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 /* max size of osd read request, limited by libceph */
50 #define CEPH_MAX_READ_SIZE CEPH_MSG_MAX_DATA_LEN
51 /* osd has a configurable limitaion of max write size.
52 * CEPH_MSG_MAX_DATA_LEN should be small enough. */
53 #define CEPH_MAX_WRITE_SIZE CEPH_MSG_MAX_DATA_LEN
54 #define CEPH_RASIZE_DEFAULT (8192*1024) /* max readahead */
55 #define CEPH_MAX_READDIR_DEFAULT 1024
56 #define CEPH_MAX_READDIR_BYTES_DEFAULT (512*1024)
57 #define CEPH_SNAPDIRNAME_DEFAULT ".snap"
59 struct ceph_mount_options {
63 int wsize; /* max write size */
64 int rsize; /* max read size */
65 int rasize; /* max readahead */
66 int congestion_kb; /* max writeback in flight */
67 int caps_wanted_delay_min, caps_wanted_delay_max;
68 int max_readdir; /* max readdir result (entires) */
69 int max_readdir_bytes; /* max readdir result (bytes) */
72 * everything above this point can be memcmp'd; everything below
73 * is handled in compare_mount_options()
76 char *snapdir_name; /* default ".snap" */
77 char *mds_namespace; /* default NULL */
78 char *server_path; /* default "/" */
79 char *fscache_uniq; /* default NULL */
82 struct ceph_fs_client {
83 struct super_block *sb;
85 struct ceph_mount_options *mount_options;
86 struct ceph_client *client;
88 unsigned long mount_state;
89 int min_caps; /* min caps i added */
91 struct ceph_mds_client *mdsc;
94 mempool_t *wb_pagevec_pool;
95 struct workqueue_struct *wb_wq;
96 struct workqueue_struct *pg_inv_wq;
97 struct workqueue_struct *trunc_wq;
98 atomic_long_t writeback_count;
100 #ifdef CONFIG_DEBUG_FS
101 struct dentry *debugfs_dentry_lru, *debugfs_caps;
102 struct dentry *debugfs_congestion_kb;
103 struct dentry *debugfs_bdi;
104 struct dentry *debugfs_mdsc, *debugfs_mdsmap;
105 struct dentry *debugfs_mds_sessions;
108 #ifdef CONFIG_CEPH_FSCACHE
109 struct fscache_cookie *fscache;
115 * File i/o capability. This tracks shared state with the metadata
116 * server that allows us to cache or writeback attributes or to read
117 * and write data. For any given inode, we should have one or more
118 * capabilities, one issued by each metadata server, and our
119 * cumulative access is the OR of all issued capabilities.
121 * Each cap is referenced by the inode's i_caps rbtree and by per-mds
122 * session capability lists.
125 struct ceph_inode_info *ci;
126 struct rb_node ci_node; /* per-ci cap tree */
127 struct ceph_mds_session *session;
128 struct list_head session_caps; /* per-session caplist */
129 u64 cap_id; /* unique cap id (mds provided) */
133 int issued; /* latest, from the mds */
134 int implemented; /* implemented superset of
135 issued (for revocation) */
138 /* caps to release */
144 u32 seq, issue_seq, mseq;
145 u32 cap_gen; /* active/stale cycle */
146 unsigned long last_used;
147 struct list_head caps_item;
150 #define CHECK_CAPS_NODELAY 1 /* do not delay any further */
151 #define CHECK_CAPS_AUTHONLY 2 /* only check auth cap */
152 #define CHECK_CAPS_FLUSH 4 /* flush any dirty caps */
154 struct ceph_cap_flush {
156 int caps; /* 0 means capsnap */
157 bool wake; /* wake up flush waiters when finish ? */
158 struct list_head g_list; // global
159 struct list_head i_list; // per inode
163 * Snapped cap state that is pending flush to mds. When a snapshot occurs,
164 * we first complete any in-process sync writes and writeback any dirty
165 * data before flushing the snapped state (tracked here) back to the MDS.
