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>
18 #include <linux/ceph/libceph.h>
20 #ifdef CONFIG_CEPH_FSCACHE
21 #include <linux/fscache.h>
24 /* f_type in struct statfs */
25 #define CEPH_SUPER_MAGIC 0x00c36400
27 /* large granularity for statfs utilization stats to facilitate
28 * large volume sizes on 32-bit machines. */
29 #define CEPH_BLOCK_SHIFT 22 /* 4 MB */
30 #define CEPH_BLOCK (1 << CEPH_BLOCK_SHIFT)
32 #define CEPH_MOUNT_OPT_DIRSTAT (1<<4) /* `cat dirname` for stats */
33 #define CEPH_MOUNT_OPT_RBYTES (1<<5) /* dir st_bytes = rbytes */
34 #define CEPH_MOUNT_OPT_NOASYNCREADDIR (1<<7) /* no dcache readdir */
35 #define CEPH_MOUNT_OPT_INO32 (1<<8) /* 32 bit inos */
36 #define CEPH_MOUNT_OPT_DCACHE (1<<9) /* use dcache for readdir etc */
37 #define CEPH_MOUNT_OPT_FSCACHE (1<<10) /* use fscache */
38 #define CEPH_MOUNT_OPT_NOPOOLPERM (1<<11) /* no pool permission check */
40 #define CEPH_MOUNT_OPT_DEFAULT (CEPH_MOUNT_OPT_RBYTES | \
41 CEPH_MOUNT_OPT_DCACHE)
43 #define ceph_set_mount_opt(fsc, opt) \
44 (fsc)->mount_options->flags |= CEPH_MOUNT_OPT_##opt;
45 #define ceph_test_mount_opt(fsc, opt) \
46 (!!((fsc)->mount_options->flags & CEPH_MOUNT_OPT_##opt))
48 #define CEPH_RSIZE_DEFAULT 0 /* max read size */
49 #define CEPH_RASIZE_DEFAULT (8192*1024) /* readahead */
50 #define CEPH_MAX_READDIR_DEFAULT 1024
51 #define CEPH_MAX_READDIR_BYTES_DEFAULT (512*1024)
52 #define CEPH_SNAPDIRNAME_DEFAULT ".snap"
54 struct ceph_mount_options {
58 int wsize; /* max write size */
59 int rsize; /* max read size */
60 int rasize; /* max readahead */
61 int congestion_kb; /* max writeback in flight */
62 int caps_wanted_delay_min, caps_wanted_delay_max;
63 int cap_release_safety;
64 int max_readdir; /* max readdir result (entires) */
65 int max_readdir_bytes; /* max readdir result (bytes) */
68 * everything above this point can be memcmp'd; everything below
69 * is handled in compare_mount_options()
72 char *snapdir_name; /* default ".snap" */
75 struct ceph_fs_client {
76 struct super_block *sb;
78 struct ceph_mount_options *mount_options;
79 struct ceph_client *client;
81 unsigned long mount_state;
82 int min_caps; /* min caps i added */
84 struct ceph_mds_client *mdsc;
87 mempool_t *wb_pagevec_pool;
88 struct workqueue_struct *wb_wq;
89 struct workqueue_struct *pg_inv_wq;
90 struct workqueue_struct *trunc_wq;
91 atomic_long_t writeback_count;
93 struct backing_dev_info backing_dev_info;
95 #ifdef CONFIG_DEBUG_FS
96 struct dentry *debugfs_dentry_lru, *debugfs_caps;
97 struct dentry *debugfs_congestion_kb;
98 struct dentry *debugfs_bdi;
99 struct dentry *debugfs_mdsc, *debugfs_mdsmap;
100 struct dentry *debugfs_mds_sessions;
103 #ifdef CONFIG_CEPH_FSCACHE
104 struct fscache_cookie *fscache;
105 struct workqueue_struct *revalidate_wq;
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 */
151 * Snapped cap state that is pending flush to mds. When a snapshot occurs,
152 * we first complete any in-process sync writes and writeback any dirty
153 * data before flushing the snapped state (tracked here) back to the MDS.
