mount.h revision 330897
1/*-
2 * SPDX-License-Identifier: BSD-3-Clause
3 *
4 * Copyright (c) 1989, 1991, 1993
5 *	The Regents of the University of California.  All rights reserved.
6 *
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
9 * are met:
10 * 1. Redistributions of source code must retain the above copyright
11 *    notice, this list of conditions and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 *    notice, this list of conditions and the following disclaimer in the
14 *    documentation and/or other materials provided with the distribution.
15 * 4. Neither the name of the University nor the names of its contributors
16 *    may be used to endorse or promote products derived from this software
17 *    without specific prior written permission.
18 *
19 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 * SUCH DAMAGE.
30 *
31 *	@(#)mount.h	8.21 (Berkeley) 5/20/95
32 * $FreeBSD: stable/11/sys/sys/mount.h 330897 2018-03-14 03:19:51Z eadler $
33 */
34
35#ifndef _SYS_MOUNT_H_
36#define _SYS_MOUNT_H_
37
38#include <sys/ucred.h>
39#include <sys/queue.h>
40#ifdef _KERNEL
41#include <sys/lock.h>
42#include <sys/lockmgr.h>
43#include <sys/_mutex.h>
44#include <sys/_sx.h>
45#endif
46
47/*
48 * NOTE: When changing statfs structure, mount structure, MNT_* flags or
49 * MNTK_* flags also update DDB show mount command in vfs_subr.c.
50 */
51
52typedef struct fsid { int32_t val[2]; } fsid_t;	/* filesystem id type */
53
54/*
55 * File identifier.
56 * These are unique per filesystem on a single machine.
57 */
58#define	MAXFIDSZ	16
59
60struct fid {
61	u_short		fid_len;		/* length of data in bytes */
62	u_short		fid_data0;		/* force longword alignment */
63	char		fid_data[MAXFIDSZ];	/* data (variable length) */
64};
65
66/*
67 * filesystem statistics
68 */
69#define	MFSNAMELEN	16		/* length of type name including null */
70#define	MNAMELEN	88		/* size of on/from name bufs */
71#define	STATFS_VERSION	0x20030518	/* current version number */
72struct statfs {
73	uint32_t f_version;		/* structure version number */
74	uint32_t f_type;		/* type of filesystem */
75	uint64_t f_flags;		/* copy of mount exported flags */
76	uint64_t f_bsize;		/* filesystem fragment size */
77	uint64_t f_iosize;		/* optimal transfer block size */
78	uint64_t f_blocks;		/* total data blocks in filesystem */
79	uint64_t f_bfree;		/* free blocks in filesystem */
80	int64_t	 f_bavail;		/* free blocks avail to non-superuser */
81	uint64_t f_files;		/* total file nodes in filesystem */
82	int64_t	 f_ffree;		/* free nodes avail to non-superuser */
83	uint64_t f_syncwrites;		/* count of sync writes since mount */
84	uint64_t f_asyncwrites;		/* count of async writes since mount */
85	uint64_t f_syncreads;		/* count of sync reads since mount */
86	uint64_t f_asyncreads;		/* count of async reads since mount */
87	uint64_t f_spare[10];		/* unused spare */
88	uint32_t f_namemax;		/* maximum filename length */
89	uid_t	  f_owner;		/* user that mounted the filesystem */
90	fsid_t	  f_fsid;		/* filesystem id */
91	char	  f_charspare[80];	    /* spare string space */
92	char	  f_fstypename[MFSNAMELEN]; /* filesystem type name */
93	char	  f_mntfromname[MNAMELEN];  /* mounted filesystem */
94	char	  f_mntonname[MNAMELEN];    /* directory on which mounted */
95};
96
97#ifdef _KERNEL
98#define	OMFSNAMELEN	16	/* length of fs type name, including null */
99#define	OMNAMELEN	(88 - 2 * sizeof(long))	/* size of on/from name bufs */
100
101/* XXX getfsstat.2 is out of date with write and read counter changes here. */
102/* XXX statfs.2 is out of date with read counter changes here. */
103struct ostatfs {
104	long	f_spare2;		/* placeholder */
105	long	f_bsize;		/* fundamental filesystem block size */
106	long	f_iosize;		/* optimal transfer block size */
107	long	f_blocks;		/* total data blocks in filesystem */
108	long	f_bfree;		/* free blocks in fs */
109	long	f_bavail;		/* free blocks avail to non-superuser */
110	long	f_files;		/* total file nodes in filesystem */
111	long	f_ffree;		/* free file nodes in fs */
112	fsid_t	f_fsid;			/* filesystem id */
113	uid_t	f_owner;		/* user that mounted the filesystem */
114	int	f_type;			/* type of filesystem */
115	int	f_flags;		/* copy of mount exported flags */
116	long	f_syncwrites;		/* count of sync writes since mount */
117	long	f_asyncwrites;		/* count of async writes since mount */
118	char	f_fstypename[OMFSNAMELEN]; /* fs type name */
119	char	f_mntonname[OMNAMELEN];	/* directory on which mounted */
120	long	f_syncreads;		/* count of sync reads since mount */
121	long	f_asyncreads;		/* count of async reads since mount */
122	short	f_spares1;		/* unused spare */
123	char	f_mntfromname[OMNAMELEN];/* mounted filesystem */
124	short	f_spares2;		/* unused spare */
125	/*
126	 * XXX on machines where longs are aligned to 8-byte boundaries, there
127	 * is an unnamed int32_t here.  This spare was after the apparent end
128	 * of the struct until we bit off the read counters from f_mntonname.
129	 */
130	long	f_spare[2];		/* unused spare */
131};
132
133TAILQ_HEAD(vnodelst, vnode);
134
135/* Mount options list */
136TAILQ_HEAD(vfsoptlist, vfsopt);
137struct vfsopt {
138	TAILQ_ENTRY(vfsopt) link;
139	char	*name;
140	void	*value;
141	int	len;
142	int	pos;
143	int	seen;
144};
145
146/*
147 * Structure per mounted filesystem.  Each mounted filesystem has an
148 * array of operations and an instance record.  The filesystems are
149 * put on a doubly linked list.
