2 * Copyright (c) 1989, 1991, 1993
3 * The Regents of the University of California. All rights reserved.
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
8 * 1. Redistributions of source code must retain the above copyright
9 * notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 * notice, this list of conditions and the following disclaimer in the
12 * documentation and/or other materials provided with the distribution.
13 * 4. Neither the name of the University nor the names of its contributors
14 * may be used to endorse or promote products derived from this software
15 * without specific prior written permission.
17 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 * @(#)mount.h 8.21 (Berkeley) 5/20/95
36 #include <sys/ucred.h>
37 #include <sys/queue.h>
40 #include <sys/lockmgr.h>
41 #include <sys/_mutex.h>
46 * NOTE: When changing statfs structure, mount structure, MNT_* flags or
47 * MNTK_* flags also update DDB show mount command in vfs_subr.c.
50 typedef struct fsid { int32_t val[2]; } fsid_t; /* filesystem id type */
54 * These are unique per filesystem on a single machine.
59 u_short fid_len; /* length of data in bytes */
60 u_short fid_data0; /* force longword alignment */
61 char fid_data[MAXFIDSZ]; /* data (variable length) */
65 * filesystem statistics
67 #define MFSNAMELEN 16 /* length of type name including null */
68 #define MNAMELEN 88 /* size of on/from name bufs */
69 #define STATFS_VERSION 0x20030518 /* current version number */
71 uint32_t f_version; /* structure version number */
72 uint32_t f_type; /* type of filesystem */
73 uint64_t f_flags; /* copy of mount exported flags */
74 uint64_t f_bsize; /* filesystem fragment size */
75 uint64_t f_iosize; /* optimal transfer block size */
76 uint64_t f_blocks; /* total data blocks in filesystem */
77 uint64_t f_bfree; /* free blocks in filesystem */
78 int64_t f_bavail; /* free blocks avail to non-superuser */
79 uint64_t f_files; /* total file nodes in filesystem */
80 int64_t f_ffree; /* free nodes avail to non-superuser */
81 uint64_t f_syncwrites; /* count of sync writes since mount */
82 uint64_t f_asyncwrites; /* count of async writes since mount */
83 uint64_t f_syncreads; /* count of sync reads since mount */
84 uint64_t f_asyncreads; /* count of async reads since mount */
85 uint64_t f_spare[10]; /* unused spare */
86 uint32_t f_namemax; /* maximum filename length */
87 uid_t f_owner; /* user that mounted the filesystem */
88 fsid_t f_fsid; /* filesystem id */
89 char f_charspare[80]; /* spare string space */
90 char f_fstypename[MFSNAMELEN]; /* filesystem type name */
91 char f_mntfromname[MNAMELEN]; /* mounted filesystem */
92 char f_mntonname[MNAMELEN]; /* directory on which mounted */
96 #define OMFSNAMELEN 16 /* length of fs type name, including null */
97 #define OMNAMELEN (88 - 2 * sizeof(long)) /* size of on/from name bufs */
99 /* XXX getfsstat.2 is out of date with write and read counter changes here. */
100 /* XXX statfs.2 is out of date with read counter changes here. */
102 long f_spare2; /* placeholder */
103 long f_bsize; /* fundamental filesystem block size */
104 long f_iosize; /* optimal transfer block size */
105 long f_blocks; /* total data blocks in filesystem */
106 long f_bfree; /* free blocks in fs */
107 long f_bavail; /* free blocks avail to non-superuser */
108 long f_files; /* total file nodes in filesystem */
109 long f_ffree; /* free file nodes in fs */
110 fsid_t f_fsid; /* filesystem id */
111 uid_t f_owner; /* user that mounted the filesystem */
112 int f_type; /* type of filesystem */
113 int f_flags; /* copy of mount exported flags */
114 long f_syncwrites; /* count of sync writes since mount */
115 long f_asyncwrites; /* count of async writes since mount */
116 char f_fstypename[OMFSNAMELEN]; /* fs type name */
117 char f_mntonname[OMNAMELEN]; /* directory on which mounted */
118 long f_syncreads; /* count of sync reads since mount */
119 long f_asyncreads; /* count of async reads since mount */
120 short f_spares1; /* unused spare */
121 char f_mntfromname[OMNAMELEN];/* mounted filesystem */
122 short f_spares2; /* unused spare */
124 * XXX on machines where longs are aligned to 8-byte boundaries, there
125 * is an unnamed int32_t here. This spare was after the apparent end
126 * of the struct until we bit off the read counters from f_mntonname.
128 long f_spare[2]; /* unused spare */
131 TAILQ_HEAD(vnodelst, vnode);
133 /* Mount options list */
134 TAILQ_HEAD(vfsoptlist, vfsopt);
136 TAILQ_ENTRY(vfsopt) link;
145 * Structure per mounted filesystem. Each mounted filesystem has an
146 * array of operations and an instance record. The filesystems are
147 * put on a doubly linked list.
153 * v - vnode freelist mutex
155 * Unmarked fields are considered stable as long as a ref is held.
