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/slab.h>
14 #include <linux/wait.h>
15 #include <linux/writeback.h>
16 #include <linux/slab.h>
19 #include "messenger.h"
21 #include "mon_client.h"
22 #include "mds_client.h"
23 #include "osd_client.h"
26 /* f_type in struct statfs */
27 #define CEPH_SUPER_MAGIC 0x00c36400
29 /* large granularity for statfs utilization stats to facilitate
30 * large volume sizes on 32-bit machines. */
31 #define CEPH_BLOCK_SHIFT 20 /* 1 MB */
32 #define CEPH_BLOCK (1 << CEPH_BLOCK_SHIFT)
37 #define CEPH_OPT_FSID (1<<0)
38 #define CEPH_OPT_NOSHARE (1<<1) /* don't share client with other sbs */
39 #define CEPH_OPT_MYIP (1<<2) /* specified my ip */
40 #define CEPH_OPT_DIRSTAT (1<<4) /* funky `cat dirname` for stats */
41 #define CEPH_OPT_RBYTES (1<<5) /* dir st_bytes = rbytes */
42 #define CEPH_OPT_NOCRC (1<<6) /* no data crc on writes */
43 #define CEPH_OPT_NOASYNCREADDIR (1<<7) /* no dcache readdir */
45 #define CEPH_OPT_DEFAULT (CEPH_OPT_RBYTES)
47 #define ceph_set_opt(client, opt) \
48 (client)->mount_args->flags |= CEPH_OPT_##opt;
49 #define ceph_test_opt(client, opt) \
50 (!!((client)->mount_args->flags & CEPH_OPT_##opt))
53 struct ceph_mount_args {
56 struct ceph_fsid fsid;
57 struct ceph_entity_addr my_addr;
59 struct ceph_entity_addr *mon_addr;
63 int osd_keepalive_timeout;
65 int rsize; /* max readahead */
66 int congestion_kb; /* max writeback in flight */
67 int caps_wanted_delay_min, caps_wanted_delay_max;
68 int cap_release_safety;
69 int max_readdir; /* max readdir size */
70 char *snapdir_name; /* default ".snap" */
78 #define CEPH_MOUNT_TIMEOUT_DEFAULT 60
79 #define CEPH_OSD_TIMEOUT_DEFAULT 60 /* seconds */
80 #define CEPH_OSD_KEEPALIVE_DEFAULT 5
81 #define CEPH_OSD_IDLE_TTL_DEFAULT 60
82 #define CEPH_MOUNT_RSIZE_DEFAULT (512*1024) /* readahead */
83 #define CEPH_MAX_READDIR_DEFAULT 1024
85 #define CEPH_MSG_MAX_FRONT_LEN (16*1024*1024)
86 #define CEPH_MSG_MAX_DATA_LEN (16*1024*1024)
88 #define CEPH_SNAPDIRNAME_DEFAULT ".snap"
89 #define CEPH_AUTH_NAME_DEFAULT "guest"
91 * Delay telling the MDS we no longer want caps, in case we reopen
92 * the file. Delay a minimum amount of time, even if we send a cap
93 * message for some other reason. Otherwise, take the oppotunity to
94 * update the mds to avoid sending another message later.
96 #define CEPH_CAPS_WANTED_DELAY_MIN_DEFAULT 5 /* cap release delay */
97 #define CEPH_CAPS_WANTED_DELAY_MAX_DEFAULT 60 /* cap release delay */
99 #define CEPH_CAP_RELEASE_SAFETY_DEFAULT (CEPH_CAPS_PER_RELEASE * 4)
105 CEPH_MOUNT_UNMOUNTING,
106 CEPH_MOUNT_UNMOUNTED,
113 static inline unsigned long time_sub(unsigned long a, unsigned long b)
115 BUG_ON(time_after(b, a));
116 return (long)a - (long)b;
120 * per-filesystem client state
122 * possibly shared by multiple mount points, if they are
123 * mounting the same ceph filesystem/cluster.