150 *
151 * Lock reference:
152 *	m - mountlist_mtx
153 *	i - interlock
154 *	v - vnode freelist mutex
155 *
156 * Unmarked fields are considered stable as long as a ref is held.
157 *
158 */
159struct mount {
160	struct mtx	mnt_mtx;		/* mount structure interlock */
161	int		mnt_gen;		/* struct mount generation */
162#define	mnt_startzero	mnt_list
163	TAILQ_ENTRY(mount) mnt_list;		/* (m) mount list */
164	struct vfsops	*mnt_op;		/* operations on fs */
165	struct vfsconf	*mnt_vfc;		/* configuration info */
166	struct vnode	*mnt_vnodecovered;	/* vnode we mounted on */
167	struct vnode	*mnt_syncer;		/* syncer vnode */
168	int		mnt_ref;		/* (i) Reference count */
169	struct vnodelst	mnt_nvnodelist;		/* (i) list of vnodes */
170	int		mnt_nvnodelistsize;	/* (i) # of vnodes */
171	struct vnodelst	mnt_activevnodelist;	/* (v) list of active vnodes */
172	int		mnt_activevnodelistsize;/* (v) # of active vnodes */
173	int		mnt_writeopcount;	/* (i) write syscalls pending */
174	int		mnt_kern_flag;		/* (i) kernel only flags */
175	uint64_t	mnt_flag;		/* (i) flags shared with user */
176	struct vfsoptlist *mnt_opt;		/* current mount options */
177	struct vfsoptlist *mnt_optnew;		/* new options passed to fs */
178	int		mnt_maxsymlinklen;	/* max size of short symlink */
179	struct statfs	mnt_stat;		/* cache of filesystem stats */
180	struct ucred	*mnt_cred;		/* credentials of mounter */
181	void *		mnt_data;		/* private data */
182	time_t		mnt_time;		/* last time written*/
183	int		mnt_iosize_max;		/* max size for clusters, etc */
184	struct netexport *mnt_export;		/* export list */
185	struct label	*mnt_label;		/* MAC label for the fs */
186	u_int		mnt_hashseed;		/* Random seed for vfs_hash */
187	int		mnt_lockref;		/* (i) Lock reference count */
188	int		mnt_secondary_writes;   /* (i) # of secondary writes */
189	int		mnt_secondary_accwrites;/* (i) secondary wr. starts */
190	struct thread	*mnt_susp_owner;	/* (i) thread owning suspension */
191#define	mnt_endzero	mnt_gjprovider
192	char		*mnt_gjprovider;	/* gjournal provider name */
193	struct lock	mnt_explock;		/* vfs_export walkers lock */
194	TAILQ_ENTRY(mount) mnt_upper_link;	/* (m) we in the all uppers */
195	TAILQ_HEAD(, mount) mnt_uppers;		/* (m) upper mounts over us*/
196};
197
198/*
199 * Definitions for MNT_VNODE_FOREACH_ALL.
200 */
201struct vnode *__mnt_vnode_next_all(struct vnode **mvp, struct mount *mp);
202struct vnode *__mnt_vnode_first_all(struct vnode **mvp, struct mount *mp);
203void          __mnt_vnode_markerfree_all(struct vnode **mvp, struct mount *mp);
204
205#define MNT_VNODE_FOREACH_ALL(vp, mp, mvp)				\
206	for (vp = __mnt_vnode_first_all(&(mvp), (mp));			\
207		(vp) != NULL; vp = __mnt_vnode_next_all(&(mvp), (mp)))
208
209#define MNT_VNODE_FOREACH_ALL_ABORT(mp, mvp)				\
210	do {								\
211		MNT_ILOCK(mp);						\
212		__mnt_vnode_markerfree_all(&(mvp), (mp));		\
213		/* MNT_IUNLOCK(mp); -- done in above function */	\
214		mtx_assert(MNT_MTX(mp), MA_NOTOWNED);			\
215	} while (0)
216
217/*
218 * Definitions for MNT_VNODE_FOREACH_ACTIVE.
219 */
220struct vnode *__mnt_vnode_next_active(struct vnode **mvp, struct mount *mp);
221struct vnode *__mnt_vnode_first_active(struct vnode **mvp, struct mount *mp);
222void          __mnt_vnode_markerfree_active(struct vnode **mvp, struct mount *);
223
224#define MNT_VNODE_FOREACH_ACTIVE(vp, mp, mvp) 				\
225	for (vp = __mnt_vnode_first_active(&(mvp), (mp)); 		\
226		(vp) != NULL; vp = __mnt_vnode_next_active(&(mvp), (mp)))
227
228#define MNT_VNODE_FOREACH_ACTIVE_ABORT(mp, mvp)				\
229	__mnt_vnode_markerfree_active(&(mvp), (mp))
230
231#define	MNT_ILOCK(mp)	mtx_lock(&(mp)->mnt_mtx)
232#define	MNT_ITRYLOCK(mp) mtx_trylock(&(mp)->mnt_mtx)
233#define	MNT_IUNLOCK(mp)	mtx_unlock(&(mp)->mnt_mtx)
234#define	MNT_MTX(mp)	(&(mp)->mnt_mtx)
235#define	MNT_REF(mp)	(mp)->mnt_ref++
236#define	MNT_REL(mp)	do {						\
237	KASSERT((mp)->mnt_ref > 0, ("negative mnt_ref"));		\
238	(mp)->mnt_ref--;						\
239	if ((mp)->mnt_ref == 0)						\
240		wakeup((mp));						\
241} while (0)
242
243#endif /* _KERNEL */
244
245/*
246 * User specifiable flags, stored in mnt_flag.