167 struct ceph_cap_snap {
169 struct list_head ci_item;
171 struct ceph_cap_flush cap_flush;
175 struct ceph_snap_context *context;
181 struct ceph_buffer *xattr_blob;
185 struct timespec mtime, atime, ctime;
189 int writing; /* a sync write is still in progress */
190 int dirty_pages; /* dirty pages awaiting writeback */
195 static inline void ceph_put_cap_snap(struct ceph_cap_snap *capsnap)
197 if (refcount_dec_and_test(&capsnap->nref)) {
198 if (capsnap->xattr_blob)
199 ceph_buffer_put(capsnap->xattr_blob);
205 * The frag tree describes how a directory is fragmented, potentially across
206 * multiple metadata servers. It is also used to indicate points where
207 * metadata authority is delegated, and whether/where metadata is replicated.
209 * A _leaf_ frag will be present in the i_fragtree IFF there is
210 * delegation info. That is, if mds >= 0 || ndist > 0.
212 #define CEPH_MAX_DIRFRAG_REP 4
214 struct ceph_inode_frag {
219 int split_by; /* i.e. 2^(split_by) children */
221 /* delegation and replication info */
222 int mds; /* -1 if same authority as parent */
223 int ndist; /* >0 if replicated */
224 int dist[CEPH_MAX_DIRFRAG_REP];
228 * We cache inode xattrs as an encoded blob until they are first used,
229 * at which point we parse them into an rbtree.
231 struct ceph_inode_xattr {
240 int should_free_name;
247 struct ceph_dentry_info {
248 struct ceph_mds_session *lease_session;
249 u32 lease_gen, lease_shared_gen;
251 unsigned long lease_renew_after, lease_renew_from;
252 struct list_head lru;
253 struct dentry *dentry;
258 struct ceph_inode_xattrs_info {
260 * (still encoded) xattr blob. we avoid the overhead of parsing
261 * this until someone actually calls getxattr, etc.
263 * blob->vec.iov_len == 4 implies there are no xattrs; blob ==
264 * NULL means we don't know.
266 struct ceph_buffer *blob, *prealloc_blob;
268 struct rb_root index;
273 u64 version, index_version;
279 struct ceph_inode_info {
280 struct ceph_vino i_vino; /* ceph ino + snap */
282 spinlock_t i_ceph_lock;
285 u64 i_inline_version;
288 unsigned i_ceph_flags;
289 atomic64_t i_release_count;
290 atomic64_t i_ordered_count;
291 atomic64_t i_complete_seq[2];
293 struct ceph_dir_layout i_dir_layout;
294 struct ceph_file_layout i_layout;
298 struct timespec i_rctime;
299 u64 i_rbytes, i_rfiles, i_rsubdirs;
300 u64 i_files, i_subdirs;
302 struct rb_root i_fragtree;
303 int i_fragtree_nsplits;
304 struct mutex i_fragtree_mutex;
306 struct ceph_inode_xattrs_info i_xattrs;
308 /* capabilities. protected _both_ by i_ceph_lock and cap->session's
310 struct rb_root i_caps; /* cap list */
311 struct ceph_cap *i_auth_cap; /* authoritative cap, if any */
312 unsigned i_dirty_caps, i_flushing_caps; /* mask of dirtied fields */
313 struct list_head i_dirty_item, i_flushing_item;
314 /* we need to track cap writeback on a per-cap-bit basis, to allow
315 * overlapping, pipelined cap flushes to the mds. we can probably
316 * reduce the tid to 8 bits if we're concerned about inode size. */
317 struct ceph_cap_flush *i_prealloc_cap_flush;
318 struct list_head i_cap_flush_list;
319 wait_queue_head_t i_cap_wq; /* threads waiting on a capability */
320 unsigned long i_hold_caps_min; /* jiffies */
321 unsigned long i_hold_caps_max; /* jiffies */
322 struct list_head i_cap_delay_list; /* for delayed cap release to mds */
323 struct ceph_cap_reservation i_cap_migration_resv;
324 struct list_head i_cap_snaps; /* snapped state pending flush to mds */
325 struct ceph_snap_context *i_head_snapc; /* set if wr_buffer_head > 0 or
326 dirty|flushing caps */
327 unsigned i_snap_caps; /* cap bits for snapped files */
329 int i_nr_by_mode[CEPH_FILE_MODE_BITS]; /* open file counts */
331 struct mutex i_truncate_mutex;
332 u32 i_truncate_seq; /* last truncate to smaller size */
333 u64 i_truncate_size; /* and the size we last truncated down to */
334 int i_truncate_pending; /* still need to call vmtruncate */
336 u64 i_max_size; /* max file size authorized by mds */
337 u64 i_reported_size; /* (max_)size reported to or requested of mds */