126 struct ceph_fsid fsid;
129 struct mutex mount_mutex; /* serialize mount attempts */
130 struct ceph_mount_args *mount_args;
132 struct super_block *sb;
134 unsigned long mount_state;
135 wait_queue_head_t auth_wq;
139 int min_caps; /* min caps i added */
141 struct ceph_messenger *msgr; /* messenger instance */
142 struct ceph_mon_client monc;
143 struct ceph_mds_client mdsc;
144 struct ceph_osd_client osdc;
147 mempool_t *wb_pagevec_pool;
148 struct workqueue_struct *wb_wq;
149 struct workqueue_struct *pg_inv_wq;
150 struct workqueue_struct *trunc_wq;
151 atomic_long_t writeback_count;
153 struct backing_dev_info backing_dev_info;
155 #ifdef CONFIG_DEBUG_FS
156 struct dentry *debugfs_monmap;
157 struct dentry *debugfs_mdsmap, *debugfs_osdmap;
158 struct dentry *debugfs_dir, *debugfs_dentry_lru, *debugfs_caps;
159 struct dentry *debugfs_congestion_kb;
160 struct dentry *debugfs_bdi;
165 * File i/o capability. This tracks shared state with the metadata
166 * server that allows us to cache or writeback attributes or to read
167 * and write data. For any given inode, we should have one or more
168 * capabilities, one issued by each metadata server, and our
169 * cumulative access is the OR of all issued capabilities.
171 * Each cap is referenced by the inode's i_caps rbtree and by per-mds
172 * session capability lists.
175 struct ceph_inode_info *ci;
176 struct rb_node ci_node; /* per-ci cap tree */
177 struct ceph_mds_session *session;
178 struct list_head session_caps; /* per-session caplist */
180 u64 cap_id; /* unique cap id (mds provided) */
181 int issued; /* latest, from the mds */
182 int implemented; /* implemented superset of issued (for revocation) */
184 u32 seq, issue_seq, mseq;
185 u32 cap_gen; /* active/stale cycle */
186 unsigned long last_used;
187 struct list_head caps_item;
190 #define CHECK_CAPS_NODELAY 1 /* do not delay any further */
191 #define CHECK_CAPS_AUTHONLY 2 /* only check auth cap */
192 #define CHECK_CAPS_FLUSH 4 /* flush any dirty caps */
195 * Snapped cap state that is pending flush to mds. When a snapshot occurs,
196 * we first complete any in-process sync writes and writeback any dirty
197 * data before flushing the snapped state (tracked here) back to the MDS.
199 struct ceph_cap_snap {
201 struct ceph_inode_info *ci;
202 struct list_head ci_item, flushing_item;
204 u64 follows, flush_tid;
206 struct ceph_snap_context *context;
217 struct timespec mtime, atime, ctime;
219 int writing; /* a sync write is still in progress */
220 int dirty_pages; /* dirty pages awaiting writeback */
223 static inline void ceph_put_cap_snap(struct ceph_cap_snap *capsnap)
225 if (atomic_dec_and_test(&capsnap->nref))
230 * The frag tree describes how a directory is fragmented, potentially across
231 * multiple metadata servers. It is also used to indicate points where
232 * metadata authority is delegated, and whether/where metadata is replicated.
234 * A _leaf_ frag will be present in the i_fragtree IFF there is
235 * delegation info. That is, if mds >= 0 || ndist > 0.
237 #define CEPH_MAX_DIRFRAG_REP 4
239 struct ceph_inode_frag {
244 int split_by; /* i.e. 2^(split_by) children */
246 /* delegation and replication info */
247 int mds; /* -1 if same authority as parent */
248 int ndist; /* >0 if replicated */
249 int dist[CEPH_MAX_DIRFRAG_REP];
253 * We cache inode xattrs as an encoded blob until they are first used,
254 * at which point we parse them into an rbtree.
