ino_t
)vino
.ino
;
416 * user-visible ino (stat, filldir)
418 #if BITS_PER_LONG == 32
419 static inline ino_t
ceph_translate_ino(struct super_block
*sb
, ino_t ino
)
424 static inline ino_t
ceph_translate_ino(struct super_block
*sb
, ino_t ino
)
426 if (ceph_test_mount_opt(ceph_sb_to_client(sb
), INO32
))
427 ino
= ceph_ino_to_ino32(ino
);
433 /* for printf-style formatting */
434 #define ceph_vinop(i) ceph_inode(i)->i_vino.ino, ceph_inode(i)->i_vino.snap
436 static inline u64
ceph_ino(struct inode
*inode
)
438 return ceph_inode(inode
)->i_vino
.ino
;
440 static inline u64
ceph_snap(struct inode
*inode
)
442 return ceph_inode(inode
)->i_vino
.snap
;
445 static inline int ceph_ino_compare(struct inode
*inode
, void *data
)
447 struct ceph_vino
*pvino
= (struct ceph_vino
*)data
;
448 struct ceph_inode_info
*ci
= ceph_inode(inode
);
449 return ci
->i_vino
.ino
== pvino
->ino
&&
450 ci
->i_vino
.snap
== pvino
->snap
;
453 static inline struct inode
*ceph_find_inode(struct super_block
*sb
,
454 struct ceph_vino vino
)
456 ino_t t
= ceph_vino_to_ino(vino
);
457 return ilookup5(sb
, t
, ceph_ino_compare
, &vino
);
464 #define CEPH_I_DIR_ORDERED (1 << 0) /* dentries in dir are ordered */
465 #define CEPH_I_NODELAY (1 << 1) /* do not delay cap release */
466 #define CEPH_I_FLUSH (1 << 2) /* do not delay flush of dirty metadata */
467 #define CEPH_I_NOFLUSH (1 << 3) /* do not flush dirty caps */
468 #define CEPH_I_POOL_PERM (1 << 4) /* pool rd/wr bits are valid */
469 #define CEPH_I_POOL_RD (1 << 5) /* can read from pool */
470 #define CEPH_I_POOL_WR (1 << 6) /* can write to pool */
471 #define CEPH_I_SEC_INITED (1 << 7) /* security initialized */
472 #define CEPH_I_CAP_DROPPED (1 << 8) /* caps were forcibly dropped */
473 #define CEPH_I_KICK_FLUSH (1 << 9) /* kick flushing caps */
474 #define CEPH_I_FLUSH_SNAPS (1 << 10) /* need flush snapss */
475 #define CEPH_I_ERROR_WRITE (1 << 11) /* have seen write errors */
478 * We set the ERROR_WRITE bit when we start seeing write errors on an inode
479 * and then clear it when they start succeeding. Note that we do a lockless
480 * check first, and only take the lock if it looks like it needs to be changed.
481 * The write submission code just takes this as a hint, so we're not too
482 * worried if a few slip through in either direction.