155 struct ceph_cap_snap {
157 struct ceph_inode_info *ci;
158 struct list_head ci_item, flushing_item;
160 u64 follows, flush_tid;
162 struct ceph_snap_context *context;
168 struct ceph_buffer *xattr_blob;
172 struct timespec mtime, atime, ctime;
174 int writing; /* a sync write is still in progress */
175 int dirty_pages; /* dirty pages awaiting writeback */
180 static inline void ceph_put_cap_snap(struct ceph_cap_snap *capsnap)
182 if (atomic_dec_and_test(&capsnap->nref)) {
183 if (capsnap->xattr_blob)
184 ceph_buffer_put(capsnap->xattr_blob);
190 * The frag tree describes how a directory is fragmented, potentially across
191 * multiple metadata servers. It is also used to indicate points where
192 * metadata authority is delegated, and whether/where metadata is replicated.
194 * A _leaf_ frag will be present in the i_fragtree IFF there is
195 * delegation info. That is, if mds >= 0 || ndist > 0.
197 #define CEPH_MAX_DIRFRAG_REP 4
199 struct ceph_inode_frag {
204 int split_by; /* i.e. 2^(split_by) children */
206 /* delegation and replication info */
207 int mds; /* -1 if same authority as parent */
208 int ndist; /* >0 if replicated */
209 int dist[CEPH_MAX_DIRFRAG_REP];
213 * We cache inode xattrs as an encoded blob until they are first used,
214 * at which point we parse them into an rbtree.
216 struct ceph_inode_xattr {
225 int should_free_name;
232 struct ceph_dentry_info {
233 struct ceph_mds_session *lease_session;
234 u32 lease_gen, lease_shared_gen;
236 unsigned long lease_renew_after, lease_renew_from;
237 struct list_head lru;
238 struct dentry *dentry;
243 struct ceph_inode_xattrs_info {
245 * (still encoded) xattr blob. we avoid the overhead of parsing
246 * this until someone actually calls getxattr, etc.
248 * blob->vec.iov_len == 4 implies there are no xattrs; blob ==
249 * NULL means we don't know.
251 struct ceph_buffer *blob, *prealloc_blob;
253 struct rb_root index;
258 u64 version, index_version;
264 struct ceph_inode_info {
265 struct ceph_vino i_vino; /* ceph ino + snap */
267 spinlock_t i_ceph_lock;
270 u64 i_inline_version;
273 unsigned i_ceph_flags;
275 atomic_t i_release_count;
276 atomic_t i_complete_count;
278 struct ceph_dir_layout i_dir_layout;
279 struct ceph_file_layout i_layout;
283 struct timespec i_rctime;
284 u64 i_rbytes, i_rfiles, i_rsubdirs;
285 u64 i_files, i_subdirs;
287 struct rb_root i_fragtree;
288 struct mutex i_fragtree_mutex;
290 struct ceph_inode_xattrs_info i_xattrs;
292 /* capabilities. protected _both_ by i_ceph_lock and cap->session's
294 struct rb_root i_caps; /* cap list */
295 struct ceph_cap *i_auth_cap; /* authoritative cap, if any */
296 unsigned i_dirty_caps, i_flushing_caps; /* mask of dirtied fields */
297 struct list_head i_dirty_item, i_flushing_item;
299 /* we need to track cap writeback on a per-cap-bit basis, to allow
300 * overlapping, pipelined cap flushes to the mds. we can probably