338 u64 i_wanted_max_size; /* offset we'd like to write too */
339 u64 i_requested_max_size; /* max_size we've requested */
341 /* held references to caps */
343 int i_rd_ref, i_rdcache_ref, i_wr_ref, i_wb_ref;
344 int i_wrbuffer_ref, i_wrbuffer_ref_head;
345 u32 i_shared_gen; /* increment each time we get FILE_SHARED */
346 u32 i_rdcache_gen; /* incremented each time we get FILE_CACHE. */
347 u32 i_rdcache_revoking; /* RDCACHE gen to async invalidate, if any */
349 struct list_head i_unsafe_dirops; /* uncommitted mds dir ops */
350 struct list_head i_unsafe_iops; /* uncommitted mds inode ops */
351 spinlock_t i_unsafe_lock;
353 struct ceph_snap_realm *i_snap_realm; /* snap realm (if caps) */
354 int i_snap_realm_counter; /* snap realm (if caps) */
355 struct list_head i_snap_realm_item;
356 struct list_head i_snap_flush_item;
358 struct work_struct i_wb_work; /* writeback work */
359 struct work_struct i_pg_inv_work; /* page invalidation work */
361 struct work_struct i_vmtruncate_work;
363 #ifdef CONFIG_CEPH_FSCACHE
364 struct fscache_cookie *fscache;
367 struct inode vfs_inode; /* at end */
370 static inline struct ceph_inode_info *ceph_inode(struct inode *inode)
372 return container_of(inode, struct ceph_inode_info, vfs_inode);
375 static inline struct ceph_fs_client *ceph_inode_to_client(struct inode *inode)
377 return (struct ceph_fs_client *)inode->i_sb->s_fs_info;
380 static inline struct ceph_fs_client *ceph_sb_to_client(struct super_block *sb)
382 return (struct ceph_fs_client *)sb->s_fs_info;
385 static inline struct ceph_vino ceph_vino(struct inode *inode)
387 return ceph_inode(inode)->i_vino;
391 * ino_t is <64 bits on many architectures, blech.
393 * i_ino (kernel inode) st_ino (userspace)
398 static inline u32 ceph_ino_to_ino32(__u64 vino)
400 u32 ino = vino & 0xffffffff;
410 static inline ino_t ceph_vino_to_ino(struct ceph_vino vino)
412 #if BITS_PER_LONG == 32
413 return ceph_ino_to_ino32(vino.ino);
415 return (ino_t)vino.ino;
420 * user-visible ino (stat, filldir)
422 #if BITS_PER_LONG == 32
423 static inline ino_t ceph_translate_ino(struct super_block *sb, ino_t ino)
428 static inline ino_t ceph_translate_ino(struct super_block *sb, ino_t ino)
430 if (ceph_test_mount_opt(ceph_sb_to_client(sb), INO32))
431 ino = ceph_ino_to_ino32(ino);
437 /* for printf-style formatting */
438 #define ceph_vinop(i) ceph_inode(i)->i_vino.ino, ceph_inode(i)->i_vino.snap
440 static inline u64 ceph_ino(struct inode *inode)
442 return ceph_inode(inode)->i_vino.ino;
444 static inline u64 ceph_snap(struct inode *inode)
446 return ceph_inode(inode)->i_vino.snap;
449 static inline int ceph_ino_compare(struct inode *inode, void *data)
451 struct ceph_vino *pvino = (struct ceph_vino *)data;
452 struct ceph_inode_info *ci = ceph_inode(inode);
453 return ci->i_vino.ino == pvino->ino &&
454 ci->i_vino.snap == pvino->snap;
457 static inline struct inode *ceph_find_inode(struct super_block *sb,
458 struct ceph_vino vino)
460 ino_t t = ceph_vino_to_ino(vino);
461 return ilookup5(sb, t, ceph_ino_compare, &vino);
468 #define CEPH_I_DIR_ORDERED (1 << 0) /* dentries in dir are ordered */
469 #define CEPH_I_NODELAY (1 << 1) /* do not delay cap release */
470 #define CEPH_I_FLUSH (1 << 2) /* do not delay flush of dirty metadata */
471 #define CEPH_I_NOFLUSH (1 << 3) /* do not flush dirty caps */
472 #define CEPH_I_POOL_PERM (1 << 4) /* pool rd/wr bits are valid */
473 #define CEPH_I_POOL_RD (1 << 5) /* can read from pool */
474 #define CEPH_I_POOL_WR (1 << 6) /* can write to pool */
475 #define CEPH_I_SEC_INITED (1 << 7) /* security initialized */
476 #define CEPH_I_CAP_DROPPED (1 << 8) /* caps were forcibly dropped */
477 #define CEPH_I_KICK_FLUSH (1 << 9) /* kick flushing caps */
478 #define CEPH_I_FLUSH_SNAPS (1 << 10) /* need flush snapss */
479 #define CEPH_I_ERROR_WRITE (1 << 11) /* have seen write errors */
482 * We set the ERROR_WRITE bit when we start seeing write errors on an inode
483 * and then clear it when they start succeeding. Note that we do a lockless
484 * check first, and only take the lock if it looks like it needs to be changed.