484 static inline void ceph_set_error_write(struct ceph_inode_info
*ci
)
486 if (!(READ_ONCE(ci
->i_ceph_flags
) & CEPH_I_ERROR_WRITE
)) {
487 spin_lock(&ci
->i_ceph_lock
);
488 ci
->i_ceph_flags
|= CEPH_I_ERROR_WRITE
;
489 spin_unlock(&ci
->i_ceph_lock
);
493 static inline void ceph_clear_error_write(struct ceph_inode_info
*ci
)
495 if (READ_ONCE(ci
->i_ceph_flags
) & CEPH_I_ERROR_WRITE
) {
496 spin_lock(&ci
->i_ceph_lock
);
497 ci
->i_ceph_flags
&= ~CEPH_I_ERROR_WRITE
;
498 spin_unlock(&ci
->i_ceph_lock
);
502 static inline void __ceph_dir_set_complete(struct ceph_inode_info
*ci
,
503 long long release_count
,
504 long long ordered_count
)
506 smp_mb__before_atomic();
507 atomic64_set(&ci
->i_complete_seq
[0], release_count
);
508 atomic64_set(&ci
->i_complete_seq
[1], ordered_count
);
511 static inline void __ceph_dir_clear_complete(struct ceph_inode_info
*ci
)
513 atomic64_inc(&ci
->i_release_count
);
516 static inline void __ceph_dir_clear_ordered(struct ceph_inode_info
*ci
)
518 atomic64_inc(&ci
->i_ordered_count
);
521 static inline bool __ceph_dir_is_complete(struct ceph_inode_info
*ci
)
523 return atomic64_read(&ci
->i_complete_seq
[0]) ==
524 atomic64_read(&ci
->i_release_count
);
527 static inline bool __ceph_dir_is_complete_ordered(struct ceph_inode_info
*ci
)
529 return atomic64_read(&ci
->i_complete_seq
[0]) ==
530 atomic64_read(&ci
->i_release_count
) &&
531 atomic64_read(&ci
->i_complete_seq
[1]) ==
532 atomic64_read(&ci
->i_ordered_count
);
535 static inline void ceph_dir_clear_complete(struct inode
*inode
)
537 __ceph_dir_clear_complete(ceph_inode(inode
));
540 static inline void ceph_dir_clear_ordered(struct inode
*inode
)
542 __ceph_dir_clear_ordered(ceph_inode(inode
));
545 static inline bool ceph_dir_is_complete_ordered(struct inode
*inode
)
547 bool ret
= __ceph_dir_is_complete_ordered(ceph_inode(inode
));
552 /* find a specific frag @f */
553 extern struct ceph_inode_frag
*__ceph_find_frag(struct ceph_inode_info
*ci
,
557 * choose fragment for value @v. copy frag content to pfrag, if leaf
560 extern u32
ceph_choose_frag(struct ceph_inode_info
*ci
, u32 v
,
561 struct ceph_inode_frag
*pfrag
,
564 static inline struct ceph_dentry_info
*ceph_dentry(struct dentry
*dentry
)
566 return (struct ceph_dentry_info
*)dentry
->d_fsdata
;
572 static inline bool __ceph_is_any_real_caps(struct ceph_inode_info
*ci
)
574 return !RB_EMPTY_ROOT(&ci
->i_caps
);
577 extern int __ceph_caps_issued(struct ceph_inode_info
*ci
, int *implemented
);
578 extern int __ceph_caps_issued_mask(struct ceph_inode_info
*ci
, int mask
, int t
);
579 extern int __ceph_caps_issued_other(struct ceph_inode_info
*ci
,
580 struct ceph_cap
*cap
);
582 static inline int ceph_caps_issued(struct ceph_inode_info
*ci
)
585 spin_lock(&ci
->i_ceph_lock
);
586 issued
= __ceph_caps_issued(ci
, NULL
);
587 spin_unlock(&ci
->i_ceph_lock
);
591 static inline int ceph_caps_issued_mask(struct ceph_inode_info
*ci
, int mask
,
595 spin_lock(&ci
->i_ceph_lock
);
596 r
= __ceph_caps_issued_mask(ci
, mask
, touch
);
597 spin_unlock(&ci
->i_ceph_lock
);
601 static inline int __ceph_caps_dirty(struct ceph_inode_info
*ci
)
603 return ci
->i_dirty_caps
| ci
->i_flushing_caps
;
605 extern struct ceph_cap_flush
*ceph_alloc_cap_flush(void);
606 extern void ceph_free_cap_flush(struct ceph_cap_flush
*cf
);
607 extern int __ceph_mark_dirty_caps(struct ceph_inode_info
*ci
, int mask
,
608 struct ceph_cap_flush
**pcf
);
610 extern int __ceph_caps_revoking_other(struct ceph_inode_info
*ci
,
611 struct ceph_cap
*ocap
, int mask
);
612 extern int ceph_caps_revoking(struct ceph_inode_info
*ci
, int mask
);
613 extern int __ceph_caps_used(struct ceph_inode_info
*ci
);
615 extern int __ceph_caps_file_wanted(struct ceph_inode_info
*ci
);
618 * wanted, by virtue of open file modes AND cap refs (buffered/cached data)
620 static inline int __ceph_caps_wanted(struct ceph_inode_info
*ci
)
622 int w
