1 #ifndef _FS_CEPH_SUPER_H
2 #define _FS_CEPH_SUPER_H
4 #include "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>
18 #include "messenger.h"
20 #include "mon_client.h"
21 #include "mds_client.h"
22 #include "osd_client.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 20 /* 1 MB */
31 #define CEPH_BLOCK (1 << CEPH_BLOCK_SHIFT)
36 #define CEPH_OPT_FSID (1<<0)
37 #define CEPH_OPT_NOSHARE (1<<1) /* don't share client with other sbs */
38 #define CEPH_OPT_MYIP (1<<2) /* specified my ip */
39 #define CEPH_OPT_DIRSTAT (1<<4) /* funky `cat dirname` for stats */
40 #define CEPH_OPT_RBYTES (1<<5) /* dir st_bytes = rbytes */
41 #define CEPH_OPT_NOCRC (1<<6) /* no data crc on writes */
42 #define CEPH_OPT_NOASYNCREADDIR (1<<7) /* no dcache readdir */
44 #define CEPH_OPT_DEFAULT (CEPH_OPT_RBYTES)
46 #define ceph_set_opt(client, opt) \
47 (client)->mount_args->flags |= CEPH_OPT_##opt;
48 #define ceph_test_opt(client, opt) \
49 (!!((client)->mount_args->flags & CEPH_OPT_##opt))
52 struct ceph_mount_args
{
55 struct ceph_entity_addr
*mon_addr
;
59 int caps_wanted_delay_min
, caps_wanted_delay_max
;
60 struct ceph_fsid fsid
;
61 struct ceph_entity_addr my_addr
;
63 int rsize
; /* max readahead */
64 int max_readdir
; /* max readdir size */
65 int congestion_kb
; /* max readdir size */
67 int osd_keepalive_timeout
;
68 char *snapdir_name
; /* default ".snap" */
71 int cap_release_safety
;
77 #define CEPH_MOUNT_TIMEOUT_DEFAULT 60
78 #define CEPH_OSD_TIMEOUT_DEFAULT 60 /* seconds */
79 #define CEPH_OSD_KEEPALIVE_DEFAULT 5
80 #define CEPH_OSD_IDLE_TTL_DEFAULT 60
81 #define CEPH_MOUNT_RSIZE_DEFAULT (512*1024) /* readahead */
83 #define CEPH_MSG_MAX_FRONT_LEN (16*1024*1024)
84 #define CEPH_MSG_MAX_DATA_LEN (16*1024*1024)
86 #define CEPH_SNAPDIRNAME_DEFAULT ".snap"
87 #define CEPH_AUTH_NAME_DEFAULT "guest"
90 * Delay telling the MDS we no longer want caps, in case we reopen
91 * the file. Delay a minimum amount of time, even if we send a cap
92 * message for some other reason. Otherwise, take the oppotunity to
93 * update the mds to avoid sending another message later.
95 #define CEPH_CAPS_WANTED_DELAY_MIN_DEFAULT 5 /* cap release delay */
96 #define CEPH_CAPS_WANTED_DELAY_MAX_DEFAULT 60 /* cap release delay */
103 CEPH_MOUNT_UNMOUNTING
,
104 CEPH_MOUNT_UNMOUNTED
,
111 static inline unsigned long time_sub(unsigned long a
, unsigned long b
)
113 BUG_ON(time_after(b
, a
));
114 return (long)a
- (long)b
;
118 * per-filesystem client state
120 * possibly shared by multiple mount points, if they are
121 * mounting the same ceph filesystem/cluster.