256 struct ceph_inode_xattr {
265 int should_free_name;
269 struct ceph_inode_xattrs_info {
271 * (still encoded) xattr blob. we avoid the overhead of parsing
272 * this until someone actually calls getxattr, etc.
274 * blob->vec.iov_len == 4 implies there are no xattrs; blob ==
275 * NULL means we don't know.
277 struct ceph_buffer *blob, *prealloc_blob;
279 struct rb_root index;
284 u64 version, index_version;
290 #define CEPH_I_COMPLETE 1 /* we have complete directory cached */
291 #define CEPH_I_NODELAY 4 /* do not delay cap release */
292 #define CEPH_I_FLUSH 8 /* do not delay flush of dirty metadata */
293 #define CEPH_I_NOFLUSH 16 /* do not flush dirty caps */
295 struct ceph_inode_info {
296 struct ceph_vino i_vino; /* ceph ino + snap */
301 unsigned i_ceph_flags;
302 unsigned long i_release_count;
304 struct ceph_file_layout i_layout;
308 struct timespec i_rctime;
309 u64 i_rbytes, i_rfiles, i_rsubdirs;
310 u64 i_files, i_subdirs;
311 u64 i_max_offset; /* largest readdir offset, set with I_COMPLETE */
313 struct rb_root i_fragtree;
314 struct mutex i_fragtree_mutex;
316 struct ceph_inode_xattrs_info i_xattrs;
318 /* capabilities. protected _both_ by i_lock and cap->session's
320 struct rb_root i_caps; /* cap list */
321 struct ceph_cap *i_auth_cap; /* authoritative cap, if any */
322 unsigned i_dirty_caps, i_flushing_caps; /* mask of dirtied fields */
323 struct list_head i_dirty_item, i_flushing_item;
325 /* we need to track cap writeback on a per-cap-bit basis, to allow
326 * overlapping, pipelined cap flushes to the mds. we can probably
327 * reduce the tid to 8 bits if we're concerned about inode size. */
328 u16 i_cap_flush_last_tid, i_cap_flush_tid[CEPH_CAP_BITS];
329 wait_queue_head_t i_cap_wq; /* threads waiting on a capability */
330 unsigned long i_hold_caps_min; /* jiffies */
331 unsigned long i_hold_caps_max; /* jiffies */
332 struct list_head i_cap_delay_list; /* for delayed cap release to mds */
333 int i_cap_exporting_mds; /* to handle cap migration between */
334 unsigned i_cap_exporting_mseq; /* mds's. */
335 unsigned i_cap_exporting_issued;
336 struct ceph_cap_reservation i_cap_migration_resv;
337 struct list_head i_cap_snaps; /* snapped state pending flush to mds */
338 struct ceph_snap_context *i_head_snapc; /* set if wr_buffer_head > 0 */
339 unsigned i_snap_caps; /* cap bits for snapped files */
341 int i_nr_by_mode[CEPH_FILE_MODE_NUM]; /* open file counts */
343 u32 i_truncate_seq; /* last truncate to smaller size */
344 u64 i_truncate_size; /* and the size we last truncated down to */
345 int i_truncate_pending; /* still need to call vmtruncate */
347 u64 i_max_size; /* max file size authorized by mds */
348 u64 i_reported_size; /* (max_)size reported to or requested of mds */