247 */
248#define	MNT_RDONLY	0x0000000000000001ULL /* read only filesystem */
249#define	MNT_SYNCHRONOUS	0x0000000000000002ULL /* fs written synchronously */
250#define	MNT_NOEXEC	0x0000000000000004ULL /* can't exec from filesystem */
251#define	MNT_NOSUID	0x0000000000000008ULL /* don't honor setuid fs bits */
252#define	MNT_NFS4ACLS	0x0000000000000010ULL /* enable NFS version 4 ACLs */
253#define	MNT_UNION	0x0000000000000020ULL /* union with underlying fs */
254#define	MNT_ASYNC	0x0000000000000040ULL /* fs written asynchronously */
255#define	MNT_SUIDDIR	0x0000000000100000ULL /* special SUID dir handling */
256#define	MNT_SOFTDEP	0x0000000000200000ULL /* using soft updates */
257#define	MNT_NOSYMFOLLOW	0x0000000000400000ULL /* do not follow symlinks */
258#define	MNT_GJOURNAL	0x0000000002000000ULL /* GEOM journal support enabled */
259#define	MNT_MULTILABEL	0x0000000004000000ULL /* MAC support for objects */
260#define	MNT_ACLS	0x0000000008000000ULL /* ACL support enabled */
261#define	MNT_NOATIME	0x0000000010000000ULL /* dont update file access time */
262#define	MNT_NOCLUSTERR	0x0000000040000000ULL /* disable cluster read */
263#define	MNT_NOCLUSTERW	0x0000000080000000ULL /* disable cluster write */
264#define	MNT_SUJ		0x0000000100000000ULL /* using journaled soft updates */
265#define	MNT_AUTOMOUNTED	0x0000000200000000ULL /* mounted by automountd(8) */
266
267/*
268 * NFS export related mount flags.
269 */
270#define	MNT_EXRDONLY	0x0000000000000080ULL	/* exported read only */
271#define	MNT_EXPORTED	0x0000000000000100ULL	/* filesystem is exported */
272#define	MNT_DEFEXPORTED	0x0000000000000200ULL	/* exported to the world */
273#define	MNT_EXPORTANON	0x0000000000000400ULL	/* anon uid mapping for all */
274#define	MNT_EXKERB	0x0000000000000800ULL	/* exported with Kerberos */
275#define	MNT_EXPUBLIC	0x0000000020000000ULL	/* public export (WebNFS) */
276
277/*
278 * Flags set by internal operations,
279 * but visible to the user.
280 * XXX some of these are not quite right.. (I've never seen the root flag set)
281 */
282#define	MNT_LOCAL	0x0000000000001000ULL /* filesystem is stored locally */
283#define	MNT_QUOTA	0x0000000000002000ULL /* quotas are enabled on fs */
284#define	MNT_ROOTFS	0x0000000000004000ULL /* identifies the root fs */
285#define	MNT_USER	0x0000000000008000ULL /* mounted by a user */
286#define	MNT_IGNORE	0x0000000000800000ULL /* do not show entry in df */
287
288/*
289 * Mask of flags that are visible to statfs().
290 * XXX I think that this could now become (~(MNT_CMDFLAGS))
291 * but the 'mount' program may need changing to handle this.
292 */
293#define	MNT_VISFLAGMASK	(MNT_RDONLY	| MNT_SYNCHRONOUS | MNT_NOEXEC	| \
294			MNT_NOSUID	| MNT_UNION	| MNT_SUJ	| \
295			MNT_ASYNC	| MNT_EXRDONLY	| MNT_EXPORTED	| \
296			MNT_DEFEXPORTED	| MNT_EXPORTANON| MNT_EXKERB	| \
297			MNT_LOCAL	| MNT_USER	| MNT_QUOTA	| \
298			MNT_ROOTFS	| MNT_NOATIME	| MNT_NOCLUSTERR| \
299			MNT_NOCLUSTERW	| MNT_SUIDDIR	| MNT_SOFTDEP	| \
300			MNT_IGNORE	| MNT_EXPUBLIC	| MNT_NOSYMFOLLOW | \
301			MNT_GJOURNAL	| MNT_MULTILABEL | MNT_ACLS	| \
302			MNT_NFS4ACLS	| MNT_AUTOMOUNTED)
303
304/* Mask of flags that can be updated. */
305#define	MNT_UPDATEMASK (MNT_NOSUID	| MNT_NOEXEC	| \
306			MNT_SYNCHRONOUS	| MNT_UNION	| MNT_ASYNC	| \
307			MNT_NOATIME | \
308			MNT_NOSYMFOLLOW	| MNT_IGNORE	| \
309			MNT_NOCLUSTERR	| MNT_NOCLUSTERW | MNT_SUIDDIR	| \
310			MNT_ACLS	| MNT_USER	| MNT_NFS4ACLS	| \
311			MNT_AUTOMOUNTED)
312
313/*
314 * External filesystem command modifier flags.
315 * Unmount can use the MNT_FORCE flag.
316 * XXX: These are not STATES and really should be somewhere else.
317 * XXX: MNT_BYFSID and MNT_NONBUSY collide with MNT_ACLS and MNT_MULTILABEL,
318 *      but because MNT_ACLS and MNT_MULTILABEL are only used for mount(2),
319 *      and MNT_BYFSID and MNT_NONBUSY are only used for unmount(2),
320 *      it's harmless.
321 */
322#define	MNT_UPDATE	0x0000000000010000ULL /* not real mount, just update */
323#define	MNT_DELEXPORT	0x0000000000020000ULL /* delete export host lists */
324#define	MNT_RELOAD	0x0000000000040000ULL /* reload filesystem data */
325#define	MNT_FORCE	0x0000000000080000ULL /* force unmount or readonly */
326#define	MNT_SNAPSHOT	0x0000000001000000ULL /* snapshot the filesystem */
327#define	MNT_NONBUSY	0x0000000004000000ULL /* check vnode use counts. */
328#define	MNT_BYFSID	0x0000000008000000ULL /* specify filesystem by ID. */
329#define MNT_CMDFLAGS   (MNT_UPDATE	| MNT_DELEXPORT	| MNT_RELOAD	| \
330			MNT_FORCE	| MNT_SNAPSHOT	| MNT_NONBUSY	| \
331			MNT_BYFSID)
332/*
333 * Internal filesystem control flags stored in mnt_kern_flag.