124 struct ceph_fsid fsid
;
127 struct mutex mount_mutex
; /* serialize mount attempts */
128 struct ceph_mount_args
*mount_args
;
130 struct super_block
*sb
;
132 unsigned long mount_state
;
133 wait_queue_head_t auth_wq
;
137 int min_caps
; /* min caps i added */
139 struct ceph_messenger
*msgr
; /* messenger instance */
140 struct ceph_mon_client monc
;
141 struct ceph_mds_client mdsc
;
142 struct ceph_osd_client osdc
;
145 mempool_t
*wb_pagevec_pool
;
146 struct workqueue_struct
*wb_wq
;
147 struct workqueue_struct
*pg_inv_wq
;
148 struct workqueue_struct
*trunc_wq
;
149 atomic_long_t writeback_count
;
151 struct backing_dev_info backing_dev_info
;
153 #ifdef CONFIG_DEBUG_FS
154 struct dentry
*debugfs_monmap
;
155 struct dentry
*debugfs_mdsmap
, *debugfs_osdmap
;
156 struct dentry
*debugfs_dir
, *debugfs_dentry_lru
, *debugfs_caps
;
157 struct dentry
*debugfs_congestion_kb
;
158 struct dentry
*debugfs_bdi
;
162 static inline struct ceph_client
*ceph_client(struct super_block
*sb
)
164 return sb
->s_fs_info
;
169 * File i/o capability. This tracks shared state with the metadata
170 * server that allows us to cache or writeback attributes or to read
171 * and write data. For any given inode, we should have one or more
172 * capabilities, one issued by each metadata server, and our
173 * cumulative access is the OR of all issued capabilities.
175 * Each cap is referenced by the inode's i_caps rbtree and by per-mds
176 * session capability lists.
179 struct ceph_inode_info
*ci
;
180 struct rb_node ci_node
; /* per-ci cap tree */
181 struct ceph_mds_session
*session
;
182 struct list_head session_caps
; /* per-session caplist */
184 u64 cap_id
; /* unique cap id (mds provided) */
185 int issued
; /* latest, from the mds */
186 int implemented
; /* implemented superset of issued (for revocation) */
188 u32 seq
, issue_seq
, mseq
;
189 u32 cap_gen
; /* active/stale cycle */
190 unsigned long last_used
;
191 struct list_head caps_item
;
194 #define CHECK_CAPS_NODELAY 1 /* do not delay any further */
195 #define CHECK_CAPS_AUTHONLY 2 /* only check auth cap */
196 #define CHECK_CAPS_FLUSH 4 /* flush any dirty caps */
199 * Snapped cap state that is pending flush to mds. When a snapshot occurs,
200 * we first complete any in-process sync writes and writeback any dirty
201 * data before flushing the snapped state (tracked here) back to the MDS.
203 struct ceph_cap_snap
{
205 struct ceph_inode_info
*ci
;
206 struct list_head ci_item
, flushing_item
;
208 u64 follows
, flush_tid
;
210 struct ceph_snap_context
*context
;
221 struct timespec mtime
, atime
, ctime
;
223 int writing
; /* a sync write is still in progress */
224 int dirty_pages
; /* dirty pages awaiting writeback */
227 static inline void ceph_put_cap_snap(struct ceph_cap_snap
*capsnap
)
229 if (atomic_dec_and_test(&capsnap
->nref
))
234 * The frag tree describes how a directory is fragmented, potentially across
235 * multiple metadata servers. It is also used to indicate points where
236 * metadata authority is delegated, and whether/where metadata is replicated.
238 * A _leaf_ frag will be present in the i_fragtree IFF there is
239 * delegation info. That is, if mds >= 0 || ndist > 0.
241 #define CEPH_MAX_DIRFRAG_REP 4
243 struct ceph_inode_frag
{
248 int split_by
; /* i.e. 2^(split_by) children */
250 /* delegation and replication info */
251 int mds
; /* -1 if same authority as parent */
252 int ndist
; /* >0 if replicated */
253 int dist
[CEPH_MAX_DIRFRAG_REP
];
257 * We cache inode xattrs as an encoded blob until they are first used,
258 * at which point we parse them into an rbtree.
260 struct ceph_inode_xattr
{
269 int should_free_name
;
273 struct ceph_inode_xattrs_info
{
275 * (still encoded) xattr blob. we avoid the overhead of parsing
276 * this until someone actually calls getxattr, etc.
278 * blob->vec.iov_len == 4 implies there are no xattrs; blob ==
279 * NULL means we don't know.