301 * reduce the tid to 8 bits if we're concerned about inode size. */
302 u16 i_cap_flush_last_tid, i_cap_flush_tid[CEPH_CAP_BITS];
303 wait_queue_head_t i_cap_wq; /* threads waiting on a capability */
304 unsigned long i_hold_caps_min; /* jiffies */
305 unsigned long i_hold_caps_max; /* jiffies */
306 struct list_head i_cap_delay_list; /* for delayed cap release to mds */
307 struct ceph_cap_reservation i_cap_migration_resv;
308 struct list_head i_cap_snaps; /* snapped state pending flush to mds */
309 struct ceph_snap_context *i_head_snapc; /* set if wr_buffer_head > 0 or
310 dirty|flushing caps */
311 unsigned i_snap_caps; /* cap bits for snapped files */
313 int i_nr_by_mode[CEPH_FILE_MODE_NUM]; /* open file counts */
315 struct mutex i_truncate_mutex;
316 u32 i_truncate_seq; /* last truncate to smaller size */
317 u64 i_truncate_size; /* and the size we last truncated down to */
318 int i_truncate_pending; /* still need to call vmtruncate */
320 u64 i_max_size; /* max file size authorized by mds */
321 u64 i_reported_size; /* (max_)size reported to or requested of mds */
322 u64 i_wanted_max_size; /* offset we'd like to write too */
323 u64 i_requested_max_size; /* max_size we've requested */
325 /* held references to caps */
327 int i_rd_ref, i_rdcache_ref, i_wr_ref, i_wb_ref;
328 int i_wrbuffer_ref, i_wrbuffer_ref_head;
329 u32 i_shared_gen; /* increment each time we get FILE_SHARED */
330 u32 i_rdcache_gen; /* incremented each time we get FILE_CACHE. */
331 u32 i_rdcache_revoking; /* RDCACHE gen to async invalidate, if any */
333 struct list_head i_unsafe_writes; /* uncommitted sync writes */
334 struct list_head i_unsafe_dirops; /* uncommitted mds dir ops */
335 spinlock_t i_unsafe_lock;
337 struct ceph_snap_realm *i_snap_realm; /* snap realm (if caps) */
338 int i_snap_realm_counter; /* snap realm (if caps) */
339 struct list_head i_snap_realm_item;
340 struct list_head i_snap_flush_item;
342 struct work_struct i_wb_work; /* writeback work */
343 struct work_struct i_pg_inv_work; /* page invalidation work */
345 struct work_struct i_vmtruncate_work;
347 #ifdef CONFIG_CEPH_FSCACHE
348 struct fscache_cookie *fscache;
349 u32 i_fscache_gen; /* sequence, for delayed fscache validate */
350 struct work_struct i_revalidate_work;
352 struct inode vfs_inode; /* at end */
355 static inline struct ceph_inode_info *ceph_inode(struct inode *inode)
357 return container_of(inode, struct ceph_inode_info, vfs_inode);
360 static inline struct ceph_fs_client *ceph_inode_to_client(struct inode *inode)
362 return (struct ceph_fs_client *)inode->i_sb->s_fs_info;
365 static inline struct ceph_fs_client *ceph_sb_to_client(struct super_block *sb)
367 return (struct ceph_fs_client *)sb->s_fs_info;
370 static inline struct ceph_vino ceph_vino(struct inode *inode)
372 return ceph_inode(inode)->i_vino;
376 * ino_t is <64 bits on many architectures, blech.