334 *
335 * MNTK_UNMOUNT locks the mount entry so that name lookup cannot proceed
336 * past the mount point.  This keeps the subtree stable during mounts
337 * and unmounts.
338 *
339 * MNTK_UNMOUNTF permits filesystems to detect a forced unmount while
340 * dounmount() is still waiting to lock the mountpoint. This allows
341 * the filesystem to cancel operations that might otherwise deadlock
342 * with the unmount attempt (used by NFS).
343 *
344 * MNTK_NOINSMNTQ is strict subset of MNTK_UNMOUNT. They are separated
345 * to allow for failed unmount attempt to restore the syncer vnode for
346 * the mount.
347 */
348#define MNTK_UNMOUNTF	0x00000001	/* forced unmount in progress */
349#define MNTK_ASYNC	0x00000002	/* filtered async flag */
350#define MNTK_SOFTDEP	0x00000004	/* async disabled by softdep */
351#define MNTK_NOINSMNTQ	0x00000008	/* insmntque is not allowed */
352#define	MNTK_DRAINING	0x00000010	/* lock draining is happening */
353#define	MNTK_REFEXPIRE	0x00000020	/* refcount expiring is happening */
354#define MNTK_EXTENDED_SHARED	0x00000040 /* Allow shared locking for more ops */
355#define	MNTK_SHARED_WRITES	0x00000080 /* Allow shared locking for writes */
356#define	MNTK_NO_IOPF	0x00000100	/* Disallow page faults during reads
357					   and writes. Filesystem shall properly
358					   handle i/o state on EFAULT. */
359#define	MNTK_VGONE_UPPER	0x00000200
360#define	MNTK_VGONE_WAITER	0x00000400
361#define	MNTK_LOOKUP_EXCL_DOTDOT	0x00000800
362#define	MNTK_MARKER		0x00001000
363#define	MNTK_UNMAPPED_BUFS	0x00002000
364#define	MNTK_USES_BCACHE	0x00004000 /* FS uses the buffer cache. */
365#define MNTK_NOASYNC	0x00800000	/* disable async */
366#define MNTK_UNMOUNT	0x01000000	/* unmount in progress */
367#define	MNTK_MWAIT	0x02000000	/* waiting for unmount to finish */
368#define	MNTK_SUSPEND	0x08000000	/* request write suspension */
369#define	MNTK_SUSPEND2	0x04000000	/* block secondary writes */
370#define	MNTK_SUSPENDED	0x10000000	/* write operations are suspended */
371#define	MNTK_NULL_NOCACHE	0x20000000 /* auto disable cache for nullfs
372					      mounts over this fs */
373#define MNTK_LOOKUP_SHARED	0x40000000 /* FS supports shared lock lookups */
374#define	MNTK_NOKNOTE	0x80000000	/* Don't send KNOTEs from VOP hooks */
375
376#ifdef _KERNEL
377static inline int
378MNT_SHARED_WRITES(struct mount *mp)
379{
380
381	return (mp != NULL && (mp->mnt_kern_flag & MNTK_SHARED_WRITES) != 0);
382}
383
384static inline int
385MNT_EXTENDED_SHARED(struct mount *mp)
386{
387
388	return (mp != NULL && (mp->mnt_kern_flag & MNTK_EXTENDED_SHARED) != 0);
389}
390#endif
391
392/*
393 * Sysctl CTL_VFS definitions.
394 *
395 * Second level identifier specifies which filesystem. Second level
396 * identifier VFS_VFSCONF returns information about all filesystems.
397 * Second level identifier VFS_GENERIC is non-terminal.
398 */
399#define	VFS_VFSCONF		0	/* get configured filesystems */
400#define	VFS_GENERIC		0	/* generic filesystem information */
401/*
402 * Third level identifiers for VFS_GENERIC are given below; third
403 * level identifiers for specific filesystems are given in their
404 * mount specific header files.
405 */
406#define VFS_MAXTYPENUM	1	/* int: highest defined filesystem type */
407#define VFS_CONF	2	/* struct: vfsconf for filesystem given
408				   as next argument */
409
410/*
411 * Flags for various system call interfaces.
412 *
413 * waitfor flags to vfs_sync() and getfsstat()
414 */
415#define MNT_WAIT	1	/* synchronously wait for I/O to complete */
416#define MNT_NOWAIT	2	/* start all I/O, but do not wait for it */
417#define MNT_LAZY	3	/* push data not written by filesystem syncer */
418#define MNT_SUSPEND	4	/* Suspend file system after sync */
419
420/*
421 * Generic file handle
422 */
423struct fhandle {
424	fsid_t	fh_fsid;	/* Filesystem id of mount point */
425	struct	fid fh_fid;	/* Filesys specific id */
426};
427typedef struct fhandle	fhandle_t;
428
429/*
430 * Old export arguments without security flavor list
431 */
432struct oexport_args {
433	int	ex_flags;		/* export related flags */
434	uid_t	ex_root;		/* mapping for root uid */
435	struct	xucred ex_anon;		/* mapping for anonymous user */
436	struct	sockaddr *ex_addr;	/* net address to which exported */
437	u_char	ex_addrlen;		/* and the net address length */
438	struct	sockaddr *ex_mask;	/* mask of valid bits in saddr */
439	u_char	ex_masklen;		/* and the smask length */
440	char	*ex_indexfile;		/* index file for WebNFS URLs */
441};
442
443/*
444 * Export arguments for local filesystem mount calls.