281 struct ceph_buffer
*blob
, *prealloc_blob
;
283 struct rb_root index
;
288 u64 version
, index_version
;
294 #define CEPH_I_COMPLETE 1 /* we have complete directory cached */
295 #define CEPH_I_NODELAY 4 /* do not delay cap release */
296 #define CEPH_I_FLUSH 8 /* do not delay flush of dirty metadata */
297 #define CEPH_I_NOFLUSH 16 /* do not flush dirty caps */
299 struct ceph_inode_info
{
300 struct ceph_vino i_vino
; /* ceph ino + snap */
305 unsigned i_ceph_flags
;
306 unsigned long i_release_count
;
308 struct ceph_file_layout i_layout
;
312 struct timespec i_rctime
;
313 u64 i_rbytes
, i_rfiles
, i_rsubdirs
;
314 u64 i_files
, i_subdirs
;
315 u64 i_max_offset
; /* largest readdir offset, set with I_COMPLETE */
317 struct rb_root i_fragtree
;
318 struct mutex i_fragtree_mutex
;
320 struct ceph_inode_xattrs_info i_xattrs
;
322 /* capabilities. protected _both_ by i_lock and cap->session's
324 struct rb_root i_caps
; /* cap list */
325 struct ceph_cap
*i_auth_cap
; /* authoritative cap, if any */
326 unsigned i_dirty_caps
, i_flushing_caps
; /* mask of dirtied fields */
327 struct list_head i_dirty_item
, i_flushing_item
;
329 /* we need to track cap writeback on a per-cap-bit basis, to allow
330 * overlapping, pipelined cap flushes to the mds. we can probably
331 * reduce the tid to 8 bits if we're concerned about inode size. */
332 u16 i_cap_flush_last_tid
, i_cap_flush_tid
[CEPH_CAP_BITS
];
333 wait_queue_head_t i_cap_wq
; /* threads waiting on a capability */
334 unsigned long i_hold_caps_min
; /* jiffies */
335 unsigned long i_hold_caps_max
; /* jiffies */
336 struct list_head i_cap_delay_list
; /* for delayed cap release to mds */
337 int i_cap_exporting_mds
; /* to handle cap migration between */
338 unsigned i_cap_exporting_mseq
; /* mds's. */
339 unsigned i_cap_exporting_issued
;
340 struct ceph_cap_reservation i_cap_migration_resv
;
341 struct list_head i_cap_snaps
; /* snapped state pending flush to mds */
342 struct ceph_snap_context
*i_head_snapc
; /* set if wr_buffer_head > 0 */
343 unsigned i_snap_caps
; /* cap bits for snapped files */
345 int i_nr_by_mode
[CEPH_FILE_MODE_NUM
]; /* open file counts */
347 u32 i_truncate_seq
; /* last truncate to smaller size */
348 u64 i_truncate_size
; /* and the size we last truncated down to */
349 int i_truncate_pending
; /* still need to call vmtruncate */
351 u64 i_max_size
; /* max file size authorized by mds */
352 u64 i_reported_size
; /* (max_)size reported to or requested of mds */
353 u64 i_wanted_max_size
; /* offset we'd like to write too */
354 u64 i_requested_max_size
; /* max_size we've requested */
356 /* held references to caps */
358 int i_rd_ref
, i_rdcache_ref
, i_wr_ref
;
359 int i_wrbuffer_ref
, i_wrbuffer_ref_head
;
360 u32 i_shared_gen
; /* increment each time we get FILE_SHARED */
361 u32 i_rdcache_gen
; /* we increment this each time we get
362 FILE_CACHE. If it's non-zero, we
363 _may_ have cached pages. */
364 u32 i_rdcache_revoking
; /* RDCACHE gen to async invalidate, if any */
366 struct list_head i_unsafe_writes
; /* uncommitted sync writes */
367 struct list_head i_unsafe_dirops
; /* uncommitted mds dir ops */
368 spinlock_t i_unsafe_lock
;
370 struct ceph_snap_realm