159 struct mtx mnt_mtx; /* mount structure interlock */
160 int mnt_gen; /* struct mount generation */
161 #define mnt_startzero mnt_list
162 TAILQ_ENTRY(mount) mnt_list; /* (m) mount list */
163 struct vfsops *mnt_op; /* operations on fs */
164 struct vfsconf *mnt_vfc; /* configuration info */
165 struct vnode *mnt_vnodecovered; /* vnode we mounted on */
166 struct vnode *mnt_syncer; /* syncer vnode */
167 int mnt_ref; /* (i) Reference count */
168 struct vnodelst mnt_nvnodelist; /* (i) list of vnodes */
169 int mnt_nvnodelistsize; /* (i) # of vnodes */
170 int mnt_writeopcount; /* (i) write syscalls pending */
171 int mnt_kern_flag; /* (i) kernel only flags */
172 uint64_t mnt_flag; /* (i) flags shared with user */
173 struct vfsoptlist *mnt_opt; /* current mount options */
174 struct vfsoptlist *mnt_optnew; /* new options passed to fs */
175 int mnt_maxsymlinklen; /* max size of short symlink */
176 struct statfs mnt_stat; /* cache of filesystem stats */
177 struct ucred *mnt_cred; /* credentials of mounter */
178 void * mnt_data; /* private data */
179 time_t mnt_time; /* last time written*/
180 int mnt_iosize_max; /* max size for clusters, etc */
181 struct netexport *mnt_export; /* export list */
182 struct label *mnt_label; /* MAC label for the fs */
183 u_int mnt_hashseed; /* Random seed for vfs_hash */
184 int mnt_lockref; /* (i) Lock reference count */
185 int mnt_secondary_writes; /* (i) # of secondary writes */
186 int mnt_secondary_accwrites;/* (i) secondary wr. starts */
187 struct thread *mnt_susp_owner; /* (i) thread owning suspension */
188 #define mnt_endzero mnt_gjprovider
189 char *mnt_gjprovider; /* gjournal provider name */
190 struct mtx mnt_listmtx;
191 struct vnodelst mnt_activevnodelist; /* (l) list of active vnodes */
192 int mnt_activevnodelistsize;/* (l) # of active vnodes */
193 struct vnodelst mnt_tmpfreevnodelist; /* (l) list of free vnodes */
194 int mnt_tmpfreevnodelistsize;/* (l) # of free vnodes */
195 struct lock mnt_explock; /* vfs_export walkers lock */
196 TAILQ_ENTRY(mount) mnt_upper_link; /* (m) we in the all uppers */
197 TAILQ_HEAD(, mount) mnt_uppers; /* (m) upper mounts over us*/
201 * Definitions for MNT_VNODE_FOREACH_ALL.
203 struct vnode *__mnt_vnode_next_all(struct vnode **mvp, struct mount *mp);
204 struct vnode *__mnt_vnode_first_all(struct vnode **mvp, struct mount *mp);
205 void __mnt_vnode_markerfree_all(struct vnode **mvp, struct mount *mp);
207 #define MNT_VNODE_FOREACH_ALL(vp, mp, mvp) \
208 for (vp = __mnt_vnode_first_all(&(mvp), (mp)); \
209 (vp) != NULL; vp = __mnt_vnode_next_all(&(mvp), (mp)))
211 #define MNT_VNODE_FOREACH_ALL_ABORT(mp, mvp) \
214 __mnt_vnode_markerfree_all(&(mvp), (mp)); \
215 /* MNT_IUNLOCK(mp); -- done in above function */ \
216 mtx_assert(MNT_MTX(mp), MA_NOTOWNED); \
220 * Definitions for MNT_VNODE_FOREACH_ACTIVE.
222 struct vnode *__mnt_vnode_next_active(struct vnode **mvp, struct mount *mp);
223 struct vnode *__mnt_vnode_first_active(struct vnode **mvp, struct mount *mp);
224 void __mnt_vnode_markerfree_active(struct vnode **mvp, struct mount *);
226 #define MNT_VNODE_FOREACH_ACTIVE(vp, mp, mvp) \
227 for (vp = __mnt_vnode_first_active(&(mvp), (mp)); \
228 (vp) != NULL; vp = __mnt_vnode_next_active(&(mvp), (mp)))
230 #define MNT_VNODE_FOREACH_ACTIVE_ABORT(mp, mvp) \
231 __mnt_vnode_markerfree_active(&(mvp), (mp))
233 #define MNT_ILOCK(mp) mtx_lock(&(mp)->mnt_mtx)
234 #define MNT_ITRYLOCK(mp) mtx_trylock(&(mp)->mnt_mtx)
235 #define MNT_IUNLOCK(mp) mtx_unlock(&(mp)->mnt_mtx)
236 #define MNT_MTX(mp) (&(mp)->mnt_mtx)
237 #define MNT_REF(mp) (mp)->mnt_ref++
238 #define MNT_REL(mp) do { \
239 KASSERT((mp)->mnt_ref > 0, ("negative mnt_ref")); \
241 if ((mp)->mnt_ref == 0) \
248 * User specifiable flags, stored in mnt_flag.