445 */
446#define	MAXSECFLAVORS	5
447struct export_args {
448	int	ex_flags;		/* export related flags */
449	uid_t	ex_root;		/* mapping for root uid */
450	struct	xucred ex_anon;		/* mapping for anonymous user */
451	struct	sockaddr *ex_addr;	/* net address to which exported */
452	u_char	ex_addrlen;		/* and the net address length */
453	struct	sockaddr *ex_mask;	/* mask of valid bits in saddr */
454	u_char	ex_masklen;		/* and the smask length */
455	char	*ex_indexfile;		/* index file for WebNFS URLs */
456	int	ex_numsecflavors;	/* security flavor count */
457	int	ex_secflavors[MAXSECFLAVORS]; /* list of security flavors */
458};
459
460/*
461 * Structure holding information for a publicly exported filesystem
462 * (WebNFS). Currently the specs allow just for one such filesystem.
463 */
464struct nfs_public {
465	int		np_valid;	/* Do we hold valid information */
466	fhandle_t	np_handle;	/* Filehandle for pub fs (internal) */
467	struct mount	*np_mount;	/* Mountpoint of exported fs */
468	char		*np_index;	/* Index file */
469};
470
471/*
472 * Filesystem configuration information. One of these exists for each
473 * type of filesystem supported by the kernel. These are searched at
474 * mount time to identify the requested filesystem.
475 *
476 * XXX: Never change the first two arguments!
477 */
478struct vfsconf {
479	u_int	vfc_version;		/* ABI version number */
480	char	vfc_name[MFSNAMELEN];	/* filesystem type name */
481	struct	vfsops *vfc_vfsops;	/* filesystem operations vector */
482	int	vfc_typenum;		/* historic filesystem type number */
483	int	vfc_refcount;		/* number mounted of this type */
484	int	vfc_flags;		/* permanent flags */
485	struct	vfsoptdecl *vfc_opts;	/* mount options */
486	TAILQ_ENTRY(vfsconf) vfc_list;	/* list of vfscons */
487};
488
489/* Userland version of the struct vfsconf. */
490struct xvfsconf {
491	struct	vfsops *vfc_vfsops;	/* filesystem operations vector */
492	char	vfc_name[MFSNAMELEN];	/* filesystem type name */
493	int	vfc_typenum;		/* historic filesystem type number */
494	int	vfc_refcount;		/* number mounted of this type */
495	int	vfc_flags;		/* permanent flags */
496	struct	vfsconf *vfc_next;	/* next in list */
497};
498
499#ifndef BURN_BRIDGES
500struct ovfsconf {
501	void	*vfc_vfsops;
502	char	vfc_name[32];
503	int	vfc_index;
504	int	vfc_refcount;
505	int	vfc_flags;
506};
507#endif
508
509/*
510 * NB: these flags refer to IMPLEMENTATION properties, not properties of
511 * any actual mounts; i.e., it does not make sense to change the flags.
512 */
513#define	VFCF_STATIC	0x00010000	/* statically compiled into kernel */
514#define	VFCF_NETWORK	0x00020000	/* may get data over the network */
515#define	VFCF_READONLY	0x00040000	/* writes are not implemented */
516#define	VFCF_SYNTHETIC	0x00080000	/* data does not represent real files */
517#define	VFCF_LOOPBACK	0x00100000	/* aliases some other mounted FS */
518#define	VFCF_UNICODE	0x00200000	/* stores file names as Unicode */
519#define	VFCF_JAIL	0x00400000	/* can be mounted from within a jail */
520#define	VFCF_DELEGADMIN	0x00800000	/* supports delegated administration */
521#define	VFCF_SBDRY	0x01000000	/* defer stop requests */
522
523typedef uint32_t fsctlop_t;
524
525struct vfsidctl {
526	int		vc_vers;	/* should be VFSIDCTL_VERS1 (below) */
527	fsid_t		vc_fsid;	/* fsid to operate on */
528	char		vc_fstypename[MFSNAMELEN];
529					/* type of fs 'nfs' or '*' */
530	fsctlop_t	vc_op;		/* operation VFS_CTL_* (below) */
531	void		*vc_ptr;	/* pointer to data structure */
532	size_t		vc_len;		/* sizeof said structure */
533	u_int32_t	vc_spare[12];	/* spare (must be zero) */
534};
535
536/* vfsidctl API version. */
537#define VFS_CTL_VERS1	0x01
538
539/*
540 * New style VFS sysctls, do not reuse/conflict with the namespace for
541 * private sysctls.
542 * All "global" sysctl ops have the 33rd bit set:
543 * 0x...1....
544 * Private sysctl ops should have the 33rd bit unset.
545 */
546#define VFS_CTL_QUERY	0x00010001	/* anything wrong? (vfsquery) */
547#define VFS_CTL_TIMEO	0x00010002	/* set timeout for vfs notification */
548#define VFS_CTL_NOLOCKS	0x00010003	/* disable file locking */
549
550struct vfsquery {
551	u_int32_t	vq_flags;
552	u_int32_t	vq_spare[31];
553};
554
555/* vfsquery flags */
556#define VQ_NOTRESP	0x0001	/* server down */
557#define VQ_NEEDAUTH	0x0002	/* server bad auth */
558#define VQ_LOWDISK	0x0004	/* we're low on space */
559#define VQ_MOUNT	0x0008	/* new filesystem arrived */
560#define VQ_UNMOUNT	0x0010	/* filesystem has left */
561#define VQ_DEAD		0x0020	/* filesystem is dead, needs force unmount */
562#define VQ_ASSIST	0x0040	/* filesystem needs assistance from external
563				   program */
564#define VQ_NOTRESPLOCK	0x0080	/* server lockd down */
565#define VQ_FLAG0100	0x0100	/* placeholder */
566#define VQ_FLAG0200	0x0200	/* placeholder */
567#define VQ_FLAG0400	0x0400	/* placeholder */
568#define VQ_FLAG0800	0x0800	/* placeholder */
569#define VQ_FLAG1000	0x1000	/* placeholder */
570#define VQ_FLAG2000	0x2000	/* placeholder */
571#define VQ_FLAG4000	0x4000	/* placeholder */
572#define VQ_FLAG8000	0x8000	/* placeholder */
573
574#ifdef _KERNEL
575/* Point a sysctl request at a vfsidctl's data. */
576#define VCTLTOREQ(vc, req)						\
577	do {								\
578		(req)->newptr = (vc)->vc_ptr;				\
579		(req)->newlen = (vc)->vc_len;				\
580		(req)->newidx = 0;					\
581	} while (0)
582#endif
583
584struct iovec;
585struct uio;
586
587#ifdef _KERNEL
588
589/*
590 * vfs_busy specific flags and mask.