*i_snap_realm
; /* snap realm (if caps) */
371 int i_snap_realm_counter
; /* snap realm (if caps) */
372 struct list_head i_snap_realm_item
;
373 struct list_head i_snap_flush_item
;
375 struct work_struct i_wb_work
; /* writeback work */
376 struct work_struct i_pg_inv_work
; /* page invalidation work */
378 struct work_struct i_vmtruncate_work
;
380 struct inode vfs_inode
; /* at end */
383 static inline struct ceph_inode_info
*ceph_inode(struct inode
*inode
)
385 return container_of(inode
, struct ceph_inode_info
, vfs_inode
);
388 static inline void ceph_i_clear(struct inode
*inode
, unsigned mask
)
390 struct ceph_inode_info
*ci
= ceph_inode(inode
);
392 spin_lock(&inode
->i_lock
);
393 ci
->i_ceph_flags
&= ~mask
;
394 spin_unlock(&inode
->i_lock
);
397 static inline void ceph_i_set(struct inode
*inode
, unsigned mask
)
399 struct ceph_inode_info
*ci
= ceph_inode(inode
);
401 spin_lock(&inode
->i_lock
);
402 ci
->i_ceph_flags
|= mask
;
403 spin_unlock(&inode
->i_lock
);
406 static inline bool ceph_i_test(struct inode
*inode
, unsigned mask
)
408 struct ceph_inode_info
*ci
= ceph_inode(inode
);
412 r
= (ci
->i_ceph_flags
& mask
) == mask
;
417 /* find a specific frag @f */
418 extern struct ceph_inode_frag
*__ceph_find_frag(struct ceph_inode_info
*ci
,
422 * choose fragment for value @v. copy frag content to pfrag, if leaf
425 extern u32
ceph_choose_frag(struct ceph_inode_info
*ci
, u32 v
,
426 struct ceph_inode_frag
*pfrag
,
432 struct ceph_dentry_info
{
433 struct ceph_mds_session
*lease_session
;
434 u32 lease_gen
, lease_shared_gen
;
436 unsigned long lease_renew_after
, lease_renew_from
;
437 struct list_head lru
;
438 struct dentry
*dentry
;
443 static inline struct ceph_dentry_info
*ceph_dentry(struct dentry
*dentry
)
445 return (struct ceph_dentry_info
*)dentry
->d_fsdata
;
448 static inline loff_t
ceph_make_fpos(unsigned frag
, unsigned off
)
450 return ((loff_t
)frag
<< 32) | (loff_t
)off
;
454 * ino_t is <64 bits on many architectures, blech.
456 * don't include snap in ino hash, at least for now.
458 static inline ino_t
ceph_vino_to_ino(struct ceph_vino vino
)
460 ino_t ino
= (ino_t
)vino
.ino
; /* ^ (vino.snap << 20); */
461 #if BITS_PER_LONG == 32
462 ino
^= vino
.ino
>> (sizeof(u64
)-sizeof(ino_t
)) * 8;
469 static inline int ceph_set_ino_cb(struct inode
*inode
, void *data
)
471 ceph_inode(inode
)->i_vino
= *(struct ceph_vino
*)data
;
472 inode
->i_ino
= ceph_vino_to_ino(*(struct ceph_vino
*)data
);
476 static inline struct ceph_vino
ceph_vino(struct inode
*inode
)
478 return ceph_inode(inode
)->i_vino
;
481 /* for printf-style formatting */
482 #define ceph_vinop(i) ceph_inode(i)->i_vino.ino, ceph_inode(i)->i_vino.snap
484 static inline u64
ceph_ino(struct inode
*inode
)
486 return ceph_inode(inode
)->i_vino
.ino
;
488 static inline u64
ceph_snap(struct inode
*inode
)
490 return ceph_inode(inode
)->i_vino
.snap
;
493 static inline int ceph_ino_compare(struct inode
*inode
, void *data
)
495 struct ceph_vino
*pvino
= (struct ceph_vino
*)data
;
496 struct ceph_inode_info
*ci
= ceph_inode(inode
);
497 return ci
->i_vino
.ino
== pvino
->ino
&&
498 ci
->i_vino
.snap
== pvino
->snap
;
501 static inline struct inode
*ceph_find_inode(struct super_block