250 #define MNT_RDONLY 0x0000000000000001ULL /* read only filesystem */
251 #define MNT_SYNCHRONOUS 0x0000000000000002ULL /* fs written synchronously */
252 #define MNT_NOEXEC 0x0000000000000004ULL /* can't exec from filesystem */
253 #define MNT_NOSUID 0x0000000000000008ULL /* don't honor setuid fs bits */
254 #define MNT_NFS4ACLS 0x0000000000000010ULL /* enable NFS version 4 ACLs */
255 #define MNT_UNION 0x0000000000000020ULL /* union with underlying fs */
256 #define MNT_ASYNC 0x0000000000000040ULL /* fs written asynchronously */
257 #define MNT_SUIDDIR 0x0000000000100000ULL /* special SUID dir handling */
258 #define MNT_SOFTDEP 0x0000000000200000ULL /* using soft updates */
259 #define MNT_NOSYMFOLLOW 0x0000000000400000ULL /* do not follow symlinks */
260 #define MNT_GJOURNAL 0x0000000002000000ULL /* GEOM journal support enabled */
261 #define MNT_MULTILABEL 0x0000000004000000ULL /* MAC support for objects */
262 #define MNT_ACLS 0x0000000008000000ULL /* ACL support enabled */
263 #define MNT_NOATIME 0x0000000010000000ULL /* dont update file access time */
264 #define MNT_NOCLUSTERR 0x0000000040000000ULL /* disable cluster read */
265 #define MNT_NOCLUSTERW 0x0000000080000000ULL /* disable cluster write */
266 #define MNT_SUJ 0x0000000100000000ULL /* using journaled soft updates */
267 #define MNT_AUTOMOUNTED 0x0000000200000000ULL /* mounted by automountd(8) */
270 * NFS export related mount flags.
272 #define MNT_EXRDONLY 0x0000000000000080ULL /* exported read only */
273 #define MNT_EXPORTED 0x0000000000000100ULL /* filesystem is exported */
274 #define MNT_DEFEXPORTED 0x0000000000000200ULL /* exported to the world */
275 #define MNT_EXPORTANON 0x0000000000000400ULL /* anon uid mapping for all */
276 #define MNT_EXKERB 0x0000000000000800ULL /* exported with Kerberos */
277 #define MNT_EXPUBLIC 0x0000000020000000ULL /* public export (WebNFS) */
280 * Flags set by internal operations,
281 * but visible to the user.
282 * XXX some of these are not quite right.. (I've never seen the root flag set)
284 #define MNT_LOCAL 0x0000000000001000ULL /* filesystem is stored locally */
285 #define MNT_QUOTA 0x0000000000002000ULL /* quotas are enabled on fs */
286 #define MNT_ROOTFS 0x0000000000004000ULL /* identifies the root fs */
287 #define MNT_USER 0x0000000000008000ULL /* mounted by a user */
288 #define MNT_IGNORE 0x0000000000800000ULL /* do not show entry in df */
291 * Mask of flags that are visible to statfs().
292 * XXX I think that this could now become (~(MNT_CMDFLAGS))
293 * but the 'mount' program may need changing to handle this.
295 #define MNT_VISFLAGMASK (MNT_RDONLY | MNT_SYNCHRONOUS | MNT_NOEXEC | \
296 MNT_NOSUID | MNT_UNION | MNT_SUJ | \
297 MNT_ASYNC | MNT_EXRDONLY | MNT_EXPORTED | \
298 MNT_DEFEXPORTED | MNT_EXPORTANON| MNT_EXKERB | \
299 MNT_LOCAL | MNT_USER | MNT_QUOTA | \
300 MNT_ROOTFS | MNT_NOATIME | MNT_NOCLUSTERR| \
301 MNT_NOCLUSTERW | MNT_SUIDDIR | MNT_SOFTDEP | \
302 MNT_IGNORE | MNT_EXPUBLIC | MNT_NOSYMFOLLOW | \
303 MNT_GJOURNAL | MNT_MULTILABEL | MNT_ACLS | \
304 MNT_NFS4ACLS | MNT_AUTOMOUNTED)
306 /* Mask of flags that can be updated. */
307 #define MNT_UPDATEMASK (MNT_NOSUID | MNT_NOEXEC | \
308 MNT_SYNCHRONOUS | MNT_UNION | MNT_ASYNC | \
310 MNT_NOSYMFOLLOW | MNT_IGNORE | \
311 MNT_NOCLUSTERR | MNT_NOCLUSTERW | MNT_SUIDDIR | \
312 MNT_ACLS | MNT_USER | MNT_NFS4ACLS | \
316 * External filesystem command modifier flags.
317 * Unmount can use the MNT_FORCE flag.
318 * XXX: These are not STATES and really should be somewhere else.