349 u64 i_wanted_max_size; /* offset we'd like to write too */
350 u64 i_requested_max_size; /* max_size we've requested */
352 /* held references to caps */
354 int i_rd_ref, i_rdcache_ref, i_wr_ref;
355 int i_wrbuffer_ref, i_wrbuffer_ref_head;
356 u32 i_shared_gen; /* increment each time we get FILE_SHARED */
357 u32 i_rdcache_gen; /* we increment this each time we get
358 FILE_CACHE. If it's non-zero, we
359 _may_ have cached pages. */
360 u32 i_rdcache_revoking; /* RDCACHE gen to async invalidate, if any */
362 struct list_head i_unsafe_writes; /* uncommitted sync writes */
363 struct list_head i_unsafe_dirops; /* uncommitted mds dir ops */
364 spinlock_t i_unsafe_lock;
366 struct ceph_snap_realm *i_snap_realm; /* snap realm (if caps) */
367 int i_snap_realm_counter; /* snap realm (if caps) */
368 struct list_head i_snap_realm_item;
369 struct list_head i_snap_flush_item;
371 struct work_struct i_wb_work; /* writeback work */
372 struct work_struct i_pg_inv_work; /* page invalidation work */
374 struct work_struct i_vmtruncate_work;
376 struct inode vfs_inode; /* at end */
379 static inline struct ceph_inode_info *ceph_inode(struct inode *inode)
381 return container_of(inode, struct ceph_inode_info, vfs_inode);
384 static inline void ceph_i_clear(struct inode *inode, unsigned mask)
386 struct ceph_inode_info *ci = ceph_inode(inode);
388 spin_lock(&inode->i_lock);
389 ci->i_ceph_flags &= ~mask;
390 spin_unlock(&inode->i_lock);
393 static inline void ceph_i_set(struct inode *inode, unsigned mask)
395 struct ceph_inode_info *ci = ceph_inode(inode);
397 spin_lock(&inode->i_lock);
398 ci->i_ceph_flags |= mask;
399 spin_unlock(&inode->i_lock);
402 static inline bool ceph_i_test(struct inode *inode, unsigned mask)
404 struct ceph_inode_info *ci = ceph_inode(inode);
408 r = (ci->i_ceph_flags & mask) == mask;
413 /* find a specific frag @f */
414 extern struct ceph_inode_frag *__ceph_find_frag(struct ceph_inode_info *ci,
418 * choose fragment for value @v. copy frag content to pfrag, if leaf
421 extern u32 ceph_choose_frag(struct ceph_inode_info *ci, u32 v,
422 struct ceph_inode_frag *pfrag,
428 struct ceph_dentry_info {
429 struct ceph_mds_session *lease_session;
430 u32 lease_gen, lease_shared_gen;
432 unsigned long lease_renew_after, lease_renew_from;
433 struct list_head lru;
434 struct dentry *dentry;
439 static inline struct ceph_dentry_info *ceph_dentry(struct dentry *dentry)
441 return (struct ceph_dentry_info *)dentry->d_fsdata;
444 static inline loff_t ceph_make_fpos(unsigned frag, unsigned off)
446 return ((loff_t)frag << 32) | (loff_t)off;
450 * ino_t is <64 bits on many architectures, blech.
452 * don't include snap in ino hash, at least for now.
