xref: /titanic_52/usr/src/boot/sys/sys/mount.h (revision 4a5d661a82b942b6538acd26209d959ce98b593a)
1 /*-
2  * Copyright (c) 1989, 1991, 1993
3  *	The Regents of the University of California.  All rights reserved.
4  *
5  * Redistribution and use in source and binary forms, with or without
6  * modification, are permitted provided that the following conditions
7  * are met:
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.
16  *
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
27  * SUCH DAMAGE.
28  *
29  *	@(#)mount.h	8.21 (Berkeley) 5/20/95
30  * $FreeBSD$
31  */
32 
33 #ifndef _SYS_MOUNT_H_
34 #define _SYS_MOUNT_H_
35 
36 #include <sys/ucred.h>
37 #include <sys/queue.h>
38 #ifdef _KERNEL
39 #include <sys/lock.h>
40 #include <sys/lockmgr.h>
41 #include <sys/_mutex.h>
42 #include <sys/_sx.h>
43 #endif
44 
45 /*
46  * NOTE: When changing statfs structure, mount structure, MNT_* flags or
47  * MNTK_* flags also update DDB show mount command in vfs_subr.c.
48  */
49 
50 typedef struct fsid { int32_t val[2]; } fsid_t;	/* filesystem id type */
51 
52 /*
53  * File identifier.
54  * These are unique per filesystem on a single machine.
55  */
56 #define	MAXFIDSZ	16
57 
58 struct fid {
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) */
62 };
63 
64 /*
65  * filesystem statistics
66  */
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 */
70 struct statfs {
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 */
93 };
94 
95 #ifdef _KERNEL
96 #define	OMFSNAMELEN	16	/* length of fs type name, including null */
97 #define	OMNAMELEN	(88 - 2 * sizeof(long))	/* size of on/from name bufs */
98 
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. */
101 struct ostatfs {
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 */
123 	/*
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.
127 	 */
128 	long	f_spare[2];		/* unused spare */
129 };
130 
131 TAILQ_HEAD(vnodelst, vnode);
132 
133 /* Mount options list */
134 TAILQ_HEAD(vfsoptlist, vfsopt);
135 struct vfsopt {
136 	TAILQ_ENTRY(vfsopt) link;
137 	char	*name;
138 	void	*value;
139 	int	len;
140 	int	pos;
141 	int	seen;
142 };
143 
144 /*
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.
148  *
149  * Lock reference:
150  *	m - mountlist_mtx
151  *	i - interlock
152  *	v - vnode freelist mutex
153  *
154  * Unmarked fields are considered stable as long as a ref is held.
155  *
156  */
157 struct mount {
158 	struct mtx	mnt_mtx;		/* mount structure interlock */
159 	int		mnt_gen;		/* struct mount generation */
160 #define	mnt_startzero	mnt_list
161 	TAILQ_ENTRY(mount) mnt_list;		/* (m) mount list */
162 	struct vfsops	*mnt_op;		/* operations on fs */
163 	struct vfsconf	*mnt_vfc;		/* configuration info */
164 	struct vnode	*mnt_vnodecovered;	/* vnode we mounted on */
165 	struct vnode	*mnt_syncer;		/* syncer vnode */
166 	int		mnt_ref;		/* (i) Reference count */
167 	struct vnodelst	mnt_nvnodelist;		/* (i) list of vnodes */
168 	int		mnt_nvnodelistsize;	/* (i) # of vnodes */
169 	struct vnodelst	mnt_activevnodelist;	/* (v) list of active vnodes */
170 	int		mnt_activevnodelistsize;/* (v) # of active vnodes */
171 	int		mnt_writeopcount;	/* (i) write syscalls pending */
172 	int		mnt_kern_flag;		/* (i) kernel only flags */
173 	uint64_t	mnt_flag;		/* (i) flags shared with user */
174 	struct vfsoptlist *mnt_opt;		/* current mount options */
175 	struct vfsoptlist *mnt_optnew;		/* new options passed to fs */
176 	int		mnt_maxsymlinklen;	/* max size of short symlink */
177 	struct statfs	mnt_stat;		/* cache of filesystem stats */
178 	struct ucred	*mnt_cred;		/* credentials of mounter */
179 	void *		mnt_data;		/* private data */
180 	time_t		mnt_time;		/* last time written*/
181 	int		mnt_iosize_max;		/* max size for clusters, etc */
182 	struct netexport *mnt_export;		/* export list */
183 	struct label	*mnt_label;		/* MAC label for the fs */
184 	u_int		mnt_hashseed;		/* Random seed for vfs_hash */
185 	int		mnt_lockref;		/* (i) Lock reference count */
186 	int		mnt_secondary_writes;   /* (i) # of secondary writes */
187 	int		mnt_secondary_accwrites;/* (i) secondary wr. starts */
188 	struct thread	*mnt_susp_owner;	/* (i) thread owning suspension */
189 #define	mnt_endzero	mnt_gjprovider
190 	char		*mnt_gjprovider;	/* gjournal provider name */
191 	struct lock	mnt_explock;		/* vfs_export walkers lock */
192 	TAILQ_ENTRY(mount) mnt_upper_link;	/* (m) we in the all uppers */
193 	TAILQ_HEAD(, mount) mnt_uppers;		/* (m) upper mounts over us*/
194 };
195 
196 /*
197  * Definitions for MNT_VNODE_FOREACH_ALL.