319 * XXX: MNT_BYFSID and MNT_NONBUSY collide with MNT_ACLS and MNT_MULTILABEL,
320 * but because MNT_ACLS and MNT_MULTILABEL are only used for mount(2),
321 * and MNT_BYFSID and MNT_NONBUSY are only used for unmount(2),
324 #define MNT_UPDATE 0x0000000000010000ULL /* not real mount, just update */
325 #define MNT_DELEXPORT 0x0000000000020000ULL /* delete export host lists */
326 #define MNT_RELOAD 0x0000000000040000ULL /* reload filesystem data */
327 #define MNT_FORCE 0x0000000000080000ULL /* force unmount or readonly */
328 #define MNT_SNAPSHOT 0x0000000001000000ULL /* snapshot the filesystem */
329 #define MNT_NONBUSY 0x0000000004000000ULL /* check vnode use counts. */
330 #define MNT_BYFSID 0x0000000008000000ULL /* specify filesystem by ID. */
331 #define MNT_CMDFLAGS (MNT_UPDATE | MNT_DELEXPORT | MNT_RELOAD | \
332 MNT_FORCE | MNT_SNAPSHOT | MNT_NONBUSY | \
335 * Internal filesystem control flags stored in mnt_kern_flag.
337 * MNTK_UNMOUNT locks the mount entry so that name lookup cannot proceed
338 * past the mount point. This keeps the subtree stable during mounts
341 * MNTK_UNMOUNTF permits filesystems to detect a forced unmount while
342 * dounmount() is still waiting to lock the mountpoint. This allows
343 * the filesystem to cancel operations that might otherwise deadlock
344 * with the unmount attempt (used by NFS).
346 * MNTK_NOINSMNTQ is strict subset of MNTK_UNMOUNT. They are separated
347 * to allow for failed unmount attempt to restore the syncer vnode for
350 #define MNTK_UNMOUNTF 0x00000001 /* forced unmount in progress */
351 #define MNTK_ASYNC 0x00000002 /* filtered async flag */
352 #define MNTK_SOFTDEP 0x00000004 /* async disabled by softdep */
353 #define MNTK_NOINSMNTQ 0x00000008 /* insmntque is not allowed */
354 #define MNTK_DRAINING 0x00000010 /* lock draining is happening */
355 #define MNTK_REFEXPIRE 0x00000020 /* refcount expiring is happening */
356 #define MNTK_EXTENDED_SHARED 0x00000040 /* Allow shared locking for more ops */
357 #define MNTK_SHARED_WRITES 0x00000080 /* Allow shared locking for writes */
358 #define MNTK_NO_IOPF 0x00000100 /* Disallow page faults during reads
359 and writes. Filesystem shall properly
360 handle i/o state on EFAULT. */
361 #define MNTK_VGONE_UPPER 0x00000200
362 #define MNTK_VGONE_WAITER 0x00000400
363 #define MNTK_LOOKUP_EXCL_DOTDOT 0x00000800
364 #define MNTK_MARKER 0x00001000
365 #define MNTK_UNMAPPED_BUFS 0x00002000
366 #define MNTK_USES_BCACHE 0x00004000 /* FS uses the buffer cache. */
367 #define MNTK_NOASYNC 0x00800000 /* disable async */
368 #define MNTK_UNMOUNT 0x01000000 /* unmount in progress */
369 #define MNTK_MWAIT 0x02000000 /* waiting for unmount to finish */
370 #define MNTK_SUSPEND 0x08000000 /* request write suspension */
371 #define MNTK_SUSPEND2 0x04000000 /* block secondary writes */
372 #define MNTK_SUSPENDED 0x10000000 /* write operations are suspended */
373 #define MNTK_NULL_NOCACHE 0x20000000 /* auto disable cache for nullfs
374 mounts over this fs */
375 #define MNTK_LOOKUP_SHARED 0x40000000 /* FS supports shared lock lookups */
376 #define MNTK_NOKNOTE 0x80000000 /* Don't send KNOTEs from VOP hooks */
380 MNT_SHARED_WRITES(struct mount *mp)
383 return (mp != NULL && (mp->mnt_kern_flag & MNTK_SHARED_WRITES) != 0);
387 MNT_EXTENDED_SHARED(struct mount *mp)
390 return (mp != NULL && (mp->mnt_kern_flag & MNTK_EXTENDED_SHARED) != 0);
395 * Sysctl CTL_VFS definitions.
397 * Second level identifier specifies which filesystem. Second level
398 * identifier VFS_VFSCONF returns information about all filesystems.
399 * Second level identifier VFS_GENERIC is non-terminal.
401 #define VFS_VFSCONF 0 /* get configured filesystems */
402 #define VFS_GENERIC 0 /* generic filesystem information */
404 * Third level identifiers for VFS_GENERIC are given below; third
405 * level identifiers for specific filesystems are given in their
406 * mount specific header files.
408 #define VFS_MAXTYPENUM 1 /* int: highest defined filesystem type */
409 #define VFS_CONF 2 /* struct: vfsconf for filesystem given
413 * Flags for various system call interfaces.