591 */
592#define	MBF_NOWAIT	0x01
593#define	MBF_MNTLSTLOCK	0x02
594#define	MBF_MASK	(MBF_NOWAIT | MBF_MNTLSTLOCK)
595
596#ifdef MALLOC_DECLARE
597MALLOC_DECLARE(M_MOUNT);
598MALLOC_DECLARE(M_STATFS);
599#endif
600extern int maxvfsconf;		/* highest defined filesystem type */
601
602TAILQ_HEAD(vfsconfhead, vfsconf);
603extern struct vfsconfhead vfsconf;
604
605/*
606 * Operations supported on mounted filesystem.
607 */
608struct mount_args;
609struct nameidata;
610struct sysctl_req;
611struct mntarg;
612
613typedef int vfs_cmount_t(struct mntarg *ma, void *data, uint64_t flags);
614typedef int vfs_unmount_t(struct mount *mp, int mntflags);
615typedef int vfs_root_t(struct mount *mp, int flags, struct vnode **vpp);
616typedef	int vfs_quotactl_t(struct mount *mp, int cmds, uid_t uid, void *arg);
617typedef	int vfs_statfs_t(struct mount *mp, struct statfs *sbp);
618typedef	int vfs_sync_t(struct mount *mp, int waitfor);
619typedef	int vfs_vget_t(struct mount *mp, ino_t ino, int flags,
620		    struct vnode **vpp);
621typedef	int vfs_fhtovp_t(struct mount *mp, struct fid *fhp,
622		    int flags, struct vnode **vpp);
623typedef	int vfs_checkexp_t(struct mount *mp, struct sockaddr *nam,
624		    int *extflagsp, struct ucred **credanonp,
625		    int *numsecflavors, int **secflavors);
626typedef	int vfs_init_t(struct vfsconf *);
627typedef	int vfs_uninit_t(struct vfsconf *);
628typedef	int vfs_extattrctl_t(struct mount *mp, int cmd,
629		    struct vnode *filename_vp, int attrnamespace,
630		    const char *attrname);
631typedef	int vfs_mount_t(struct mount *mp);
632typedef int vfs_sysctl_t(struct mount *mp, fsctlop_t op,
633		    struct sysctl_req *req);
634typedef void vfs_susp_clean_t(struct mount *mp);
635typedef void vfs_notify_lowervp_t(struct mount *mp, struct vnode *lowervp);
636typedef void vfs_purge_t(struct mount *mp);
637
638struct vfsops {
639	vfs_mount_t		*vfs_mount;
640	vfs_cmount_t		*vfs_cmount;
641	vfs_unmount_t		*vfs_unmount;
642	vfs_root_t		*vfs_root;
643	vfs_quotactl_t		*vfs_quotactl;
644	vfs_statfs_t		*vfs_statfs;
645	vfs_sync_t		*vfs_sync;
646	vfs_vget_t		*vfs_vget;
647	vfs_fhtovp_t		*vfs_fhtovp;
648	vfs_checkexp_t		*vfs_checkexp;
649	vfs_init_t		*vfs_init;
650	vfs_uninit_t		*vfs_uninit;
651	vfs_extattrctl_t	*vfs_extattrctl;
652	vfs_sysctl_t		*vfs_sysctl;
653	vfs_susp_clean_t	*vfs_susp_clean;
654	vfs_notify_lowervp_t	*vfs_reclaim_lowervp;
655	vfs_notify_lowervp_t	*vfs_unlink_lowervp;
656	vfs_purge_t		*vfs_purge;
657	vfs_mount_t		*vfs_spare[6];	/* spares for ABI compat */
658};
659
660vfs_statfs_t	__vfs_statfs;
661
662#define	VFS_PROLOGUE(MP)	do {					\
663	struct mount *mp__;						\
664	int _prev_stops;						\
665									\
666	mp__ = (MP);							\
667	_prev_stops = sigdeferstop((mp__ != NULL &&			\
668	    (mp__->mnt_vfc->vfc_flags & VFCF_SBDRY) != 0) ?		\
669	    SIGDEFERSTOP_SILENT : SIGDEFERSTOP_NOP);
670
671#define	VFS_EPILOGUE(MP)						\
672	sigallowstop(_prev_stops);					\
673} while (0)
674
675#define	VFS_MOUNT(MP) ({						\
676	int _rc;							\
677									\
678	VFS_PROLOGUE(MP);						\
679	_rc = (*(MP)->mnt_op->vfs_mount)(MP);				\
680	VFS_EPILOGUE(MP);						\
681	_rc; })
682
683#define	VFS_UNMOUNT(MP, FORCE) ({					\
684	int _rc;							\
685									\
686	VFS_PROLOGUE(MP);						\
687	_rc = (*(MP)->mnt_op->vfs_unmount)(MP, FORCE);			\
688	VFS_EPILOGUE(MP);						\
689	_rc; })
690