198  */
199 struct vnode *__mnt_vnode_next_all(struct vnode **mvp, struct mount *mp);
200 struct vnode *__mnt_vnode_first_all(struct vnode **mvp, struct mount *mp);
201 void          __mnt_vnode_markerfree_all(struct vnode **mvp, struct mount *mp);
202 
203 #define MNT_VNODE_FOREACH_ALL(vp, mp, mvp)				\
204 	for (vp = __mnt_vnode_first_all(&(mvp), (mp));			\
205 		(vp) != NULL; vp = __mnt_vnode_next_all(&(mvp), (mp)))
206 
207 #define MNT_VNODE_FOREACH_ALL_ABORT(mp, mvp)				\
208 	do {								\
209 		MNT_ILOCK(mp);						\
210 		__mnt_vnode_markerfree_all(&(mvp), (mp));		\
211 		/* MNT_IUNLOCK(mp); -- done in above function */	\
212 		mtx_assert(MNT_MTX(mp), MA_NOTOWNED);			\
213 	} while (0)
214 
215 /*
216  * Definitions for MNT_VNODE_FOREACH_ACTIVE.
217  */
218 struct vnode *__mnt_vnode_next_active(struct vnode **mvp, struct mount *mp);
219 struct vnode *__mnt_vnode_first_active(struct vnode **mvp, struct mount *mp);
220 void          __mnt_vnode_markerfree_active(struct vnode **mvp, struct mount *);
221 
222 #define MNT_VNODE_FOREACH_ACTIVE(vp, mp, mvp) 				\
223 	for (vp = __mnt_vnode_first_active(&(mvp), (mp)); 		\
224 		(vp) != NULL; vp = __mnt_vnode_next_active(&(mvp), (mp)))
225 
226 #define MNT_VNODE_FOREACH_ACTIVE_ABORT(mp, mvp)				\
227 	__mnt_vnode_markerfree_active(&(mvp), (mp))
228 
229 #define	MNT_ILOCK(mp)	mtx_lock(&(mp)->mnt_mtx)
230 #define	MNT_ITRYLOCK(mp) mtx_trylock(&(mp)->mnt_mtx)
231 #define	MNT_IUNLOCK(mp)	mtx_unlock(&(mp)->mnt_mtx)
232 #define	MNT_MTX(mp)	(&(mp)->mnt_mtx)
233 #define	MNT_REF(mp)	(mp)->mnt_ref++
234 #define	MNT_REL(mp)	do {						\
235 	KASSERT((mp)->mnt_ref > 0, ("negative mnt_ref"));		\
236 	(mp)->mnt_ref--;						\
237 	if ((mp)->mnt_ref == 0)						\
238 		wakeup((mp));						\
239 } while (0)
240 
241 #endif /* _KERNEL */
242 
243 /*
244  * User specifiable flags, stored in mnt_flag.