415 * waitfor flags to vfs_sync() and getfsstat()
417 #define MNT_WAIT 1 /* synchronously wait for I/O to complete */
418 #define MNT_NOWAIT 2 /* start all I/O, but do not wait for it */
419 #define MNT_LAZY 3 /* push data not written by filesystem syncer */
420 #define MNT_SUSPEND 4 /* Suspend file system after sync */
423 * Generic file handle
426 fsid_t fh_fsid; /* Filesystem id of mount point */
427 struct fid fh_fid; /* Filesys specific id */
429 typedef struct fhandle fhandle_t;
432 * Old export arguments without security flavor list
434 struct oexport_args {
435 int ex_flags; /* export related flags */
436 uid_t ex_root; /* mapping for root uid */
437 struct xucred ex_anon; /* mapping for anonymous user */
438 struct sockaddr *ex_addr; /* net address to which exported */
439 u_char ex_addrlen; /* and the net address length */
440 struct sockaddr *ex_mask; /* mask of valid bits in saddr */
441 u_char ex_masklen; /* and the smask length */
442 char *ex_indexfile; /* index file for WebNFS URLs */
446 * Export arguments for local filesystem mount calls.
448 #define MAXSECFLAVORS 5
450 int ex_flags; /* export related flags */
451 uid_t ex_root; /* mapping for root uid */
452 struct xucred ex_anon; /* mapping for anonymous user */
453 struct sockaddr *ex_addr; /* net address to which exported */
454 u_char ex_addrlen; /* and the net address length */
455 struct sockaddr *ex_mask; /* mask of valid bits in saddr */
456 u_char ex_masklen; /* and the smask length */
457 char *ex_indexfile; /* index file for WebNFS URLs */
458 int ex_numsecflavors; /* security flavor count */
459 int ex_secflavors[MAXSECFLAVORS]; /* list of security flavors */
463 * Structure holding information for a publicly exported filesystem
464 * (WebNFS). Currently the specs allow just for one such filesystem.
467 int np_valid; /* Do we hold valid information */
468 fhandle_t np_handle; /* Filehandle for pub fs (internal) */
469 struct mount *np_mount; /* Mountpoint of exported fs */
470 char *np_index; /* Index file */
474 * Filesystem configuration information. One of these exists for each
475 * type of filesystem supported by the kernel. These are searched at
476 * mount time to identify the requested filesystem.
478 * XXX: Never change the first two arguments!
481 u_int vfc_version; /* ABI version number */
482 char vfc_name[MFSNAMELEN]; /* filesystem type name */
483 struct vfsops *vfc_vfsops; /* filesystem operations vector */
484 int vfc_typenum; /* historic filesystem type number */
485 int vfc_refcount; /* number mounted of this type */
486 int vfc_flags; /* permanent flags */
487 struct vfsoptdecl *vfc_opts; /* mount options */
488 TAILQ_ENTRY(vfsconf) vfc_list; /* list of vfscons */
491 /* Userland version of the struct vfsconf. */
493 struct vfsops *vfc_vfsops; /* filesystem operations vector */
494 char vfc_name[MFSNAMELEN]; /* filesystem type name */
495 int vfc_typenum; /* historic filesystem type number */
496 int vfc_refcount; /* number mounted of this type */
497 int vfc_flags; /* permanent flags */
498 struct vfsconf *vfc_next; /* next in list */
512 * NB: these flags refer to IMPLEMENTATION properties, not properties of
513 * any actual mounts; i.e., it does not make sense to change the flags.
515 #define VFCF_STATIC 0x00010000 /* statically compiled into kernel */
516 #define VFCF_NETWORK 0x00020000 /* may get data over the network */
517 #define VFCF_READONLY 0x00040000 /* writes are not implemented */
518 #define VFCF_SYNTHETIC 0x00080000 /* data does not represent real files */
519 #define VFCF_LOOPBACK 0x00100000 /* aliases some other mounted FS */
520 #define VFCF_UNICODE 0x00200000 /* stores file names as Unicode */
521 #define VFCF_JAIL 0x00400000 /* can be mounted from within a jail */
522 #define VFCF_DELEGADMIN 0x00800000 /* supports delegated administration */
523 #define VFCF_SBDRY 0x01000000 /* defer stop requests */
525 typedef uint32_t fsctlop_t;
528 int vc_vers; /* should be VFSIDCTL_VERS1 (below) */
529 fsid_t vc_fsid; /* fsid to operate on */
530 char vc_fstypename[MFSNAMELEN];
531 /* type of fs 'nfs' or '*' */
532 fsctlop_t vc_op; /* operation VFS_CTL_* (below) */
533 void *vc_ptr; /* pointer to data structure */
534 size_t vc_len; /* sizeof said structure */
535 u_int32_t vc_spare[12]; /* spare (must be zero) */
538 /* vfsidctl API version. */
539 #define VFS_CTL_VERS1 0x01
542 * New style VFS sysctls, do not reuse/conflict with the namespace for
544 * All "global" sysctl ops have the 33rd bit set:
546 * Private sysctl ops should have the 33rd bit unset.