378 * i_ino (kernel inode) st_ino (userspace)
383 static inline u32 ceph_ino_to_ino32(__u64 vino)
385 u32 ino = vino & 0xffffffff;
395 static inline ino_t ceph_vino_to_ino(struct ceph_vino vino)
397 #if BITS_PER_LONG == 32
398 return ceph_ino_to_ino32(vino.ino);
400 return (ino_t)vino.ino;
405 * user-visible ino (stat, filldir)
407 #if BITS_PER_LONG == 32
408 static inline ino_t ceph_translate_ino(struct super_block *sb, ino_t ino)
413 static inline ino_t ceph_translate_ino(struct super_block *sb, ino_t ino)
415 if (ceph_test_mount_opt(ceph_sb_to_client(sb), INO32))
416 ino = ceph_ino_to_ino32(ino);
422 /* for printf-style formatting */
423 #define ceph_vinop(i) ceph_inode(i)->i_vino.ino, ceph_inode(i)->i_vino.snap
425 static inline u64 ceph_ino(struct inode *inode)
427 return ceph_inode(inode)->i_vino.ino;
429 static inline u64 ceph_snap(struct inode *inode)
431 return ceph_inode(inode)->i_vino.snap;
434 static inline int ceph_ino_compare(struct inode *inode, void *data)
436 struct ceph_vino *pvino = (struct ceph_vino *)data;
437 struct ceph_inode_info *ci = ceph_inode(inode);
438 return ci->i_vino.ino == pvino->ino &&
439 ci->i_vino.snap == pvino->snap;
442 static inline struct inode *ceph_find_inode(struct super_block *sb,
443 struct ceph_vino vino)
445 ino_t t = ceph_vino_to_ino(vino);
446 return ilookup5(sb, t, ceph_ino_compare, &vino);
453 #define CEPH_I_DIR_ORDERED (1 << 0) /* dentries in dir are ordered */
454 #define CEPH_I_NODELAY (1 << 1) /* do not delay cap release */
455 #define CEPH_I_FLUSH (1 << 2) /* do not delay flush of dirty metadata */
456 #define CEPH_I_NOFLUSH (1 << 3) /* do not flush dirty caps */
457 #define CEPH_I_POOL_PERM (1 << 4) /* pool rd/wr bits are valid */
458 #define CEPH_I_POOL_RD (1 << 5) /* can read from pool */
459 #define CEPH_I_POOL_WR (1 << 6) /* can write to pool */
462 static inline void __ceph_dir_set_complete(struct ceph_inode_info *ci,
463 int release_count, int ordered_count)
465 atomic_set(&ci->i_complete_count, release_count);
466 if (ci->i_ordered_count == ordered_count)
467 ci->i_ceph_flags |= CEPH_I_DIR_ORDERED;
469 ci->i_ceph_flags &= ~CEPH_I_DIR_ORDERED;
472 static inline void __ceph_dir_clear_complete(struct ceph_inode_info *ci)
474 atomic_inc(&ci->i_release_count);
477 static inline bool __ceph_dir_is_complete(struct ceph_inode_info *ci)
479 return atomic_read(&ci->i_complete_count) ==
480 atomic_read(&ci->i_release_count);
483 static inline bool __ceph_dir_is_complete_ordered(struct ceph_inode_info *ci)
485 return __ceph_dir_is_complete(ci) &&
486 (ci->i_ceph_flags & CEPH_I_DIR_ORDERED);
489 static inline void ceph_dir_clear_complete(struct inode *inode)
491 __ceph_dir_clear_complete(ceph_inode(inode));
494 static inline void ceph_dir_clear_ordered(struct inode *inode)
496 struct ceph_inode_info *ci = ceph_inode(inode);
497 spin_lock(&ci->i_ceph_lock);
498 ci->i_ordered_count++;
499 ci->i_ceph_flags &= ~CEPH_I_DIR_ORDERED;
500 spin_unlock(&ci->i_ceph_lock);
503 static inline bool ceph_dir_is_complete_ordered(struct inode *inode)
505 struct ceph_inode_info *ci = ceph_inode(inode);
507 spin_lock(&ci->i_ceph_lock);
508 ret = __ceph_dir_is_complete_ordered(ci);
509 spin_unlock(&ci->i_ceph_lock);
513 /* find a specific frag @f */
514 extern struct ceph_inode_frag *__ceph_find_frag(struct ceph_inode_info *ci,
518 * choose fragment for value @v. copy frag content to pfrag, if leaf
521 extern u32 ceph_choose_frag(struct ceph_inode_info *ci, u32 v,
522 struct ceph_inode_frag *pfrag,