= __ceph_caps_file_wanted(ci
) | __ceph_caps_used(ci
);
623 if (w
& CEPH_CAP_FILE_BUFFER
)
624 w
|= CEPH_CAP_FILE_EXCL
; /* we want EXCL if dirty data */
628 /* what the mds thinks we want */
629 extern int __ceph_caps_mds_wanted(struct ceph_inode_info
*ci
, bool check
);
631 extern void ceph_caps_init(struct ceph_mds_client
*mdsc
);
632 extern void ceph_caps_finalize(struct ceph_mds_client
*mdsc
);
633 extern void ceph_adjust_min_caps(struct ceph_mds_client
*mdsc
, int delta
);
634 extern void ceph_reserve_caps(struct ceph_mds_client
*mdsc
,
635 struct ceph_cap_reservation
*ctx
, int need
);
636 extern int ceph_unreserve_caps(struct ceph_mds_client
*mdsc
,
637 struct ceph_cap_reservation
*ctx
);
638 extern void ceph_reservation_status(struct ceph_fs_client
*client
,
639 int *total
, int *avail
, int *used
,
640 int *reserved
, int *min
);
645 * we keep buffered readdir results attached to file->private_data
647 #define CEPH_F_SYNC 1
648 #define CEPH_F_ATEND 2
650 struct ceph_file_info
{
651 short fmode
; /* initialized on open */
652 short flags
; /* CEPH_F_* */
654 /* readdir: position within the dir */
656 struct ceph_mds_request
*last_readdir
;
658 /* readdir: position within a frag */
659 unsigned next_offset
; /* offset of next chunk (last_name's + 1) */
660 char *last_name
; /* last entry in previous chunk */
661 long long dir_release_count
;
662 long long dir_ordered_count
;
663 int readdir_cache_idx
;
665 /* used for -o dirstat read() on directory thing */
670 struct ceph_readdir_cache_control
{
672 struct dentry
**dentries
;
677 * A "snap realm" describes a subset of the file hierarchy sharing
678 * the same set of snapshots that apply to it. The realms themselves
679 * are organized into a hierarchy, such that children inherit (some of)
680 * the snapshots of their parents.
682 * All inodes within the realm that have capabilities are linked into a
685 struct ceph_snap_realm
{
692 u64 parent_since
; /* snapid when our current parent became so */
694 u64
*prior_parent_snaps
; /* snaps inherited from any parents we */
695 u32 num_prior_parent_snaps
; /* had prior to parent_since */
696 u64
*snaps
; /* snaps specific to this realm */
699 struct ceph_snap_realm
*parent
;
700 struct list_head children
; /* list of child realms */
701 struct list_head child_item
;
703 struct list_head empty_item
; /* if i have ref==0 */
705 struct list_head dirty_item
; /* if realm needs new context */
707 /* the current set of snaps for this realm */
708 struct ceph_snap_context
*cached_context
;
710 struct list_head inodes_with_caps
;
711 spinlock_t inodes_with_caps_lock
;
714 static inline int default_congestion_kb(void)
721 * congestion size, scale with available memory.
733 * This allows larger machines to have larger/more transfers.
734 * Limit the default to 256M
736 congestion_kb
= (16*int_sqrt(totalram_pages
)) << (PAGE_SHIFT
-10);
737 if (congestion_kb
> 256*1024)
738 congestion_kb
= 256*1024;
740 return congestion_kb
;
746 struct ceph_snap_realm
*ceph_lookup_snap_realm(struct ceph_mds_client
*mdsc
,
748 extern void ceph_get_snap_realm(struct ceph_mds_client
*mdsc
,
749 struct ceph_snap_realm
*realm
);
750 extern void ceph_put_snap_realm(struct ceph_mds_client
*mdsc
,
751 struct ceph_snap_realm
*realm
);
752 extern int ceph_update_snap_trace(struct ceph_mds_client
*m
,
753 void *p
, void *e
, bool deletion
,
754 struct ceph_snap_realm
**realm_ret
);
755 extern void ceph_handle_snap(struct ceph_mds_client
*mdsc
,
756 struct ceph_mds_session
*session
,
757 struct ceph_msg
*msg
);
758 extern void ceph_queue_cap_snap(struct ceph_inode_info
*ci
);
759 extern int __ceph_finish_cap_snap(struct ceph_inode_info
*ci
,
760 struct ceph_cap_snap
*capsnap
);
761 extern void ceph_cleanup_empty_realms(struct ceph_mds_client
*mdsc
);
764 * a cap_snap is "pending" if it is still awaiting an in-progress
765 * sync write (that may/may not still update size, mtime, etc.).