548 #define VFS_CTL_QUERY 0x00010001 /* anything wrong? (vfsquery) */
549 #define VFS_CTL_TIMEO 0x00010002 /* set timeout for vfs notification */
550 #define VFS_CTL_NOLOCKS 0x00010003 /* disable file locking */
554 u_int32_t vq_spare[31];
558 #define VQ_NOTRESP 0x0001 /* server down */
559 #define VQ_NEEDAUTH 0x0002 /* server bad auth */
560 #define VQ_LOWDISK 0x0004 /* we're low on space */
561 #define VQ_MOUNT 0x0008 /* new filesystem arrived */
562 #define VQ_UNMOUNT 0x0010 /* filesystem has left */
563 #define VQ_DEAD 0x0020 /* filesystem is dead, needs force unmount */
564 #define VQ_ASSIST 0x0040 /* filesystem needs assistance from external
566 #define VQ_NOTRESPLOCK 0x0080 /* server lockd down */
567 #define VQ_FLAG0100 0x0100 /* placeholder */
568 #define VQ_FLAG0200 0x0200 /* placeholder */
569 #define VQ_FLAG0400 0x0400 /* placeholder */
570 #define VQ_FLAG0800 0x0800 /* placeholder */
571 #define VQ_FLAG1000 0x1000 /* placeholder */
572 #define VQ_FLAG2000 0x2000 /* placeholder */
573 #define VQ_FLAG4000 0x4000 /* placeholder */
574 #define VQ_FLAG8000 0x8000 /* placeholder */
577 /* Point a sysctl request at a vfsidctl's data. */
578 #define VCTLTOREQ(vc, req) \
580 (req)->newptr = (vc)->vc_ptr; \
581 (req)->newlen = (vc)->vc_len; \
592 * vfs_busy specific flags and mask.
594 #define MBF_NOWAIT 0x01
595 #define MBF_MNTLSTLOCK 0x02
596 #define MBF_MASK (MBF_NOWAIT | MBF_MNTLSTLOCK)
598 #ifdef MALLOC_DECLARE
599 MALLOC_DECLARE(M_MOUNT);
601 extern int maxvfsconf; /* highest defined filesystem type */
603 TAILQ_HEAD(vfsconfhead, vfsconf);
604 extern struct vfsconfhead vfsconf;
607 * Operations supported on mounted filesystem.
614 typedef int vfs_cmount_t(struct mntarg *ma, void *data, uint64_t flags);
615 typedef int vfs_unmount_t(struct mount *mp, int mntflags);
616 typedef int vfs_root_t(struct mount *mp, int flags, struct vnode **vpp);
617 typedef int vfs_quotactl_t(struct mount *mp, int cmds, uid_t uid, void *arg);
618 typedef int vfs_statfs_t(struct mount *mp, struct statfs *sbp);
619 typedef int vfs_sync_t(struct mount *mp, int waitfor);
620 typedef int vfs_vget_t(struct mount *mp, ino_t ino, int flags,
622 typedef int vfs_fhtovp_t(struct mount *mp, struct fid *fhp,
623 int flags, struct vnode **vpp);
624 typedef int vfs_checkexp_t(struct mount *mp, struct sockaddr *nam,
625 int *extflagsp, struct ucred **credanonp,
626 int *numsecflavors, int **secflavors);
627 typedef int vfs_init_t(struct vfsconf *);
628 typedef int vfs_uninit_t(struct vfsconf *);
629 typedef int vfs_extattrctl_t(struct mount *mp, int cmd,
630 struct vnode *filename_vp, int attrnamespace,
631 const char *attrname);
632 typedef int vfs_mount_t(struct mount *mp);
633 typedef int vfs_sysctl_t(struct mount *mp, fsctlop_t op,
634 struct sysctl_req *req);
635 typedef void vfs_susp_clean_t(struct mount *mp);
636 typedef void vfs_notify_lowervp_t(struct mount *mp, struct vnode *lowervp);
637 typedef void vfs_purge_t(struct mount *mp);
640 vfs_mount_t *vfs_mount;
641 vfs_cmount_t *vfs_cmount;
642 vfs_unmount_t *vfs_unmount;
643 vfs_root_t *vfs_root;
644 vfs_quotactl_t *vfs_quotactl;
645 vfs_statfs_t *vfs_statfs;
646 vfs_sync_t *vfs_sync;
647 vfs_vget_t *vfs_vget;
648 vfs_fhtovp_t *vfs_fhtovp;
649 vfs_checkexp_t *vfs_checkexp;
650 vfs_init_t *vfs_init;
651 vfs_uninit_t *vfs_uninit;
652 vfs_extattrctl_t *vfs_extattrctl;
653 vfs_sysctl_t *vfs_sysctl;
654 vfs_susp_clean_t *vfs_susp_clean;
655 vfs_notify_lowervp_t *vfs_reclaim_lowervp;
656 vfs_notify_lowervp_t *vfs_unlink_lowervp;
657 vfs_purge_t *vfs_purge;
658 vfs_mount_t *vfs_spare[6]; /* spares for ABI compat */
661 vfs_statfs_t __vfs_statfs;
663 #define VFS_PROLOGUE(MP) do { \
664 struct mount *mp__; \
668 _prev_stops = sigdeferstop((mp__ != NULL && \
669 (mp__->mnt_vfc->vfc_flags & VFCF_SBDRY) != 0) ? \