525 static inline struct ceph_dentry_info *ceph_dentry(struct dentry *dentry)
527 return (struct ceph_dentry_info *)dentry->d_fsdata;
530 static inline loff_t ceph_make_fpos(unsigned frag, unsigned off)
532 return ((loff_t)frag << 32) | (loff_t)off;
538 static inline bool __ceph_is_any_real_caps(struct ceph_inode_info *ci)
540 return !RB_EMPTY_ROOT(&ci->i_caps);
543 extern int __ceph_caps_issued(struct ceph_inode_info *ci, int *implemented);
544 extern int __ceph_caps_issued_mask(struct ceph_inode_info *ci, int mask, int t);
545 extern int __ceph_caps_issued_other(struct ceph_inode_info *ci,
546 struct ceph_cap *cap);
548 static inline int ceph_caps_issued(struct ceph_inode_info *ci)
551 spin_lock(&ci->i_ceph_lock);
552 issued = __ceph_caps_issued(ci, NULL);
553 spin_unlock(&ci->i_ceph_lock);
557 static inline int ceph_caps_issued_mask(struct ceph_inode_info *ci, int mask,
561 spin_lock(&ci->i_ceph_lock);
562 r = __ceph_caps_issued_mask(ci, mask, touch);
563 spin_unlock(&ci->i_ceph_lock);
567 static inline int __ceph_caps_dirty(struct ceph_inode_info *ci)
569 return ci->i_dirty_caps | ci->i_flushing_caps;
571 extern int __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask);
573 extern int __ceph_caps_revoking_other(struct ceph_inode_info *ci,
574 struct ceph_cap *ocap, int mask);
575 extern int ceph_caps_revoking(struct ceph_inode_info *ci, int mask);
576 extern int __ceph_caps_used(struct ceph_inode_info *ci);
578 extern int __ceph_caps_file_wanted(struct ceph_inode_info *ci);
581 * wanted, by virtue of open file modes AND cap refs (buffered/cached data)
583 static inline int __ceph_caps_wanted(struct ceph_inode_info *ci)
585 int w = __ceph_caps_file_wanted(ci) | __ceph_caps_used(ci);
586 if (w & CEPH_CAP_FILE_BUFFER)
587 w |= CEPH_CAP_FILE_EXCL; /* we want EXCL if dirty data */
591 /* what the mds thinks we want */
592 extern int __ceph_caps_mds_wanted(struct ceph_inode_info *ci);
594 extern void ceph_caps_init(struct ceph_mds_client *mdsc);
595 extern void ceph_caps_finalize(struct ceph_mds_client *mdsc);
596 extern void ceph_adjust_min_caps(struct ceph_mds_client *mdsc, int delta);
597 extern void ceph_reserve_caps(struct ceph_mds_client *mdsc,
598 struct ceph_cap_reservation *ctx, int need);
599 extern int ceph_unreserve_caps(struct ceph_mds_client *mdsc,
600 struct ceph_cap_reservation *ctx);
601 extern void ceph_reservation_status(struct ceph_fs_client *client,
602 int *total, int *avail, int *used,
603 int *reserved, int *min);
608 * we keep buffered readdir results attached to file->private_data
610 #define CEPH_F_SYNC 1
611 #define CEPH_F_ATEND 2
613 struct ceph_file_info {
614 short fmode; /* initialized on open */
615 short flags; /* CEPH_F_* */
617 /* readdir: position within the dir */
619 struct ceph_mds_request *last_readdir;
621 /* readdir: position within a frag */
622 unsigned offset; /* offset of last chunk, adjusted for . and .. */
623 unsigned next_offset; /* offset of next chunk (last_name's + 1) */
624 char *last_name; /* last entry in previous chunk */
625 struct dentry *dentry; /* next dentry (for dcache readdir) */
626 int dir_release_count;
627 int dir_ordered_count;
629 /* used for -o dirstat read() on directory thing */
637 * A "snap realm" describes a subset of the file hierarchy sharing
638 * the same set of snapshots that apply to it. The realms themselves
639 * are organized into a hierarchy, such that children inherit (some of)
640 * the snapshots of their parents.