*sb
,
502 struct ceph_vino vino
)
504 ino_t t
= ceph_vino_to_ino(vino
);
505 return ilookup5(sb
, t
, ceph_ino_compare
, &vino
);
512 static inline bool __ceph_is_any_real_caps(struct ceph_inode_info
*ci
)
514 return !RB_EMPTY_ROOT(&ci
->i_caps
);
517 extern int __ceph_caps_issued(struct ceph_inode_info
*ci
, int *implemented
);
518 extern int __ceph_caps_issued_mask(struct ceph_inode_info
*ci
, int mask
, int t
);
519 extern int __ceph_caps_issued_other(struct ceph_inode_info
*ci
,
520 struct ceph_cap
*cap
);
522 static inline int ceph_caps_issued(struct ceph_inode_info
*ci
)
525 spin_lock(&ci
->vfs_inode
.i_lock
);
526 issued
= __ceph_caps_issued(ci
, NULL
);
527 spin_unlock(&ci
->vfs_inode
.i_lock
);
531 static inline int ceph_caps_issued_mask(struct ceph_inode_info
*ci
, int mask
,
535 spin_lock(&ci
->vfs_inode
.i_lock
);
536 r
= __ceph_caps_issued_mask(ci
, mask
, touch
);
537 spin_unlock(&ci
->vfs_inode
.i_lock
);
541 static inline int __ceph_caps_dirty(struct ceph_inode_info
*ci
)
543 return ci
->i_dirty_caps
| ci
->i_flushing_caps
;
545 extern void __ceph_mark_dirty_caps(struct ceph_inode_info
*ci
, int mask
);
547 extern int ceph_caps_revoking(struct ceph_inode_info
*ci
, int mask
);
548 extern int __ceph_caps_used(struct ceph_inode_info
*ci
);
550 extern int __ceph_caps_file_wanted(struct ceph_inode_info
*ci
);
553 * wanted, by virtue of open file modes AND cap refs (buffered/cached data)
555 static inline int __ceph_caps_wanted(struct ceph_inode_info
*ci
)
557 int w
= __ceph_caps_file_wanted(ci
) | __ceph_caps_used(ci
);
558 if (w
& CEPH_CAP_FILE_BUFFER
)
559 w
|= CEPH_CAP_FILE_EXCL
; /* we want EXCL if dirty data */
563 /* what the mds thinks we want */
564 extern int __ceph_caps_mds_wanted(struct ceph_inode_info
*ci
);
566 extern void ceph_caps_init(void);
567 extern void ceph_caps_finalize(void);
568 extern void ceph_adjust_min_caps(int delta
);
569 extern int ceph_reserve_caps(struct ceph_cap_reservation
*ctx
, int need
);
570 extern int ceph_unreserve_caps(struct ceph_cap_reservation
*ctx
);
571 extern void ceph_reservation_status(struct ceph_client
*client
,
572 int *total
, int *avail
, int *used
,
573 int *reserved
, int *min
);
575 static inline struct ceph_client
*ceph_inode_to_client(struct inode
*inode
)
577 return (struct ceph_client
*)inode
->i_sb
->s_fs_info
;
580 static inline struct ceph_client
*ceph_sb_to_client(struct super_block
*sb
)
582 return (struct ceph_client
*)sb
->s_fs_info
;
587 * we keep buffered readdir results attached to file->private_data
589 struct ceph_file_info
{
590 int fmode
; /* initialized on open */
592 /* readdir: position within the dir */
594 struct ceph_mds_request
*last_readdir
;
597 /* readdir: position within a frag */
598 unsigned offset
; /* offset of last chunk, adjusted for . and .. */
599 u64 next_offset
; /* offset of next chunk (last_name's + 1) */
600 char *last_name
; /* last entry in previous chunk */
601 struct dentry
*dentry
; /* next dentry (for dcache readdir) */
602 unsigned long dir_release_count
;
604 /* used for -o dirstat read() on directory thing */
616 * A "snap context" is the set of existing snapshots when we
617 * write data. It is used by the OSD to guide its COW behavior.