670 SIGDEFERSTOP_SILENT : SIGDEFERSTOP_NOP);
672 #define VFS_EPILOGUE(MP) \
673 sigallowstop(_prev_stops); \
676 #define VFS_MOUNT(MP) ({ \
680 _rc = (*(MP)->mnt_op->vfs_mount)(MP); \
684 #define VFS_UNMOUNT(MP, FORCE) ({ \
688 _rc = (*(MP)->mnt_op->vfs_unmount)(MP, FORCE); \
692 #define VFS_ROOT(MP, FLAGS, VPP) ({ \
696 _rc = (*(MP)->mnt_op->vfs_root)(MP, FLAGS, VPP); \
700 #define VFS_QUOTACTL(MP, C, U, A) ({ \
704 _rc = (*(MP)->mnt_op->vfs_quotactl)(MP, C, U, A); \
708 #define VFS_STATFS(MP, SBP) ({ \
712 _rc = __vfs_statfs((MP), (SBP)); \
716 #define VFS_SYNC(MP, WAIT) ({ \
720 _rc = (*(MP)->mnt_op->vfs_sync)(MP, WAIT); \
724 #define VFS_VGET(MP, INO, FLAGS, VPP) ({ \
728 _rc = (*(MP)->mnt_op->vfs_vget)(MP, INO, FLAGS, VPP); \
732 #define VFS_FHTOVP(MP, FIDP, FLAGS, VPP) ({ \
736 _rc = (*(MP)->mnt_op->vfs_fhtovp)(MP, FIDP, FLAGS, VPP); \
740 #define VFS_CHECKEXP(MP, NAM, EXFLG, CRED, NUMSEC, SEC) ({ \
744 _rc = (*(MP)->mnt_op->vfs_checkexp)(MP, NAM, EXFLG, CRED, NUMSEC,\
749 #define VFS_EXTATTRCTL(MP, C, FN, NS, N) ({ \
753 _rc = (*(MP)->mnt_op->vfs_extattrctl)(MP, C, FN, NS, N); \
757 #define VFS_SYSCTL(MP, OP, REQ) ({ \
761 _rc = (*(MP)->mnt_op->vfs_sysctl)(MP, OP, REQ); \
765 #define VFS_SUSP_CLEAN(MP) do { \
766 if (*(MP)->mnt_op->vfs_susp_clean != NULL) { \
768 (*(MP)->mnt_op->vfs_susp_clean)(MP); \
773 #define VFS_RECLAIM_LOWERVP(MP, VP) do { \
774 if (*(MP)->mnt_op->vfs_reclaim_lowervp != NULL) { \
776 (*(MP)->mnt_op->vfs_reclaim_lowervp)((MP), (VP)); \
781 #define VFS_UNLINK_LOWERVP(MP, VP) do { \
782 if (*(MP)->mnt_op->vfs_unlink_lowervp != NULL) { \
784 (*(MP)->mnt_op->vfs_unlink_lowervp)((MP), (VP)); \
789 #define VFS_PURGE(MP) do { \
790 if (*(MP)->mnt_op->vfs_purge != NULL) { \
792 (*(MP)->mnt_op->vfs_purge)(MP); \
797 #define VFS_KNOTE_LOCKED(vp, hint) do \
799 if (((vp)->v_vflag & VV_NOKNOTE) == 0) \
800 VN_KNOTE((vp), (hint), KNF_LISTLOCKED); \
803 #define VFS_KNOTE_UNLOCKED(vp, hint) do \
805 if (((vp)->v_vflag & VV_NOKNOTE) == 0) \
806 VN_KNOTE((vp), (hint), 0); \
809 #define VFS_NOTIFY_UPPER_RECLAIM 1
810 #define VFS_NOTIFY_UPPER_UNLINK 2
812 #include <sys/module.h>
817 #define VFS_VERSION_00 0x19660120
818 #define VFS_VERSION_01 0x20121030
819 #define VFS_VERSION VFS_VERSION_01
821 #define VFS_SET(vfsops, fsname, flags) \
822 static struct vfsconf fsname ## _vfsconf = { \
823 .vfc_version = VFS_VERSION, \
824 .vfc_name = #fsname, \
825 .vfc_vfsops = &vfsops, \
827 .vfc_flags = flags, \
829 static moduledata_t fsname ## _mod = { \
832 & fsname ## _vfsconf \
834 DECLARE_MODULE(fsname, fsname ## _mod, SI_SUB_VFS, SI_ORDER_MIDDLE)
837 * exported vnode operations
840 int dounmount(struct mount *, int, struct thread *);
842 int kernel_mount(struct mntarg *ma, uint64_t flags);
843 int kernel_vmount(int flags, ...);
844 struct mntarg *mount_arg(struct mntarg *ma, const char *name, const void *val, int len);
845 struct mntarg *mount_argb(struct mntarg *ma, int flag, const char *name);
846 struct mntarg *mount_argf(struct mntarg *ma, const char *name, const char *fmt, ...);
847 struct mntarg *mount_argsu(struct mntarg *ma, const char *name, const void *val, int len);
848 void statfs_scale_blocks(struct statfs *sf, long max_size);
849 struct vfsconf *vfs_byname(const char *);
850 struct vfsconf *vfs_byname_kld(const char *, struct thread *td, int *);
851 void vfs_mount_destroy(struct mount *);
852 void vfs_event_signal(fsid_t *, u_int32_t, intptr_t);
853 void vfs_freeopts(struct vfsoptlist *opts);
854 void vfs_deleteopt(struct vfsoptlist *opts, const char *name);
855 int vfs_buildopts(struct uio *auio, struct vfsoptlist **options);
856 int vfs_flagopt(struct vfsoptlist *opts, const char *name, uint64_t *w,
858 int vfs_getopt(struct vfsoptlist *, const char *, void **, int *);