485 * The write submission code just takes this as a hint, so we're not too
486 * worried if a few slip through in either direction.
488 static inline void ceph_set_error_write(struct ceph_inode_info *ci)
490 if (!(READ_ONCE(ci->i_ceph_flags) & CEPH_I_ERROR_WRITE)) {
491 spin_lock(&ci->i_ceph_lock);
492 ci->i_ceph_flags |= CEPH_I_ERROR_WRITE;
493 spin_unlock(&ci->i_ceph_lock);
497 static inline void ceph_clear_error_write(struct ceph_inode_info *ci)
499 if (READ_ONCE(ci->i_ceph_flags) & CEPH_I_ERROR_WRITE) {
500 spin_lock(&ci->i_ceph_lock);
501 ci->i_ceph_flags &= ~CEPH_I_ERROR_WRITE;
502 spin_unlock(&ci->i_ceph_lock);
506 static inline void __ceph_dir_set_complete(struct ceph_inode_info *ci,
507 long long release_count,
508 long long ordered_count)
510 smp_mb__before_atomic();
511 atomic64_set(&ci->i_complete_seq[0], release_count);
512 atomic64_set(&ci->i_complete_seq[1], ordered_count);
515 static inline void __ceph_dir_clear_complete(struct ceph_inode_info *ci)
517 atomic64_inc(&ci->i_release_count);
520 static inline void __ceph_dir_clear_ordered(struct ceph_inode_info *ci)
522 atomic64_inc(&ci->i_ordered_count);
525 static inline bool __ceph_dir_is_complete(struct ceph_inode_info *ci)
527 return atomic64_read(&ci->i_complete_seq[0]) ==
528 atomic64_read(&ci->i_release_count);
531 static inline bool __ceph_dir_is_complete_ordered(struct ceph_inode_info *ci)
533 return atomic64_read(&ci->i_complete_seq[0]) ==
534 atomic64_read(&ci->i_release_count) &&
535 atomic64_read(&ci->i_complete_seq[1]) ==
536 atomic64_read(&ci->i_ordered_count);
539 static inline void ceph_dir_clear_complete(struct inode *inode)
541 __ceph_dir_clear_complete(ceph_inode(inode));
544 static inline void ceph_dir_clear_ordered(struct inode *inode)
546 __ceph_dir_clear_ordered(ceph_inode(inode));
549 static inline bool ceph_dir_is_complete_ordered(struct inode *inode)
551 bool ret = __ceph_dir_is_complete_ordered(ceph_inode(inode));
556 /* find a specific frag @f */
557 extern struct ceph_inode_frag *__ceph_find_frag(struct ceph_inode_info *ci,
561 * choose fragment for value @v. copy frag content to pfrag, if leaf
564 extern u32 ceph_choose_frag(struct ceph_inode_info *ci, u32 v,
565 struct ceph_inode_frag *pfrag,
568 static inline struct ceph_dentry_info *ceph_dentry(struct dentry *dentry)
570 return (struct ceph_dentry_info *)dentry->d_fsdata;
576 static inline bool __ceph_is_any_real_caps(struct ceph_inode_info *ci)
578 return !RB_EMPTY_ROOT(&ci->i_caps);
581 extern int __ceph_caps_issued(struct ceph_inode_info *ci, int *implemented);
582 extern int __ceph_caps_issued_mask(struct ceph_inode_info *ci, int mask, int t);
583 extern int __ceph_caps_issued_other(struct ceph_inode_info *ci,
584 struct ceph_cap *cap);
586 static inline int ceph_caps_issued(struct ceph_inode_info *ci)
589 spin_lock(&ci->i_ceph_lock);
590 issued = __ceph_caps_issued(ci, NULL);