619 * The ceph_snap_context is refcounted, and attached to each dirty
620 * page, indicating which context the dirty data belonged when it was
623 struct ceph_snap_context
{
630 static inline struct ceph_snap_context
*
631 ceph_get_snap_context(struct ceph_snap_context
*sc
)
634 printk("get_snap_context %p %d -> %d\n", sc, atomic_read(&sc->nref),
635 atomic_read(&sc->nref)+1);
638 atomic_inc(&sc
->nref
);
642 static inline void ceph_put_snap_context(struct ceph_snap_context
*sc
)
647 printk("put_snap_context %p %d -> %d\n", sc, atomic_read(&sc->nref),
648 atomic_read(&sc->nref)-1);
650 if (atomic_dec_and_test(&sc
->nref
)) {
651 /*printk(" deleting snap_context %p\n", sc);*/
657 * A "snap realm" describes a subset of the file hierarchy sharing
658 * the same set of snapshots that apply to it. The realms themselves
659 * are organized into a hierarchy, such that children inherit (some of)
660 * the snapshots of their parents.
662 * All inodes within the realm that have capabilities are linked into a
665 struct ceph_snap_realm
{
672 u64 parent_since
; /* snapid when our current parent became so */
674 u64
*prior_parent_snaps
; /* snaps inherited from any parents we */
675 int num_prior_parent_snaps
; /* had prior to parent_since */
676 u64
*snaps
; /* snaps specific to this realm */
679 struct ceph_snap_realm
*parent
;
680 struct list_head children
; /* list of child realms */
681 struct list_head child_item
;
683 struct list_head empty_item
; /* if i have ref==0 */
685 /* the current set of snaps for this realm */
686 struct ceph_snap_context
*cached_context
;
688 struct list_head inodes_with_caps
;
689 spinlock_t inodes_with_caps_lock
;
695 * calculate the number of pages a given length and offset map onto,
696 * if we align the data.
698 static inline int calc_pages_for(u64 off
, u64 len
)
700 return ((off
+len
+PAGE_CACHE_SIZE
-1) >> PAGE_CACHE_SHIFT
) -
701 (off
>> PAGE_CACHE_SHIFT
);
707 struct ceph_snap_realm
*ceph_lookup_snap_realm(struct ceph_mds_client
*mdsc
,
709 extern void ceph_get_snap_realm(struct ceph_mds_client
*mdsc
,
710 struct ceph_snap_realm
*realm
);
711 extern void ceph_put_snap_realm(struct ceph_mds_client
*mdsc
,
712 struct ceph_snap_realm
*realm
);
713 extern int ceph_update_snap_trace(struct ceph_mds_client
*m
,
714 void *p
, void *e
, bool deletion
);
715 extern void ceph_handle_snap(struct ceph_mds_client
*mdsc
,
716 struct ceph_mds_session
*session
,
717 struct ceph_msg
*msg
);
718 extern void ceph_queue_cap_snap(struct ceph_inode_info
*ci
);
719 extern int __ceph_finish_cap_snap(struct ceph_inode_info
*ci
,
720 struct ceph_cap_snap
*capsnap
);
721 extern void ceph_cleanup_empty_realms(struct ceph_mds_client
*mdsc
);
724 * a cap_snap is "pending" if it is still awaiting an in-progress
725 * sync write (that may/may not still update size, mtime, etc.).