767 static inline bool __ceph_have_pending_cap_snap(struct ceph_inode_info
*ci
)
769 return !list_empty(&ci
->i_cap_snaps
) &&
770 list_last_entry(&ci
->i_cap_snaps
, struct ceph_cap_snap
,
775 extern const struct inode_operations ceph_file_iops
;
777 extern struct inode
*ceph_alloc_inode(struct super_block
*sb
);
778 extern void ceph_destroy_inode(struct inode
*inode
);
779 extern int ceph_drop_inode(struct inode
*inode
);
781 extern struct inode
*ceph_get_inode(struct super_block
*sb
,
782 struct ceph_vino vino
);
783 extern struct inode
*ceph_get_snapdir(struct inode
*parent
);
784 extern int ceph_fill_file_size(struct inode
*inode
, int issued
,
785 u32 truncate_seq
, u64 truncate_size
, u64 size
);
786 extern void ceph_fill_file_time(struct inode
*inode
, int issued
,
787 u64 time_warp_seq
, struct timespec
*ctime
,
788 struct timespec
*mtime
, struct timespec
*atime
);
789 extern int ceph_fill_trace(struct super_block
*sb
,
790 struct ceph_mds_request
*req
);
791 extern int ceph_readdir_prepopulate(struct ceph_mds_request
*req
,
792 struct ceph_mds_session
*session
);
794 extern int ceph_inode_holds_cap(struct inode
*inode
, int mask
);
796 extern bool ceph_inode_set_size(struct inode
*inode
, loff_t size
);
797 extern void __ceph_do_pending_vmtruncate(struct inode
*inode
);
798 extern void ceph_queue_vmtruncate(struct inode
*inode
);
800 extern void ceph_queue_invalidate(struct inode
*inode
);
801 extern void ceph_queue_writeback(struct inode
*inode
);
803 extern int __ceph_do_getattr(struct inode
*inode
, struct page
*locked_page
,
804 int mask
, bool force
);
805 static inline int ceph_do_getattr(struct inode
*inode
, int mask
, bool force
)
807 return __ceph_do_getattr(inode
, NULL
, mask
, force
);
809 extern int ceph_permission(struct inode
*inode
, int mask
);
810 extern int __ceph_setattr(struct inode
*inode
, struct iattr
*attr
);
811 extern int ceph_setattr(struct dentry
*dentry
, struct iattr
*attr
);
812 extern int ceph_getattr(const struct path
*path
, struct kstat
*stat
,
813 u32 request_mask
, unsigned int flags
);
816 int __ceph_setxattr(struct inode
*, const char *, const void *, size_t, int);
817 ssize_t
__ceph_getxattr(struct inode
*, const char *, void *, size_t);
818 extern ssize_t
ceph_listxattr(struct dentry
*, char *, size_t);
819 extern void __ceph_build_xattrs_blob(struct ceph_inode_info
*ci
);
820 extern void __ceph_destroy_xattrs(struct ceph_inode_info
*ci
);
821 extern void __init
ceph_xattr_init(void);
822 extern void ceph_xattr_exit(void);
823 extern const struct xattr_handler
*ceph_xattr_handlers
[];
825 #ifdef CONFIG_SECURITY
826 extern bool ceph_security_xattr_deadlock(struct inode
*in
);
827 extern bool ceph_security_xattr_wanted(struct inode
*in
);
829 static inline bool ceph_security_xattr_deadlock(struct inode
*in
)
833 static inline bool ceph_security_xattr_wanted(struct inode
*in
)
840 struct ceph_acls_info
{
843 struct ceph_pagelist
*pagelist
;
846 #ifdef CONFIG_CEPH_FS_POSIX_ACL
848 struct posix_acl
*ceph_get_acl(struct inode
*, int);
849 int ceph_set_acl(struct inode
*inode
, struct posix_acl
*acl
, int type
);
850 int ceph_pre_init_acls(struct inode
*dir
, umode_t
*mode
,
851 struct ceph_acls_info