454 static inline ino_t ceph_vino_to_ino(struct ceph_vino vino)
456 ino_t ino = (ino_t)vino.ino; /* ^ (vino.snap << 20); */
457 #if BITS_PER_LONG == 32
458 ino ^= vino.ino >> (sizeof(u64)-sizeof(ino_t)) * 8;
465 static inline int ceph_set_ino_cb(struct inode *inode, void *data)
467 ceph_inode(inode)->i_vino = *(struct ceph_vino *)data;
468 inode->i_ino = ceph_vino_to_ino(*(struct ceph_vino *)data);
472 static inline struct ceph_vino ceph_vino(struct inode *inode)
474 return ceph_inode(inode)->i_vino;
477 /* for printf-style formatting */
478 #define ceph_vinop(i) ceph_inode(i)->i_vino.ino, ceph_inode(i)->i_vino.snap
480 static inline u64 ceph_ino(struct inode *inode)
482 return ceph_inode(inode)->i_vino.ino;
484 static inline u64 ceph_snap(struct inode *inode)
486 return ceph_inode(inode)->i_vino.snap;
489 static inline int ceph_ino_compare(struct inode *inode, void *data)
491 struct ceph_vino *pvino = (struct ceph_vino *)data;
492 struct ceph_inode_info *ci = ceph_inode(inode);
493 return ci->i_vino.ino == pvino->ino &&
494 ci->i_vino.snap == pvino->snap;
497 static inline struct inode *ceph_find_inode(struct super_block *sb,
498 struct ceph_vino vino)
500 ino_t t = ceph_vino_to_ino(vino);
501 return ilookup5(sb, t, ceph_ino_compare, &vino);
508 static inline bool __ceph_is_any_real_caps(struct ceph_inode_info *ci)
510 return !RB_EMPTY_ROOT(&ci->i_caps);
513 extern int __ceph_caps_issued(struct ceph_inode_info *ci, int *implemented);
514 extern int __ceph_caps_issued_mask(struct ceph_inode_info *ci, int mask, int t);
515 extern int __ceph_caps_issued_other(struct ceph_inode_info *ci,
516 struct ceph_cap *cap);
518 static inline int ceph_caps_issued(struct ceph_inode_info *ci)
521 spin_lock(&ci->vfs_inode.i_lock);
522 issued = __ceph_caps_issued(ci, NULL);
523 spin_unlock(&ci->vfs_inode.i_lock);
527 static inline int ceph_caps_issued_mask(struct ceph_inode_info *ci, int mask,
531 spin_lock(&ci->vfs_inode.i_lock);
532 r = __ceph_caps_issued_mask(ci, mask, touch);
533 spin_unlock(&ci->vfs_inode.i_lock);
537 static inline int __ceph_caps_dirty(struct ceph_inode_info *ci)
539 return ci->i_dirty_caps | ci->i_flushing_caps;
541 extern void __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask);
543 extern int ceph_caps_revoking(struct ceph_inode_info *ci, int mask);
544 extern int __ceph_caps_used(struct ceph_inode_info *ci);
546 extern int __ceph_caps_file_wanted(struct ceph_inode_info *ci);
549 * wanted, by virtue of open file modes AND cap refs (buffered/cached data)
551 static inline int __ceph_caps_wanted(struct ceph_inode_info *ci)
553 int w = __ceph_caps_file_wanted(ci) | __ceph_caps_used(ci);
554 if (w & CEPH_CAP_FILE_BUFFER)
555 w |= CEPH_CAP_FILE_EXCL; /* we want EXCL if dirty data */
559 /* what the mds thinks we want */
560 extern int __ceph_caps_mds_wanted(struct ceph_inode_info *ci);
562 extern void ceph_caps_init(void);
563 extern void ceph_caps_finalize(void);
564 extern void ceph_adjust_min_caps(int delta);
565 extern int ceph_reserve_caps(struct ceph_cap_reservation *ctx, int need);
566 extern int ceph_unreserve_caps(struct ceph_cap_reservation *ctx);
567 extern void ceph_reservation_status(struct ceph_client *client,
568 int *total, int *avail, int *used,
569 int *reserved, int *min);
571 static inline struct ceph_client *ceph_inode_to_client(struct inode *inode)
573 return (struct ceph_client *)inode->i_sb->s_fs_info;
576 static inline struct ceph_client *ceph_sb_to_client(struct super_block *sb)
578 return (struct ceph_client *)sb->s_fs_info;
583 * we keep buffered readdir results attached to file->private_data
585 struct ceph_file_info {
586 int fmode; /* initialized on open */
588 /* readdir: position within the dir */
590 struct ceph_mds_request *last_readdir;
593 /* readdir: position within a frag */
594 unsigned offset; /* offset of last chunk, adjusted for . and .. */
595 u64 next_offset; /* offset of next chunk (last_name's + 1) */
596 char *last_name; /* last entry in previous chunk */
597 struct dentry *dentry; /* next dentry (for dcache readdir) */
598 unsigned long dir_release_count;
600 /* used for -o dirstat read() on directory thing */
612 * A "snap context" is the set of existing snapshots when we
613 * write data. It is used by the OSD to guide its COW behavior.