591 spin_unlock(&ci->i_ceph_lock);
595 static inline int ceph_caps_issued_mask(struct ceph_inode_info *ci, int mask,
599 spin_lock(&ci->i_ceph_lock);
600 r = __ceph_caps_issued_mask(ci, mask, touch);
601 spin_unlock(&ci->i_ceph_lock);
605 static inline int __ceph_caps_dirty(struct ceph_inode_info *ci)
607 return ci->i_dirty_caps | ci->i_flushing_caps;
609 extern struct ceph_cap_flush *ceph_alloc_cap_flush(void);
610 extern void ceph_free_cap_flush(struct ceph_cap_flush *cf);
611 extern int __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask,
612 struct ceph_cap_flush **pcf);
614 extern int __ceph_caps_revoking_other(struct ceph_inode_info *ci,
615 struct ceph_cap *ocap, int mask);
616 extern int ceph_caps_revoking(struct ceph_inode_info *ci, int mask);
617 extern int __ceph_caps_used(struct ceph_inode_info *ci);
619 extern int __ceph_caps_file_wanted(struct ceph_inode_info *ci);
622 * wanted, by virtue of open file modes AND cap refs (buffered/cached data)
624 static inline int __ceph_caps_wanted(struct ceph_inode_info *ci)
626 int w = __ceph_caps_file_wanted(ci) | __ceph_caps_used(ci);
627 if (w & CEPH_CAP_FILE_BUFFER)
628 w |= CEPH_CAP_FILE_EXCL; /* we want EXCL if dirty data */
632 /* what the mds thinks we want */
633 extern int __ceph_caps_mds_wanted(struct ceph_inode_info *ci, bool check);
635 extern void ceph_caps_init(struct ceph_mds_client *mdsc);
636 extern void ceph_caps_finalize(struct ceph_mds_client *mdsc);
637 extern void ceph_adjust_min_caps(struct ceph_mds_client *mdsc, int delta);
638 extern void ceph_reserve_caps(struct ceph_mds_client *mdsc,
639 struct ceph_cap_reservation *ctx, int need);
640 extern int ceph_unreserve_caps(struct ceph_mds_client *mdsc,
641 struct ceph_cap_reservation *ctx);
642 extern void ceph_reservation_status(struct ceph_fs_client *client,
643 int *total, int *avail, int *used,
644 int *reserved, int *min);
649 * we keep buffered readdir results attached to file->private_data
651 #define CEPH_F_SYNC 1
652 #define CEPH_F_ATEND 2
654 struct ceph_file_info {
655 short fmode; /* initialized on open */
656 short flags; /* CEPH_F_* */
658 /* readdir: position within the dir */
660 struct ceph_mds_request *last_readdir;
662 /* readdir: position within a frag */
663 unsigned next_offset; /* offset of next chunk (last_name's + 1) */
664 char *last_name; /* last entry in previous chunk */
665 long long dir_release_count;
666 long long dir_ordered_count;
667 int readdir_cache_idx;
669 /* used for -o dirstat read() on directory thing */
674 struct ceph_readdir_cache_control {
676 struct dentry **dentries;
681 * A "snap realm" describes a subset of the file hierarchy sharing
682 * the same set of snapshots that apply to it. The realms themselves
683 * are organized into a hierarchy, such that children inherit (some of)
684 * the snapshots of their parents.