*info
);
852 void ceph_init_inode_acls(struct inode
*inode
, struct ceph_acls_info
*info
);
853 void ceph_release_acls_info(struct ceph_acls_info
*info
);
855 static inline void ceph_forget_all_cached_acls(struct inode
*inode
)
857 forget_all_cached_acls(inode
);
862 #define ceph_get_acl NULL
863 #define ceph_set_acl NULL
865 static inline int ceph_pre_init_acls(struct inode
*dir
, umode_t
*mode
,
866 struct ceph_acls_info
*info
)
870 static inline void ceph_init_inode_acls(struct inode
*inode
,
871 struct ceph_acls_info
*info
)
874 static inline void ceph_release_acls_info(struct ceph_acls_info
*info
)
877 static inline int ceph_acl_chmod(struct dentry
*dentry
, struct inode
*inode
)
882 static inline void ceph_forget_all_cached_acls(struct inode
*inode
)
889 extern const char *ceph_cap_string(int c
);
890 extern void ceph_handle_caps(struct ceph_mds_session
*session
,
891 struct ceph_msg
*msg
);
892 extern struct ceph_cap
*ceph_get_cap(struct ceph_mds_client
*mdsc
,
893 struct ceph_cap_reservation
*ctx
);
894 extern void ceph_add_cap(struct inode
*inode
,
895 struct ceph_mds_session
*session
, u64 cap_id
,
896 int fmode
, unsigned issued
, unsigned wanted
,
897 unsigned cap
, unsigned seq
, u64 realmino
, int flags
,
898 struct ceph_cap
**new_cap
);
899 extern void __ceph_remove_cap(struct ceph_cap
*cap
, bool queue_release
);
900 extern void ceph_put_cap(struct ceph_mds_client
*mdsc
,
901 struct ceph_cap
*cap
);
902 extern int ceph_is_any_caps(struct inode
*inode
);
904 extern void ceph_queue_caps_release(struct inode
*inode
);
905 extern int ceph_write_inode(struct inode
*inode
, struct writeback_control
*wbc
);
906 extern int ceph_fsync(struct file
*file
, loff_t start
, loff_t end
,
908 extern void ceph_early_kick_flushing_caps(struct ceph_mds_client
*mdsc
,
909 struct ceph_mds_session
*session
);
910 extern void ceph_kick_flushing_caps(struct ceph_mds_client
*mdsc
,
911 struct ceph_mds_session
*session
);
912 extern struct ceph_cap
*ceph_get_cap_for_mds(struct ceph_inode_info
*ci
,
914 extern int ceph_get_cap_mds(struct inode
*inode
);
915 extern void ceph_get_cap_refs(struct ceph_inode_info
*ci
, int caps
);
916 extern void ceph_put_cap_refs(struct ceph_inode_info
*ci
, int had
);
917 extern void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info
*ci
, int nr
,
918 struct ceph_snap_context
*snapc
);
919 extern void ceph_flush_snaps(struct ceph_inode_info
*ci
,
920 struct ceph_mds_session
**psession
);
921 extern bool __ceph_should_report_size(struct ceph_inode_info
*ci
);
922 extern void ceph_check_caps(struct ceph_inode_info
*ci
, int flags
,
923 struct ceph_mds_session
*session
);
924 extern void ceph_check_delayed_caps(struct ceph_mds_client
*mdsc
);
925 extern void ceph_flush_dirty_caps(struct ceph_mds_client
*mdsc
);
927 extern int ceph_encode_inode_release(void **p
, struct inode
*inode
,
928 int mds
, int drop
, int unless
, int force
);
929 extern int ceph_encode_dentry_release(void **p
, struct dentry
*dn
,
931 int mds
, int drop
, int unless
);
933 extern int ceph_get_caps(struct ceph_inode_info
*ci
, int need
, int want
,
934 loff_t endoff
, int *got
, struct page
**pinned_page
);
935 extern int ceph_try_get_caps(struct ceph_inode_info