615 * The ceph_snap_context is refcounted, and attached to each dirty
616 * page, indicating which context the dirty data belonged when it was
619 struct ceph_snap_context {
626 static inline struct ceph_snap_context *
627 ceph_get_snap_context(struct ceph_snap_context *sc)
630 printk("get_snap_context %p %d -> %d\n", sc, atomic_read(&sc->nref),
631 atomic_read(&sc->nref)+1);
634 atomic_inc(&sc->nref);
638 static inline void ceph_put_snap_context(struct ceph_snap_context *sc)
643 printk("put_snap_context %p %d -> %d\n", sc, atomic_read(&sc->nref),
644 atomic_read(&sc->nref)-1);
646 if (atomic_dec_and_test(&sc->nref)) {
647 /*printk(" deleting snap_context %p\n", sc);*/
653 * A "snap realm" describes a subset of the file hierarchy sharing
654 * the same set of snapshots that apply to it. The realms themselves
655 * are organized into a hierarchy, such that children inherit (some of)
656 * the snapshots of their parents.
658 * All inodes within the realm that have capabilities are linked into a
661 struct ceph_snap_realm {
668 u64 parent_since; /* snapid when our current parent became so */
670 u64 *prior_parent_snaps; /* snaps inherited from any parents we */
671 int num_prior_parent_snaps; /* had prior to parent_since */
672 u64 *snaps; /* snaps specific to this realm */
675 struct ceph_snap_realm *parent;
676 struct list_head children; /* list of child realms */
677 struct list_head child_item;
679 struct list_head empty_item; /* if i have ref==0 */
681 /* the current set of snaps for this realm */
682 struct ceph_snap_context *cached_context;
684 struct list_head inodes_with_caps;
685 spinlock_t inodes_with_caps_lock;
691 * calculate the number of pages a given length and offset map onto,
692 * if we align the data.
694 static inline int calc_pages_for(u64 off, u64 len)
696 return ((off+len+PAGE_CACHE_SIZE-1) >> PAGE_CACHE_SHIFT) -
697 (off >> PAGE_CACHE_SHIFT);
703 struct ceph_snap_realm *ceph_lookup_snap_realm(struct ceph_mds_client *mdsc,
705 extern void ceph_get_snap_realm(struct ceph_mds_client *mdsc,
706 struct ceph_snap_realm *realm);
707 extern void ceph_put_snap_realm(struct ceph_mds_client *mdsc,
708 struct ceph_snap_realm *realm);
709 extern int ceph_update_snap_trace(struct ceph_mds_client *m,
710 void *p, void *e, bool deletion);
711 extern void ceph_handle_snap(struct ceph_mds_client *mdsc,
712 struct ceph_mds_session *session,
713 struct ceph_msg *msg);
714 extern void ceph_queue_cap_snap(struct ceph_inode_info *ci);
715 extern int __ceph_finish_cap_snap(struct ceph_inode_info *ci,
716 struct ceph_cap_snap *capsnap);
717 extern void ceph_cleanup_empty_realms(struct ceph_mds_client *mdsc);
720 * a cap_snap is "pending" if it is still awaiting an in-progress
721 * sync write (that may/may not still update size, mtime, etc.).