642 * All inodes within the realm that have capabilities are linked into a
645 struct ceph_snap_realm {
652 u64 parent_since; /* snapid when our current parent became so */
654 u64 *prior_parent_snaps; /* snaps inherited from any parents we */
655 u32 num_prior_parent_snaps; /* had prior to parent_since */
656 u64 *snaps; /* snaps specific to this realm */
659 struct ceph_snap_realm *parent;
660 struct list_head children; /* list of child realms */
661 struct list_head child_item;
663 struct list_head empty_item; /* if i have ref==0 */
665 struct list_head dirty_item; /* if realm needs new context */
667 /* the current set of snaps for this realm */
668 struct ceph_snap_context *cached_context;
670 struct list_head inodes_with_caps;
671 spinlock_t inodes_with_caps_lock;
674 static inline int default_congestion_kb(void)
681 * congestion size, scale with available memory.
693 * This allows larger machines to have larger/more transfers.
694 * Limit the default to 256M
696 congestion_kb = (16*int_sqrt(totalram_pages)) << (PAGE_SHIFT-10);
697 if (congestion_kb > 256*1024)
698 congestion_kb = 256*1024;
700 return congestion_kb;
706 extern struct ceph_snap_context *ceph_empty_snapc;
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 struct ceph_snap_realm **realm_ret);
716 extern void ceph_handle_snap(struct ceph_mds_client *mdsc,
717 struct ceph_mds_session *session,
718 struct ceph_msg *msg);
719 extern void ceph_queue_cap_snap(struct ceph_inode_info *ci);
720 extern int __ceph_finish_cap_snap(struct ceph_inode_info *ci,
721 struct ceph_cap_snap *capsnap);
722 extern void ceph_cleanup_empty_realms(struct ceph_mds_client *mdsc);
723 extern int ceph_snap_init(void);
724 extern void ceph_snap_exit(void);
727 * a cap_snap is "pending" if it is still awaiting an in-progress
728 * sync write (that may/may not still update size, mtime, etc.).
730 static inline bool __ceph_have_pending_cap_snap(struct ceph_inode_info *ci)
732 return !list_empty(&ci->i_cap_snaps) &&
733 list_last_entry(&ci->i_cap_snaps, struct ceph_cap_snap,
738 extern const struct inode_operations ceph_file_iops;
740 extern struct inode *ceph_alloc_inode(struct super_block *sb);
741 extern void ceph_destroy_inode(struct inode *inode);
742 extern int ceph_drop_inode(struct inode *inode);
744 extern struct inode *ceph_get_inode(struct super_block *sb,
745 struct ceph_vino vino);
746 extern struct inode *ceph_get_snapdir(struct inode *parent);
747 extern int ceph_fill_file_size(struct inode *inode, int issued,
748 u32 truncate_seq, u64 truncate_size, u64 size);
749 extern void ceph_fill_file_time(struct inode *inode, int issued,
750 u64 time_warp_seq, struct timespec *ctime,
751 struct timespec *mtime, struct timespec *atime);
752 extern int ceph_fill_trace(struct super_block *sb,
753 struct ceph_mds_request *req,
754 struct ceph_mds_session *session);
755 extern int ceph_readdir_prepopulate(struct ceph_mds_request *req,
756 struct ceph_mds_session *session);
758 extern int ceph_inode_holds_cap(struct inode *inode, int mask);
760 extern int ceph_inode_set_size(struct inode *inode, loff_t size);
761 extern void __ceph_do_pending_vmtruncate(struct inode *inode);
762 extern void ceph_queue_vmtruncate(struct inode *inode);
764 extern void ceph_queue_invalidate(struct inode *inode);
765 extern void ceph_queue_writeback(struct inode *inode);
767 extern int __ceph_do_getattr(struct inode *inode, struct page *locked_page,
768 int mask, bool force);
769 static inline int ceph_do_getattr(struct inode *inode, int mask, bool force)
771 return __ceph_do_getattr(inode, NULL, mask, force);
773 extern int ceph_permission(struct inode *inode, int mask);