691#define	VFS_ROOT(MP, FLAGS, VPP) ({					\
692	int _rc;							\
693									\
694	VFS_PROLOGUE(MP);						\
695	_rc = (*(MP)->mnt_op->vfs_root)(MP, FLAGS, VPP);		\
696	VFS_EPILOGUE(MP);						\
697	_rc; })
698
699#define	VFS_QUOTACTL(MP, C, U, A) ({					\
700	int _rc;							\
701									\
702	VFS_PROLOGUE(MP);						\
703	_rc = (*(MP)->mnt_op->vfs_quotactl)(MP, C, U, A);		\
704	VFS_EPILOGUE(MP);						\
705	_rc; })
706
707#define	VFS_STATFS(MP, SBP) ({						\
708	int _rc;							\
709									\
710	VFS_PROLOGUE(MP);						\
711	_rc = __vfs_statfs((MP), (SBP));				\
712	VFS_EPILOGUE(MP);						\
713	_rc; })
714
715#define	VFS_SYNC(MP, WAIT) ({						\
716	int _rc;							\
717									\
718	VFS_PROLOGUE(MP);						\
719	_rc = (*(MP)->mnt_op->vfs_sync)(MP, WAIT);			\
720	VFS_EPILOGUE(MP);						\
721	_rc; })
722
723#define	VFS_VGET(MP, INO, FLAGS, VPP) ({				\
724	int _rc;							\
725									\
726	VFS_PROLOGUE(MP);						\
727	_rc = (*(MP)->mnt_op->vfs_vget)(MP, INO, FLAGS, VPP);		\
728	VFS_EPILOGUE(MP);						\
729	_rc; })
730
731#define	VFS_FHTOVP(MP, FIDP, FLAGS, VPP) ({				\
732	int _rc;							\
733									\
734	VFS_PROLOGUE(MP);						\
735	_rc = (*(MP)->mnt_op->vfs_fhtovp)(MP, FIDP, FLAGS, VPP);	\
736	VFS_EPILOGUE(MP);						\
737	_rc; })
738
739#define	VFS_CHECKEXP(MP, NAM, EXFLG, CRED, NUMSEC, SEC) ({		\
740	int _rc;							\
741									\
742	VFS_PROLOGUE(MP);						\
743	_rc = (*(MP)->mnt_op->vfs_checkexp)(MP, NAM, EXFLG, CRED, NUMSEC,\
744	    SEC);							\
745	VFS_EPILOGUE(MP);						\
746	_rc; })
747
748#define	VFS_EXTATTRCTL(MP, C, FN, NS, N) ({				\
749	int _rc;							\
750									\
751	VFS_PROLOGUE(MP);						\
752	_rc = (*(MP)->mnt_op->vfs_extattrctl)(MP, C, FN, NS, N);	\
753	VFS_EPILOGUE(MP);						\
754	_rc; })
755
756#define	VFS_SYSCTL(MP, OP, REQ) ({					\
757	int _rc;							\
758									\
759	VFS_PROLOGUE(MP);						\
760	_rc = (*(MP)->mnt_op->vfs_sysctl)(MP, OP, REQ);			\
761	VFS_EPILOGUE(MP);						\
762	_rc; })
763
764#define	VFS_SUSP_CLEAN(MP) do {						\
765	if (*(MP)->mnt_op->vfs_susp_clean != NULL) {			\
766		VFS_PROLOGUE(MP);					\
767		(*(MP)->mnt_op->vfs_susp_clean)(MP);			\
768		VFS_EPILOGUE(MP);					\
769	}								\
770} while (0)
771
772#define	VFS_RECLAIM_LOWERVP(MP, VP) do {				\
773	if (*(MP)->mnt_op->vfs_reclaim_lowervp != NULL) {		\
774		VFS_PROLOGUE(MP);					\
775		(*(MP)->mnt_op->vfs_reclaim_lowervp)((MP), (VP));	\
776		VFS_EPILOGUE(MP);					\
777	}								\
778} while (0)
779
780#define	VFS_UNLINK_LOWERVP(MP, VP) do {					\
781	if (*(MP)->mnt_op->vfs_unlink_lowervp != NULL) {		\
782		VFS_PROLOGUE(MP);					\
783		(*(MP)->mnt_op->vfs_unlink_lowervp)((MP), (VP));	\
784		VFS_EPILOGUE(MP);					\
785	}								\
786} while (0)
787
788#define	VFS_PURGE(MP) do {						\
789	if (*(MP)->mnt_op->vfs_purge != NULL) {				\
790		VFS_PROLOGUE(MP);					\
791		(*(MP)->mnt_op->vfs_purge)(MP);				\
792		VFS_EPILOGUE(MP);					\
793	}								\
794} while (0)
795
796#define VFS_KNOTE_LOCKED(vp, hint) do					\
797{									\
798	if (((vp)->v_vflag & VV_NOKNOTE) == 0)				\
799		VN_KNOTE((vp), (hint), KNF_LISTLOCKED);			\
800} while (0)
801
802#define VFS_KNOTE_UNLOCKED(vp, hint) do					\
803{									\
804	if (((vp)->v_vflag & VV_NOKNOTE) == 0)				\
805		VN_KNOTE((vp), (hint), 0);				\
806} while (0)
807
808#define	VFS_NOTIFY_UPPER_RECLAIM	1
809#define	VFS_NOTIFY_UPPER_UNLINK		2
810
811#include <sys/module.h>
812
813/*
814 * Version numbers.