727 static inline bool __ceph_have_pending_cap_snap(struct ceph_inode_info
*ci
)
729 return !list_empty(&ci
->i_cap_snaps
) &&
730 list_entry(ci
->i_cap_snaps
.prev
, struct ceph_cap_snap
,
736 extern struct kmem_cache
*ceph_inode_cachep
;
737 extern struct kmem_cache
*ceph_cap_cachep
;
738 extern struct kmem_cache
*ceph_dentry_cachep
;
739 extern struct kmem_cache
*ceph_file_cachep
;
741 extern const char *ceph_msg_type_name(int type
);
742 extern int ceph_check_fsid(struct ceph_client
*client
, struct ceph_fsid
*fsid
);
744 #define FSID_FORMAT "%02x%02x%02x%02x-%02x%02x-%02x%02x-%02x%02x-" \
745 "%02x%02x%02x%02x%02x%02x"
746 #define PR_FSID(f) (f)->fsid[0], (f)->fsid[1], (f)->fsid[2], (f)->fsid[3], \
747 (f)->fsid[4], (f)->fsid[5], (f)->fsid[6], (f)->fsid[7], \
748 (f)->fsid[8], (f)->fsid[9], (f)->fsid[10], (f)->fsid[11], \
749 (f)->fsid[12], (f)->fsid[13], (f)->fsid[14], (f)->fsid[15]
752 extern const struct inode_operations ceph_file_iops
;
754 extern struct inode
*ceph_alloc_inode(struct super_block
*sb
);
755 extern void ceph_destroy_inode(struct inode
*inode
);
757 extern struct inode
*ceph_get_inode(struct super_block
*sb
,
758 struct ceph_vino vino
);
759 extern struct inode
*ceph_get_snapdir(struct inode
*parent
);
760 extern int ceph_fill_file_size(struct inode
*inode
, int issued
,
761 u32 truncate_seq
, u64 truncate_size
, u64 size
);
762 extern void ceph_fill_file_time(struct inode
*inode
, int issued
,
763 u64 time_warp_seq
, struct timespec
*ctime
,
764 struct timespec
*mtime
, struct timespec
*atime
);
765 extern int ceph_fill_trace(struct super_block
*sb
,
766 struct ceph_mds_request
*req
,
767 struct ceph_mds_session
*session
);
768 extern int ceph_readdir_prepopulate(struct ceph_mds_request
*req
,
769 struct ceph_mds_session
*session
);
771 extern int ceph_inode_holds_cap(struct inode
*inode
, int mask
);
773 extern int ceph_inode_set_size(struct inode
*inode
, loff_t size
);
774 extern void __ceph_do_pending_vmtruncate(struct inode
*inode
);
775 extern void ceph_queue_vmtruncate(struct inode
*inode
);
777 extern void ceph_queue_invalidate(struct inode
*inode
);
778 extern void ceph_queue_writeback(struct inode
*inode
);
780 extern int ceph_do_getattr(struct inode
*inode
, int mask
);
781 extern int ceph_permission(struct inode
*inode
, int mask
);
782 extern int ceph_setattr(struct dentry
*dentry
, struct iattr
*attr
);
783 extern int ceph_getattr(struct vfsmount
*mnt
, struct dentry
*dentry
,
787 extern int ceph_setxattr(struct dentry
*, const char *, const void *,
789 extern ssize_t
ceph_getxattr(struct dentry
*, const char *, void *, size_t);
790 extern ssize_t
ceph_listxattr(struct dentry
*, char *, size_t);
791 extern int ceph_removexattr(struct dentry
*, const char *);
792 extern void __ceph_build_xattrs_blob(struct ceph_inode_info
*ci
);
793 extern void __ceph_destroy_xattrs(struct ceph_inode_info
*ci
);
796 extern const char *ceph_cap_string(int c
);
797 extern void ceph_handle_caps(struct ceph_mds_session
*session
,
798 struct ceph_msg
*msg
);
799 extern int ceph_add_cap(struct inode
*inode
,
800 struct ceph_mds_session
*session
, u64 cap_id
,
801 int fmode
, unsigned issued
, unsigned wanted
,
802 unsigned cap
, unsigned seq
, u64 realmino
, int flags
,
803 struct ceph_cap_reservation
*caps_reservation
);
804 extern void __ceph_remove_cap(struct ceph_cap
*cap
);
805 static inline void ceph_remove_cap(struct ceph_cap
*cap
)
807 struct inode
*inode
= &cap
->ci
->vfs_inode
;
808 spin_lock(&inode
->i_lock
);
809 __ceph_remove_cap(cap
);
810 spin_unlock(&inode
->i_lock
);