774 extern int ceph_setattr(struct dentry *dentry, struct iattr *attr);
775 extern int ceph_getattr(struct vfsmount *mnt, struct dentry *dentry,
779 extern int ceph_setxattr(struct dentry *, const char *, const void *,
781 int __ceph_setxattr(struct dentry *, const char *, const void *, size_t, int);
782 ssize_t __ceph_getxattr(struct inode *, const char *, void *, size_t);
783 int __ceph_removexattr(struct dentry *, const char *);
784 extern ssize_t ceph_getxattr(struct dentry *, const char *, void *, size_t);
785 extern ssize_t ceph_listxattr(struct dentry *, char *, size_t);
786 extern int ceph_removexattr(struct dentry *, const char *);
787 extern void __ceph_build_xattrs_blob(struct ceph_inode_info *ci);
788 extern void __ceph_destroy_xattrs(struct ceph_inode_info *ci);
789 extern void __init ceph_xattr_init(void);
790 extern void ceph_xattr_exit(void);
791 extern const struct xattr_handler *ceph_xattr_handlers[];
794 struct ceph_acls_info {
797 struct ceph_pagelist *pagelist;
800 #ifdef CONFIG_CEPH_FS_POSIX_ACL
802 struct posix_acl *ceph_get_acl(struct inode *, int);
803 int ceph_set_acl(struct inode *inode, struct posix_acl *acl, int type);
804 int ceph_pre_init_acls(struct inode *dir, umode_t *mode,
805 struct ceph_acls_info *info);
806 void ceph_init_inode_acls(struct inode *inode, struct ceph_acls_info *info);
807 void ceph_release_acls_info(struct ceph_acls_info *info);
809 static inline void ceph_forget_all_cached_acls(struct inode *inode)
811 forget_all_cached_acls(inode);
816 #define ceph_get_acl NULL
817 #define ceph_set_acl NULL
819 static inline int ceph_pre_init_acls(struct inode *dir, umode_t *mode,
820 struct ceph_acls_info *info)
824 static inline void ceph_init_inode_acls(struct inode *inode,
825 struct ceph_acls_info *info)
828 static inline void ceph_release_acls_info(struct ceph_acls_info *info)
831 static inline int ceph_acl_chmod(struct dentry *dentry, struct inode *inode)
836 static inline void ceph_forget_all_cached_acls(struct inode *inode)
843 extern const char *ceph_cap_string(int c);
844 extern void ceph_handle_caps(struct ceph_mds_session *session,
845 struct ceph_msg *msg);
846 extern struct ceph_cap *ceph_get_cap(struct ceph_mds_client *mdsc,
847 struct ceph_cap_reservation *ctx);
848 extern void ceph_add_cap(struct inode *inode,
849 struct ceph_mds_session *session, u64 cap_id,
850 int fmode, unsigned issued, unsigned wanted,
851 unsigned cap, unsigned seq, u64 realmino, int flags,
852 struct ceph_cap **new_cap);
853 extern void __ceph_remove_cap(struct ceph_cap *cap, bool queue_release);
854 extern void ceph_put_cap(struct ceph_mds_client *mdsc,
855 struct ceph_cap *cap);
856 extern int ceph_is_any_caps(struct inode *inode);
858 extern void ceph_queue_caps_release(struct inode *inode);
859 extern int ceph_write_inode(struct inode *inode, struct writeback_control *wbc);
860 extern int ceph_fsync(struct file *file, loff_t start, loff_t end,
862 extern void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc,
863 struct ceph_mds_session *session);
864 extern struct ceph_cap *ceph_get_cap_for_mds(struct ceph_inode_info *ci,
866 extern int ceph_get_cap_mds(struct inode *inode);
867 extern void ceph_get_cap_refs(struct ceph_inode_info *ci, int caps);
868 extern void ceph_put_cap_refs(struct ceph_inode_info *ci, int had);
869 extern void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr,
870 struct ceph_snap_context *snapc);
871 extern void __ceph_flush_snaps(struct ceph_inode_info *ci,
872 struct ceph_mds_session **psession,
874 extern void ceph_check_caps(struct ceph_inode_info *ci, int flags,
875 struct ceph_mds_session *session);