*ci
,
936 int need
, int want
, int *got
);
938 /* for counting open files by mode */
939 extern void __ceph_get_fmode(struct ceph_inode_info
*ci
, int mode
);
940 extern void ceph_put_fmode(struct ceph_inode_info
*ci
, int mode
);
943 extern const struct address_space_operations ceph_aops
;
944 extern int ceph_mmap(struct file
*file
, struct vm_area_struct
*vma
);
945 extern int ceph_uninline_data(struct file
*filp
, struct page
*locked_page
);
946 extern int ceph_pool_perm_check(struct ceph_inode_info
*ci
, int need
);
947 extern void ceph_pool_perm_destroy(struct ceph_mds_client
* mdsc
);
950 extern const struct file_operations ceph_file_fops
;
952 extern int ceph_renew_caps(struct inode
*inode
);
953 extern int ceph_open(struct inode
*inode
, struct file
*file
);
954 extern int ceph_atomic_open(struct inode
*dir
, struct dentry
*dentry
,
955 struct file
*file
, unsigned flags
, umode_t mode
,
957 extern int ceph_release(struct inode
*inode
, struct file
*filp
);
958 extern void ceph_fill_inline_data(struct inode
*inode
, struct page
*locked_page
,
959 char *data
, size_t len
);
962 extern const struct file_operations ceph_dir_fops
;
963 extern const struct file_operations ceph_snapdir_fops
;
964 extern const struct inode_operations ceph_dir_iops
;
965 extern const struct inode_operations ceph_snapdir_iops
;
966 extern const struct dentry_operations ceph_dentry_ops
;
968 extern loff_t
ceph_make_fpos(unsigned high
, unsigned off
, bool hash_order
);
969 extern int ceph_handle_notrace_create(struct inode
*dir
, struct dentry
*dentry
);
970 extern int ceph_handle_snapdir(struct ceph_mds_request
*req
,
971 struct dentry
*dentry
, int err
);
972 extern struct dentry
*ceph_finish_lookup(struct ceph_mds_request
*req
,
973 struct dentry
*dentry
, int err
);
975 extern void ceph_dentry_lru_add(struct dentry
*dn
);
976 extern void ceph_dentry_lru_touch(struct dentry
*dn
);
977 extern void ceph_dentry_lru_del(struct dentry
*dn
);
978 extern void ceph_invalidate_dentry_lease(struct dentry
*dentry
);
979 extern unsigned ceph_dentry_hash(struct inode
*dir
, struct dentry
*dn
);
980 extern void ceph_readdir_cache_release(struct ceph_readdir_cache_control
*ctl
);
983 extern long ceph_ioctl(struct file
*file
, unsigned int cmd
, unsigned long arg
);
986 extern const struct export_operations ceph_export_ops
;
989 extern __init
void ceph_flock_init(void);
990 extern int ceph_lock(struct file
*file
, int cmd
, struct file_lock
*fl
);
991 extern int ceph_flock(struct file
*file
, int cmd
, struct file_lock
*fl
);
992 extern void ceph_count_locks(struct inode
*inode
, int *p_num
, int *f_num
);
993 extern int ceph_encode_locks_to_buffer(struct inode
*inode
,
994 struct ceph_filelock
*flocks
,
996 int num_flock_locks
);
997 extern int ceph_locks_to_pagelist(struct ceph_filelock
*flocks
,
998 struct ceph_pagelist
*pagelist
,
999 int num_fcntl_locks
, int num_flock_locks
);
1000 extern int lock_to_ceph_filelock(struct file_lock
*fl
, struct ceph_filelock
*c
);
1003 extern int ceph_fs_debugfs_init(struct ceph_fs_client
*client
);
1004 extern void ceph_fs_debugfs_cleanup(struct ceph_fs_client
*client
);
1006 #endif /* _FS_CEPH_SUPER_H */