686 * All inodes within the realm that have capabilities are linked into a
689 struct ceph_snap_realm {
696 u64 parent_since; /* snapid when our current parent became so */
698 u64 *prior_parent_snaps; /* snaps inherited from any parents we */
699 u32 num_prior_parent_snaps; /* had prior to parent_since */
700 u64 *snaps; /* snaps specific to this realm */
703 struct ceph_snap_realm *parent;
704 struct list_head children; /* list of child realms */
705 struct list_head child_item;
707 struct list_head empty_item; /* if i have ref==0 */
709 struct list_head dirty_item; /* if realm needs new context */
711 /* the current set of snaps for this realm */
712 struct ceph_snap_context *cached_context;
714 struct list_head inodes_with_caps;
715 spinlock_t inodes_with_caps_lock;
718 static inline int default_congestion_kb(void)
725 * congestion size, scale with available memory.
737 * This allows larger machines to have larger/more transfers.
738 * Limit the default to 256M
740 congestion_kb = (16*int_sqrt(totalram_pages)) << (PAGE_SHIFT-10);
741 if (congestion_kb > 256*1024)
742 congestion_kb = 256*1024;
744 return congestion_kb;
750 struct ceph_snap_realm *ceph_lookup_snap_realm(struct ceph_mds_client *mdsc,
752 extern void ceph_get_snap_realm(struct ceph_mds_client *mdsc,
753 struct ceph_snap_realm *realm);
754 extern void ceph_put_snap_realm(struct ceph_mds_client *mdsc,
755 struct ceph_snap_realm *realm);
756 extern int ceph_update_snap_trace(struct ceph_mds_client *m,
757 void *p, void *e, bool deletion,
758 struct ceph_snap_realm **realm_ret);
759 extern void ceph_handle_snap(struct ceph_mds_client *mdsc,
760 struct ceph_mds_session *session,
761 struct ceph_msg *msg);
762 extern void ceph_queue_cap_snap(struct ceph_inode_info *ci);
763 extern int __ceph_finish_cap_snap(struct ceph_inode_info *ci,
764 struct ceph_cap_snap *capsnap);
765 extern void ceph_cleanup_empty_realms(struct ceph_mds_client *mdsc);
768 * a cap_snap is "pending" if it is still awaiting an in-progress
769 * sync write (that may/may not still update size, mtime, etc.).
771 static inline bool __ceph_have_pending_cap_snap(struct ceph_inode_info *ci)
773 return !list_empty(&ci->i_cap_snaps) &&
774 list_last_entry(&ci->i_cap_snaps, struct ceph_cap_snap,
779 extern const struct inode_operations ceph_file_iops;
781 extern struct inode *ceph_alloc_inode(struct super_block *sb);
782 extern void ceph_destroy_inode(struct inode *inode);
783 extern int ceph_drop_inode(struct inode *inode);
785 extern struct inode *ceph_get_inode(struct super_block *sb,
786 struct ceph_vino vino);
787 extern struct inode *ceph_get_snapdir(struct inode *parent);
788 extern int ceph_fill_file_size(struct inode *inode, int issued,
789 u32 truncate_seq, u64 truncate_size, u64 size);
790 extern void ceph_fill_file_time(struct inode *inode, int issued,
791 u64 time_warp_seq, struct timespec *ctime,
792 struct timespec *mtime, struct timespec *atime);
793 extern int ceph_fill_trace(struct super_block *sb,
794 struct ceph_mds_request *req);
795 extern int ceph_readdir_prepopulate(struct ceph_mds_request *req,
796 struct ceph_mds_session *session);
798 extern int ceph_inode_holds_cap(struct inode *inode, int mask);
800 extern bool ceph_inode_set_size(struct inode *inode, loff_t size);
801 extern void __ceph_do_pending_vmtruncate(struct inode *inode);
802 extern void ceph_queue_vmtruncate(struct inode *inode);
804 extern void ceph_queue_invalidate(struct inode *inode);
805 extern void ceph_queue_writeback(struct inode *inode);
807 extern int __ceph_do_getattr(struct inode *inode, struct page *locked_page,
808 int mask, bool force);
809 static inline int ceph_do_getattr(struct inode *inode, int mask, bool force)
811 return __ceph_do_getattr(inode, NULL, mask, force);
813 extern int ceph_permission(struct inode *inode, int mask);