876 extern void ceph_check_delayed_caps(struct ceph_mds_client *mdsc);
877 extern void ceph_flush_dirty_caps(struct ceph_mds_client *mdsc);
879 extern int ceph_encode_inode_release(void **p, struct inode *inode,
880 int mds, int drop, int unless, int force);
881 extern int ceph_encode_dentry_release(void **p, struct dentry *dn,
882 int mds, int drop, int unless);
884 extern int ceph_get_caps(struct ceph_inode_info *ci, int need, int want,
885 loff_t endoff, int *got, struct page **pinned_page);
887 /* for counting open files by mode */
888 static inline void __ceph_get_fmode(struct ceph_inode_info *ci, int mode)
890 ci->i_nr_by_mode[mode]++;
892 extern void ceph_put_fmode(struct ceph_inode_info *ci, int mode);
895 extern const struct address_space_operations ceph_aops;
896 extern int ceph_mmap(struct file *file, struct vm_area_struct *vma);
897 extern int ceph_uninline_data(struct file *filp, struct page *locked_page);
898 extern int ceph_pool_perm_check(struct ceph_inode_info *ci, int need);
899 extern void ceph_pool_perm_destroy(struct ceph_mds_client* mdsc);
902 extern const struct file_operations ceph_file_fops;
904 extern int ceph_open(struct inode *inode, struct file *file);
905 extern int ceph_atomic_open(struct inode *dir, struct dentry *dentry,
906 struct file *file, unsigned flags, umode_t mode,
908 extern int ceph_release(struct inode *inode, struct file *filp);
909 extern void ceph_fill_inline_data(struct inode *inode, struct page *locked_page,
910 char *data, size_t len);
912 extern const struct file_operations ceph_dir_fops;
913 extern const struct file_operations ceph_snapdir_fops;
914 extern const struct inode_operations ceph_dir_iops;
915 extern const struct inode_operations ceph_snapdir_iops;
916 extern const struct dentry_operations ceph_dentry_ops, ceph_snap_dentry_ops,
917 ceph_snapdir_dentry_ops;
919 extern int ceph_handle_notrace_create(struct inode *dir, struct dentry *dentry);
920 extern int ceph_handle_snapdir(struct ceph_mds_request *req,
921 struct dentry *dentry, int err);
922 extern struct dentry *ceph_finish_lookup(struct ceph_mds_request *req,
923 struct dentry *dentry, int err);
925 extern void ceph_dentry_lru_add(struct dentry *dn);
926 extern void ceph_dentry_lru_touch(struct dentry *dn);
927 extern void ceph_dentry_lru_del(struct dentry *dn);
928 extern void ceph_invalidate_dentry_lease(struct dentry *dentry);
929 extern unsigned ceph_dentry_hash(struct inode *dir, struct dentry *dn);
930 extern struct inode *ceph_get_dentry_parent_inode(struct dentry *dentry);
933 * our d_ops vary depending on whether the inode is live,
934 * snapshotted (read-only), or a virtual ".snap" directory.
936 int ceph_init_dentry(struct dentry *dentry);
940 extern long ceph_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
943 extern const struct export_operations ceph_export_ops;
946 extern __init void ceph_flock_init(void);
947 extern int ceph_lock(struct file *file, int cmd, struct file_lock *fl);
948 extern int ceph_flock(struct file *file, int cmd, struct file_lock *fl);
949 extern void ceph_count_locks(struct inode *inode, int *p_num, int *f_num);
950 extern int ceph_encode_locks_to_buffer(struct inode *inode,
951 struct ceph_filelock *flocks,
953 int num_flock_locks);
954 extern int ceph_locks_to_pagelist(struct ceph_filelock *flocks,
955 struct ceph_pagelist *pagelist,
956 int num_fcntl_locks, int num_flock_locks);
957 extern int lock_to_ceph_filelock(struct file_lock *fl, struct ceph_filelock *c);
960 extern int ceph_fs_debugfs_init(struct ceph_fs_client *client);
961 extern void ceph_fs_debugfs_cleanup(struct ceph_fs_client *client);
963 #endif /* _FS_CEPH_SUPER_H */