723 static inline bool __ceph_have_pending_cap_snap(struct ceph_inode_info *ci)
725 return !list_empty(&ci->i_cap_snaps) &&
726 list_entry(ci->i_cap_snaps.prev, struct ceph_cap_snap,
732 extern struct kmem_cache *ceph_inode_cachep;
733 extern struct kmem_cache *ceph_cap_cachep;
734 extern struct kmem_cache *ceph_dentry_cachep;
735 extern struct kmem_cache *ceph_file_cachep;
737 extern const char *ceph_msg_type_name(int type);
738 extern int ceph_check_fsid(struct ceph_client *client, struct ceph_fsid *fsid);
740 #define FSID_FORMAT "%02x%02x%02x%02x-%02x%02x-%02x%02x-%02x%02x-" \
741 "%02x%02x%02x%02x%02x%02x"
742 #define PR_FSID(f) (f)->fsid[0], (f)->fsid[1], (f)->fsid[2], (f)->fsid[3], \
743 (f)->fsid[4], (f)->fsid[5], (f)->fsid[6], (f)->fsid[7], \
744 (f)->fsid[8], (f)->fsid[9], (f)->fsid[10], (f)->fsid[11], \
745 (f)->fsid[12], (f)->fsid[13], (f)->fsid[14], (f)->fsid[15]
748 extern const struct inode_operations ceph_file_iops;
750 extern struct inode *ceph_alloc_inode(struct super_block *sb);
751 extern void ceph_destroy_inode(struct inode *inode);
753 extern struct inode *ceph_get_inode(struct super_block *sb,
754 struct ceph_vino vino);
755 extern struct inode *ceph_get_snapdir(struct inode *parent);
756 extern int ceph_fill_file_size(struct inode *inode, int issued,
757 u32 truncate_seq, u64 truncate_size, u64 size);
758 extern void ceph_fill_file_time(struct inode *inode, int issued,
759 u64 time_warp_seq, struct timespec *ctime,
760 struct timespec *mtime, struct timespec *atime);
761 extern int ceph_fill_trace(struct super_block *sb,
762 struct ceph_mds_request *req,
763 struct ceph_mds_session *session);
764 extern int ceph_readdir_prepopulate(struct ceph_mds_request *req,
765 struct ceph_mds_session *session);
767 extern int ceph_inode_holds_cap(struct inode *inode, int mask);
769 extern int ceph_inode_set_size(struct inode *inode, loff_t size);
770 extern void __ceph_do_pending_vmtruncate(struct inode *inode);
771 extern void ceph_queue_vmtruncate(struct inode *inode);
773 extern void ceph_queue_invalidate(struct inode *inode);
774 extern void ceph_queue_writeback(struct inode *inode);
776 extern int ceph_do_getattr(struct inode *inode, int mask);
777 extern int ceph_permission(struct inode *inode, int mask);
778 extern int ceph_setattr(struct dentry *dentry, struct iattr *attr);
779 extern int ceph_getattr(struct vfsmount *mnt, struct dentry *dentry,
783 extern int ceph_setxattr(struct dentry *, const char *, const void *,
785 extern ssize_t ceph_getxattr(struct dentry *, const char *, void *, size_t);
786 extern ssize_t ceph_listxattr(struct dentry *, char *, size_t);
787 extern int ceph_removexattr(struct dentry *, const char *);
788 extern void __ceph_build_xattrs_blob(struct ceph_inode_info *ci);
789 extern void __ceph_destroy_xattrs(struct ceph_inode_info *ci);
792 extern const char *ceph_cap_string(int c);
793 extern void ceph_handle_caps(struct ceph_mds_session *session,
794 struct ceph_msg *msg);
795 extern int ceph_add_cap(struct inode *inode,
796 struct ceph_mds_session *session, u64 cap_id,
797 int fmode, unsigned issued, unsigned wanted,
798 unsigned cap, unsigned seq, u64 realmino, int flags,
799 struct ceph_cap_reservation *caps_reservation);
800 extern void __ceph_remove_cap(struct ceph_cap *cap);
801 static inline void ceph_remove_cap(struct ceph_cap *cap)
803 struct inode *inode = &cap->ci->vfs_inode;
804 spin_lock(&inode->i_lock);
805 __ceph_remove_cap(cap);
806 spin_unlock(&inode->i_lock);
808 extern void ceph_put_cap(struct ceph_cap *cap);