814 extern int __ceph_setattr(struct inode *inode, struct iattr *attr);
815 extern int ceph_setattr(struct dentry *dentry, struct iattr *attr);
816 extern int ceph_getattr(const struct path *path, struct kstat *stat,
817 u32 request_mask, unsigned int flags);
820 int __ceph_setxattr(struct inode *, const char *, const void *, size_t, int);
821 ssize_t __ceph_getxattr(struct inode *, const char *, void *, size_t);
822 extern ssize_t ceph_listxattr(struct dentry *, char *, size_t);
823 extern void __ceph_build_xattrs_blob(struct ceph_inode_info *ci);
824 extern void __ceph_destroy_xattrs(struct ceph_inode_info *ci);
825 extern void __init ceph_xattr_init(void);
826 extern void ceph_xattr_exit(void);
827 extern const struct xattr_handler *ceph_xattr_handlers[];
829 #ifdef CONFIG_SECURITY
830 extern bool ceph_security_xattr_deadlock(struct inode *in);
831 extern bool ceph_security_xattr_wanted(struct inode *in);
833 static inline bool ceph_security_xattr_deadlock(struct inode *in)
837 static inline bool ceph_security_xattr_wanted(struct inode *in)
844 struct ceph_acls_info {
847 struct ceph_pagelist *pagelist;
850 #ifdef CONFIG_CEPH_FS_POSIX_ACL
852 struct posix_acl *ceph_get_acl(struct inode *, int);
853 int ceph_set_acl(struct inode *inode, struct posix_acl *acl, int type);
854 int ceph_pre_init_acls(struct inode *dir, umode_t *mode,
855 struct ceph_acls_info *info);
856 void ceph_init_inode_acls(struct inode *inode, struct ceph_acls_info *info);
857 void ceph_release_acls_info(struct ceph_acls_info *info);
859 static inline void ceph_forget_all_cached_acls(struct inode *inode)
861 forget_all_cached_acls(inode);
866 #define ceph_get_acl NULL
867 #define ceph_set_acl NULL
869 static inline int ceph_pre_init_acls(struct inode *dir, umode_t *mode,
870 struct ceph_acls_info *info)
874 static inline void ceph_init_inode_acls(struct inode *inode,
875 struct ceph_acls_info *info)
878 static inline void ceph_release_acls_info(struct ceph_acls_info *info)
881 static inline int ceph_acl_chmod(struct dentry *dentry, struct inode *inode)
886 static inline void ceph_forget_all_cached_acls(struct inode *inode)
893 extern const char *ceph_cap_string(int c);
894 extern void ceph_handle_caps(struct ceph_mds_session *session,
895 struct ceph_msg *msg);
896 extern struct ceph_cap *ceph_get_cap(struct ceph_mds_client *mdsc,
897 struct ceph_cap_reservation *ctx);
898 extern void ceph_add_cap(struct inode *inode,
899 struct ceph_mds_session *session, u64 cap_id,
900 int fmode, unsigned issued, unsigned wanted,
901 unsigned cap, unsigned seq, u64 realmino, int flags,
902 struct ceph_cap **new_cap);
903 extern void __ceph_remove_cap(struct ceph_cap *cap, bool queue_release);
904 extern void ceph_put_cap(struct ceph_mds_client *mdsc,
905 struct ceph_cap *cap);
906 extern int ceph_is_any_caps(struct inode *inode);
908 extern void ceph_queue_caps_release(struct inode *inode);
909 extern int ceph_write_inode(struct inode *inode, struct writeback_control *wbc);
910 extern int ceph_fsync(struct file *file, loff_t start, loff_t end,
912 extern void ceph_early_kick_flushing_caps(struct ceph_mds_client *mdsc,
913 struct ceph_mds_session *session);
914 extern void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc,
915 struct ceph_mds_session *session);
916 extern struct ceph_cap *ceph_get_cap_for_mds(struct ceph_inode_info *ci,
918 extern int ceph_get_cap_mds(struct inode *inode);
919 extern void ceph_get_cap_refs(struct ceph_inode_info *ci, int caps);
920 extern void ceph_put_cap_refs(struct ceph_inode_info *ci, int had);
921 extern void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr,
922 struct ceph_snap_context *snapc);
923 extern void ceph_flush_snaps(struct ceph_inode_info *ci,
924 struct ceph_mds_session **psession);