815 */
816#define VFS_VERSION_00	0x19660120
817#define VFS_VERSION_01	0x20121030
818#define VFS_VERSION	VFS_VERSION_01
819
820#define VFS_SET(vfsops, fsname, flags) \
821	static struct vfsconf fsname ## _vfsconf = {		\
822		.vfc_version = VFS_VERSION,			\
823		.vfc_name = #fsname,				\
824		.vfc_vfsops = &vfsops,				\
825		.vfc_typenum = -1,				\
826		.vfc_flags = flags,				\
827	};							\
828	static moduledata_t fsname ## _mod = {			\
829		#fsname,					\
830		vfs_modevent,					\
831		& fsname ## _vfsconf				\
832	};							\
833	DECLARE_MODULE(fsname, fsname ## _mod, SI_SUB_VFS, SI_ORDER_MIDDLE)
834
835/*
836 * exported vnode operations
837 */
838
839int	dounmount(struct mount *, int, struct thread *);
840
841int	kernel_mount(struct mntarg *ma, uint64_t flags);
842int	kernel_vmount(int flags, ...);
843struct mntarg *mount_arg(struct mntarg *ma, const char *name, const void *val, int len);
844struct mntarg *mount_argb(struct mntarg *ma, int flag, const char *name);
845struct mntarg *mount_argf(struct mntarg *ma, const char *name, const char *fmt, ...);
846struct mntarg *mount_argsu(struct mntarg *ma, const char *name, const void *val, int len);
847void	statfs_scale_blocks(struct statfs *sf, long max_size);
848struct vfsconf *vfs_byname(const char *);
849struct vfsconf *vfs_byname_kld(const char *, struct thread *td, int *);
850void	vfs_mount_destroy(struct mount *);
851void	vfs_event_signal(fsid_t *, u_int32_t, intptr_t);
852void	vfs_freeopts(struct vfsoptlist *opts);
853void	vfs_deleteopt(struct vfsoptlist *opts, const char *name);
854int	vfs_buildopts(struct uio *auio, struct vfsoptlist **options);
855int	vfs_flagopt(struct vfsoptlist *opts, const char *name, uint64_t *w,
856	    uint64_t val);
857int	vfs_getopt(struct vfsoptlist *, const char *, void **, int *);
858int	vfs_getopt_pos(struct vfsoptlist *opts, const char *name);
859int	vfs_getopt_size(struct vfsoptlist *opts, const char *name,
860	    off_t *value);
861char	*vfs_getopts(struct vfsoptlist *, const char *, int *error);
862int	vfs_copyopt(struct vfsoptlist *, const char *, void *, int);
863int	vfs_filteropt(struct vfsoptlist *, const char **legal);
864void	vfs_opterror(struct vfsoptlist *opts, const char *fmt, ...);
865int	vfs_scanopt(struct vfsoptlist *opts, const char *name, const char *fmt, ...);
866int	vfs_setopt(struct vfsoptlist *opts, const char *name, void *value,
867	    int len);
868int	vfs_setopt_part(struct vfsoptlist *opts, const char *name, void *value,
869	    int len);
870int	vfs_setopts(struct vfsoptlist *opts, const char *name,
871	    const char *value);
872int	vfs_setpublicfs			    /* set publicly exported fs */
873	    (struct mount *, struct netexport *, struct export_args *);
874void	vfs_msync(struct mount *, int);
875int	vfs_busy(struct mount *, int);
876int	vfs_export			 /* process mount export info */
877	    (struct mount *, struct export_args *);
878void	vfs_allocate_syncvnode(struct mount *);
879void	vfs_deallocate_syncvnode(struct mount *);
880int	vfs_donmount(struct thread *td, uint64_t fsflags,
881	    struct uio *fsoptions);
882void	vfs_getnewfsid(struct mount *);
883struct cdev *vfs_getrootfsid(struct mount *);
884struct	mount *vfs_getvfs(fsid_t *);      /* return vfs given fsid */
885struct	mount *vfs_busyfs(fsid_t *);
886int	vfs_modevent(module_t, int, void *);
887void	vfs_mount_error(struct mount *, const char *, ...);
888void	vfs_mountroot(void);			/* mount our root filesystem */
889void	vfs_mountedfrom(struct mount *, const char *from);
890void	vfs_notify_upper(struct vnode *, int);
891void	vfs_oexport_conv(const struct oexport_args *oexp,
892	    struct export_args *exp);
893void	vfs_ref(struct mount *);
894void	vfs_rel(struct mount *);
895struct mount *vfs_mount_alloc(struct vnode *, struct vfsconf *, const char *,
896	    struct ucred *);
897int	vfs_suser(struct mount *, struct thread *);
898void	vfs_unbusy(struct mount *);
899void	vfs_unmountall(void);
900extern	TAILQ_HEAD(mntlist, mount) mountlist;	/* mounted filesystem list */
901extern	struct mtx mountlist_mtx;
902extern	struct nfs_public nfs_pub;
903extern	struct sx vfsconf_sx;
904#define	vfsconf_lock()		sx_xlock(&vfsconf_sx)
905#define	vfsconf_unlock()	sx_xunlock(&vfsconf_sx)
906#define	vfsconf_slock()		sx_slock(&vfsconf_sx)
907#define	vfsconf_sunlock()	sx_sunlock(&vfsconf_sx)
908
909/*
910 * Declarations for these vfs default operations are located in
911 * kern/vfs_default.c.  They will be automatically used to replace
912 * null entries in VFS ops tables when registering a new filesystem
913 * type in the global table.
914 */
915vfs_root_t		vfs_stdroot;
916vfs_quotactl_t		vfs_stdquotactl;
917vfs_statfs_t		vfs_stdstatfs;
918vfs_sync_t		vfs_stdsync;
919vfs_sync_t		vfs_stdnosync;
920vfs_vget_t		vfs_stdvget;
921vfs_fhtovp_t		vfs_stdfhtovp;
922vfs_checkexp_t		vfs_stdcheckexp;
923vfs_init_t		vfs_stdinit;
924vfs_uninit_t		vfs_stduninit;
925vfs_extattrctl_t	vfs_stdextattrctl;
926vfs_sysctl_t		vfs_stdsysctl;
927
928void	syncer_suspend(void);
929void	syncer_resume(void);
930
931#else /* !_KERNEL */
932
933#include <sys/cdefs.h>
934
935struct stat;
936
937__BEGIN_DECLS
938int	fhopen(const struct fhandle *, int);
939int	fhstat(const struct fhandle *, struct stat *);
940int	fhstatfs(const struct fhandle *, struct statfs *);
941int	fstatfs(int, struct statfs *);
942int	getfh(const char *, fhandle_t *);
943int	getfsstat(struct statfs *, long, int);
944int	getmntinfo(struct statfs **, int);
945int	lgetfh(const char *, fhandle_t *);
946int	mount(const char *, const char *, int, void *);
947int	nmount(struct iovec *, unsigned int, int);
948int	statfs(const char *, struct statfs *);
949int	unmount(const char *, int);
950
951/* C library stuff */
952int	getvfsbyname(const char *, struct xvfsconf *);
953__END_DECLS
954
955#endif /* _KERNEL */
956
957#endif /* !_SYS_MOUNT_H_ */
958