859 int vfs_getopt_pos(struct vfsoptlist *opts, const char *name);
860 int vfs_getopt_size(struct vfsoptlist *opts, const char *name,
862 char *vfs_getopts(struct vfsoptlist *, const char *, int *error);
863 int vfs_copyopt(struct vfsoptlist *, const char *, void *, int);
864 int vfs_filteropt(struct vfsoptlist *, const char **legal);
865 void vfs_opterror(struct vfsoptlist *opts, const char *fmt, ...);
866 int vfs_scanopt(struct vfsoptlist *opts, const char *name, const char *fmt, ...);
867 int vfs_setopt(struct vfsoptlist *opts, const char *name, void *value,
869 int vfs_setopt_part(struct vfsoptlist *opts, const char *name, void *value,
871 int vfs_setopts(struct vfsoptlist *opts, const char *name,
873 int vfs_setpublicfs /* set publicly exported fs */
874 (struct mount *, struct netexport *, struct export_args *);
875 void vfs_msync(struct mount *, int);
876 int vfs_busy(struct mount *, int);
877 int vfs_export /* process mount export info */
878 (struct mount *, struct export_args *);
879 void vfs_allocate_syncvnode(struct mount *);
880 void vfs_deallocate_syncvnode(struct mount *);
881 int vfs_donmount(struct thread *td, uint64_t fsflags,
882 struct uio *fsoptions);
883 void vfs_getnewfsid(struct mount *);
884 struct cdev *vfs_getrootfsid(struct mount *);
885 struct mount *vfs_getvfs(fsid_t *); /* return vfs given fsid */
886 struct mount *vfs_busyfs(fsid_t *);
887 int vfs_modevent(module_t, int, void *);
888 void vfs_mount_error(struct mount *, const char *, ...);
889 void vfs_mountroot(void); /* mount our root filesystem */
890 void vfs_mountedfrom(struct mount *, const char *from);
891 void vfs_notify_upper(struct vnode *, int);
892 void vfs_oexport_conv(const struct oexport_args *oexp,
893 struct export_args *exp);
894 void vfs_ref(struct mount *);
895 void vfs_rel(struct mount *);
896 struct mount *vfs_mount_alloc(struct vnode *, struct vfsconf *, const char *,
898 int vfs_suser(struct mount *, struct thread *);
899 void vfs_unbusy(struct mount *);
900 void vfs_unmountall(void);
901 extern TAILQ_HEAD(mntlist, mount) mountlist; /* mounted filesystem list */
902 extern struct mtx mountlist_mtx;
903 extern struct nfs_public nfs_pub;
904 extern struct sx vfsconf_sx;
905 #define vfsconf_lock() sx_xlock(&vfsconf_sx)
906 #define vfsconf_unlock() sx_xunlock(&vfsconf_sx)
907 #define vfsconf_slock() sx_slock(&vfsconf_sx)
908 #define vfsconf_sunlock() sx_sunlock(&vfsconf_sx)
911 * Declarations for these vfs default operations are located in
912 * kern/vfs_default.c. They will be automatically used to replace
913 * null entries in VFS ops tables when registering a new filesystem
914 * type in the global table.
916 vfs_root_t vfs_stdroot;
917 vfs_quotactl_t vfs_stdquotactl;
918 vfs_statfs_t vfs_stdstatfs;
919 vfs_sync_t vfs_stdsync;
920 vfs_sync_t vfs_stdnosync;
921 vfs_vget_t vfs_stdvget;
922 vfs_fhtovp_t vfs_stdfhtovp;
923 vfs_checkexp_t vfs_stdcheckexp;
924 vfs_init_t vfs_stdinit;
925 vfs_uninit_t vfs_stduninit;
926 vfs_extattrctl_t vfs_stdextattrctl;
927 vfs_sysctl_t vfs_stdsysctl;
929 void syncer_suspend(void);
930 void syncer_resume(void);
934 #include <sys/cdefs.h>
939 int fhopen(const struct fhandle *, int);
940 int fhstat(const struct fhandle *, struct stat *);
941 int fhstatfs(const struct fhandle *, struct statfs *);
942 int fstatfs(int, struct statfs *);
943 int getfh(const char *, fhandle_t *);
944 int getfsstat(struct statfs *, long, int);
945 int getmntinfo(struct statfs **, int);
946 int lgetfh(const char *, fhandle_t *);
947 int mount(const char *, const char *, int, void *);
948 int nmount(struct iovec *, unsigned int, int);
949 int statfs(const char *, struct statfs *);
950 int unmount(const char *, int);
952 /* C library stuff */
953 int getvfsbyname(const char *, struct xvfsconf *);
958 #endif /* !_SYS_MOUNT_H_ */