925 extern bool __ceph_should_report_size(struct ceph_inode_info *ci);
926 extern void ceph_check_caps(struct ceph_inode_info *ci, int flags,
927 struct ceph_mds_session *session);
928 extern void ceph_check_delayed_caps(struct ceph_mds_client *mdsc);
929 extern void ceph_flush_dirty_caps(struct ceph_mds_client *mdsc);
931 extern int ceph_encode_inode_release(void **p, struct inode *inode,
932 int mds, int drop, int unless, int force);
933 extern int ceph_encode_dentry_release(void **p, struct dentry *dn,
935 int mds, int drop, int unless);
937 extern int ceph_get_caps(struct ceph_inode_info *ci, int need, int want,
938 loff_t endoff, int *got, struct page **pinned_page);
939 extern int ceph_try_get_caps(struct ceph_inode_info *ci,
940 int need, int want, int *got);
942 /* for counting open files by mode */
943 extern void __ceph_get_fmode(struct ceph_inode_info *ci, int mode);
944 extern void ceph_put_fmode(struct ceph_inode_info *ci, int mode);
947 extern const struct address_space_operations ceph_aops;
948 extern int ceph_mmap(struct file *file, struct vm_area_struct *vma);
949 extern int ceph_uninline_data(struct file *filp, struct page *locked_page);
950 extern int ceph_pool_perm_check(struct ceph_inode_info *ci, int need);
951 extern void ceph_pool_perm_destroy(struct ceph_mds_client* mdsc);
954 extern const struct file_operations ceph_file_fops;
956 extern int ceph_renew_caps(struct inode *inode);
957 extern int ceph_open(struct inode *inode, struct file *file);
958 extern int ceph_atomic_open(struct inode *dir, struct dentry *dentry,
959 struct file *file, unsigned flags, umode_t mode,
961 extern int ceph_release(struct inode *inode, struct file *filp);
962 extern void ceph_fill_inline_data(struct inode *inode, struct page *locked_page,
963 char *data, size_t len);
966 extern const struct file_operations ceph_dir_fops;
967 extern const struct file_operations ceph_snapdir_fops;
968 extern const struct inode_operations ceph_dir_iops;
969 extern const struct inode_operations ceph_snapdir_iops;
970 extern const struct dentry_operations ceph_dentry_ops;
972 extern loff_t ceph_make_fpos(unsigned high, unsigned off, bool hash_order);
973 extern int ceph_handle_notrace_create(struct inode *dir, struct dentry *dentry);
974 extern int ceph_handle_snapdir(struct ceph_mds_request *req,
975 struct dentry *dentry, int err);
976 extern struct dentry *ceph_finish_lookup(struct ceph_mds_request *req,
977 struct dentry *dentry, int err);
979 extern void ceph_dentry_lru_add(struct dentry *dn);
980 extern void ceph_dentry_lru_touch(struct dentry *dn);
981 extern void ceph_dentry_lru_del(struct dentry *dn);
982 extern void ceph_invalidate_dentry_lease(struct dentry *dentry);
983 extern unsigned ceph_dentry_hash(struct inode *dir, struct dentry *dn);
984 extern void ceph_readdir_cache_release(struct ceph_readdir_cache_control *ctl);
987 extern long ceph_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
990 extern const struct export_operations ceph_export_ops;
993 extern __init void ceph_flock_init(void);
994 extern int ceph_lock(struct file *file, int cmd, struct file_lock *fl);
995 extern int ceph_flock(struct file *file, int cmd, struct file_lock *fl);
996 extern void ceph_count_locks(struct inode *inode, int *p_num, int *f_num);
997 extern int ceph_encode_locks_to_buffer(struct inode *inode,
998 struct ceph_filelock *flocks,
1000 int num_flock_locks);
1001 extern int ceph_locks_to_pagelist(struct ceph_filelock *flocks,
1002 struct ceph_pagelist *pagelist,
1003 int num_fcntl_locks, int num_flock_locks);
1004 extern int lock_to_ceph_filelock(struct file_lock *fl, struct ceph_filelock *c);
1007 extern int ceph_fs_debugfs_init(struct ceph_fs_client *client);
1008 extern void ceph_fs_debugfs_cleanup(struct ceph_fs_client *client);
1010 #endif /* _FS_CEPH_SUPER_H */