812 extern void ceph_put_cap(struct ceph_cap
*cap
);
814 extern void ceph_queue_caps_release(struct inode
*inode
);
815 extern int ceph_write_inode(struct inode
*inode
, struct writeback_control
*wbc
);
816 extern int ceph_fsync(struct file
*file
, struct dentry
*dentry
, int datasync
);
817 extern void ceph_kick_flushing_caps(struct ceph_mds_client
*mdsc
,
818 struct ceph_mds_session
*session
);
819 extern int ceph_get_cap_mds(struct inode
*inode
);
820 extern void ceph_get_cap_refs(struct ceph_inode_info
*ci
, int caps
);
821 extern void ceph_put_cap_refs(struct ceph_inode_info
*ci
, int had
);
822 extern void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info
*ci
, int nr
,
823 struct ceph_snap_context
*snapc
);
824 extern void __ceph_flush_snaps(struct ceph_inode_info
*ci
,
825 struct ceph_mds_session
**psession
);
826 extern void ceph_check_caps(struct ceph_inode_info
*ci
, int flags
,
827 struct ceph_mds_session
*session
);
828 extern void ceph_check_delayed_caps(struct ceph_mds_client
*mdsc
);
829 extern void ceph_flush_dirty_caps(struct ceph_mds_client
*mdsc
);
831 extern int ceph_encode_inode_release(void **p
, struct inode
*inode
,
832 int mds
, int drop
, int unless
, int force
);
833 extern int ceph_encode_dentry_release(void **p
, struct dentry
*dn
,
834 int mds
, int drop
, int unless
);
836 extern int ceph_get_caps(struct ceph_inode_info
*ci
, int need
, int want
,
837 int *got
, loff_t endoff
);
839 /* for counting open files by mode */
840 static inline void __ceph_get_fmode(struct ceph_inode_info
*ci
, int mode
)
842 ci
->i_nr_by_mode
[mode
]++;
844 extern void ceph_put_fmode(struct ceph_inode_info
*ci
, int mode
);
847 extern const struct address_space_operations ceph_aops
;
848 extern int ceph_mmap(struct file
*file
, struct vm_area_struct
*vma
);
851 extern const struct file_operations ceph_file_fops
;
852 extern const struct address_space_operations ceph_aops
;
853 extern int ceph_open(struct inode
*inode
, struct file
*file
);
854 extern struct dentry
*ceph_lookup_open(struct inode
*dir
, struct dentry
*dentry
,
855 struct nameidata
*nd
, int mode
,
857 extern int ceph_release(struct inode
*inode
, struct file
*filp
);
858 extern void ceph_release_page_vector(struct page
**pages
, int num_pages
);
861 extern const struct file_operations ceph_dir_fops
;
862 extern const struct inode_operations ceph_dir_iops
;
863 extern struct dentry_operations ceph_dentry_ops
, ceph_snap_dentry_ops
,
864 ceph_snapdir_dentry_ops
;
866 extern int ceph_handle_notrace_create(struct inode
*dir
, struct dentry
*dentry
);
867 extern struct dentry
*ceph_finish_lookup(struct ceph_mds_request
*req
,
868 struct dentry
*dentry
, int err
);
870 extern void ceph_dentry_lru_add(struct dentry
*dn
);
871 extern void ceph_dentry_lru_touch(struct dentry
*dn
);
872 extern void ceph_dentry_lru_del(struct dentry
*dn
);
875 * our d_ops vary depending on whether the inode is live,
876 * snapshotted (read-only), or a virtual ".snap" directory.
878 int ceph_init_dentry(struct dentry
*dentry
);
882 extern long ceph_ioctl(struct file
*file
, unsigned int cmd
, unsigned long arg
);
885 extern const struct export_operations ceph_export_ops
;
888 extern int ceph_debugfs_init(void);
889 extern void ceph_debugfs_cleanup(void);
890 extern int ceph_debugfs_client_init(struct ceph_client
*client
);
891 extern void ceph_debugfs_client_cleanup(struct ceph_client
*client
);
893 static inline struct inode
*get_dentry_parent_inode(struct dentry
*dentry
)
895 if (dentry
&& dentry
->d_parent
)
896 return dentry
->d_parent
->d_inode
;
901 #endif /* _FS_CEPH_SUPER_H */