810 extern void ceph_queue_caps_release(struct inode *inode);
811 extern int ceph_write_inode(struct inode *inode, struct writeback_control *wbc);
812 extern int ceph_fsync(struct file *file, struct dentry *dentry, int datasync);
813 extern void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc,
814 struct ceph_mds_session *session);
815 extern int ceph_get_cap_mds(struct inode *inode);
816 extern void ceph_get_cap_refs(struct ceph_inode_info *ci, int caps);
817 extern void ceph_put_cap_refs(struct ceph_inode_info *ci, int had);
818 extern void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr,
819 struct ceph_snap_context *snapc);
820 extern void __ceph_flush_snaps(struct ceph_inode_info *ci,
821 struct ceph_mds_session **psession);
822 extern void ceph_check_caps(struct ceph_inode_info *ci, int flags,
823 struct ceph_mds_session *session);
824 extern void ceph_check_delayed_caps(struct ceph_mds_client *mdsc);
825 extern void ceph_flush_dirty_caps(struct ceph_mds_client *mdsc);
827 extern int ceph_encode_inode_release(void **p, struct inode *inode,
828 int mds, int drop, int unless, int force);
829 extern int ceph_encode_dentry_release(void **p, struct dentry *dn,
830 int mds, int drop, int unless);
832 extern int ceph_get_caps(struct ceph_inode_info *ci, int need, int want,
833 int *got, loff_t endoff);
835 /* for counting open files by mode */
836 static inline void __ceph_get_fmode(struct ceph_inode_info *ci, int mode)
838 ci->i_nr_by_mode[mode]++;
840 extern void ceph_put_fmode(struct ceph_inode_info *ci, int mode);
843 extern const struct address_space_operations ceph_aops;
844 extern int ceph_mmap(struct file *file, struct vm_area_struct *vma);
847 extern const struct file_operations ceph_file_fops;
848 extern const struct address_space_operations ceph_aops;
849 extern int ceph_open(struct inode *inode, struct file *file);
850 extern struct dentry *ceph_lookup_open(struct inode *dir, struct dentry *dentry,
851 struct nameidata *nd, int mode,
853 extern int ceph_release(struct inode *inode, struct file *filp);
854 extern void ceph_release_page_vector(struct page **pages, int num_pages);
857 extern const struct file_operations ceph_dir_fops;
858 extern const struct inode_operations ceph_dir_iops;
859 extern struct dentry_operations ceph_dentry_ops, ceph_snap_dentry_ops,
860 ceph_snapdir_dentry_ops;
862 extern int ceph_handle_notrace_create(struct inode *dir, struct dentry *dentry);
863 extern struct dentry *ceph_finish_lookup(struct ceph_mds_request *req,
864 struct dentry *dentry, int err);
866 extern void ceph_dentry_lru_add(struct dentry *dn);
867 extern void ceph_dentry_lru_touch(struct dentry *dn);
868 extern void ceph_dentry_lru_del(struct dentry *dn);
869 extern void ceph_invalidate_dentry_lease(struct dentry *dentry);
872 * our d_ops vary depending on whether the inode is live,
873 * snapshotted (read-only), or a virtual ".snap" directory.
875 int ceph_init_dentry(struct dentry *dentry);
879 extern long ceph_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
882 extern const struct export_operations ceph_export_ops;
885 extern int ceph_debugfs_init(void);
886 extern void ceph_debugfs_cleanup(void);
887 extern int ceph_debugfs_client_init(struct ceph_client *client);
888 extern void ceph_debugfs_client_cleanup(struct ceph_client *client);
890 static inline struct inode *get_dentry_parent_inode(struct dentry *dentry)
892 if (dentry && dentry->d_parent)
893 return dentry->d_parent->d_inode;
898 #endif /* _FS_CEPH_SUPER_H */