245  */
246 #define	MNT_RDONLY	0x0000000000000001ULL /* read only filesystem */
247 #define	MNT_SYNCHRONOUS	0x0000000000000002ULL /* fs written synchronously */
248 #define	MNT_NOEXEC	0x0000000000000004ULL /* can't exec from filesystem */
249 #define	MNT_NOSUID	0x0000000000000008ULL /* don't honor setuid fs bits */
250 #define	MNT_NFS4ACLS	0x0000000000000010ULL /* enable NFS version 4 ACLs */
251 #define	MNT_UNION	0x0000000000000020ULL /* union with underlying fs */
252 #define	MNT_ASYNC	0x0000000000000040ULL /* fs written asynchronously */
253 #define	MNT_SUIDDIR	0x0000000000100000ULL /* special SUID dir handling */
254 #define	MNT_SOFTDEP	0x0000000000200000ULL /* using soft updates */
255 #define	MNT_NOSYMFOLLOW	0x0000000000400000ULL /* do not follow symlinks */
256 #define	MNT_GJOURNAL	0x0000000002000000ULL /* GEOM journal support enabled */
257 #define	MNT_MULTILABEL	0x0000000004000000ULL /* MAC support for objects */
258 #define	MNT_ACLS	0x0000000008000000ULL /* ACL support enabled */
259 #define	MNT_NOATIME	0x0000000010000000ULL /* dont update file access time */
260 #define	MNT_NOCLUSTERR	0x0000000040000000ULL /* disable cluster read */
261 #define	MNT_NOCLUSTERW	0x0000000080000000ULL /* disable cluster write */
262 #define	MNT_SUJ		0x0000000100000000ULL /* using journaled soft updates */
263 #define	MNT_AUTOMOUNTED	0x0000000200000000ULL /* mounted by automountd(8) */
264 
265 /*
266  * NFS export related mount flags.
267  */
268 #define	MNT_EXRDONLY	0x0000000000000080ULL	/* exported read only */
269 #define	MNT_EXPORTED	0x0000000000000100ULL	/* filesystem is exported */
270 #define	MNT_DEFEXPORTED	0x0000000000000200ULL	/* exported to the world */
271 #define	MNT_EXPORTANON	0x0000000000000400ULL	/* anon uid mapping for all */
272 #define	MNT_EXKERB	0x0000000000000800ULL	/* exported with Kerberos */
273 #define	MNT_EXPUBLIC	0x0000000020000000ULL	/* public export (WebNFS) */
274 
275 /*
276  * Flags set by internal operations,
277  * but visible to the user.
278  * XXX some of these are not quite right.. (I've never seen the root flag set)
279  */
280 #define	MNT_LOCAL	0x0000000000001000ULL /* filesystem is stored locally */
281 #define	MNT_QUOTA	0x0000000000002000ULL /* quotas are enabled on fs */
282 #define	MNT_ROOTFS	0x0000000000004000ULL /* identifies the root fs */
283 #define	MNT_USER	0x0000000000008000ULL /* mounted by a user */
284 #define	MNT_IGNORE	0x0000000000800000ULL /* do not show entry in df */
285 
286 /*
287  * Mask of flags that are visible to statfs().
288  * XXX I think that this could now become (~(MNT_CMDFLAGS))
289  * but the 'mount' program may need changing to handle this.
290  */
291 #define	MNT_VISFLAGMASK	(MNT_RDONLY	| MNT_SYNCHRONOUS | MNT_NOEXEC	| \
292 			MNT_NOSUID	| MNT_UNION	| MNT_SUJ	| \
293 			MNT_ASYNC	| MNT_EXRDONLY	| MNT_EXPORTED	| \
294 			MNT_DEFEXPORTED	| MNT_EXPORTANON| MNT_EXKERB	| \
295 			MNT_LOCAL	| MNT_USER	| MNT_QUOTA	| \
296 			MNT_ROOTFS	| MNT_NOATIME	| MNT_NOCLUSTERR| \
297 			MNT_NOCLUSTERW	| MNT_SUIDDIR	| MNT_SOFTDEP	| \
298 			MNT_IGNORE	| MNT_EXPUBLIC	| MNT_NOSYMFOLLOW | \
299 			MNT_GJOURNAL	| MNT_MULTILABEL | MNT_ACLS	| \
300 			MNT_NFS4ACLS	| MNT_AUTOMOUNTED)
301 
302 /* Mask of flags that can be updated. */
303 #define	MNT_UPDATEMASK (MNT_NOSUID	| MNT_NOEXEC	| \
304 			MNT_SYNCHRONOUS	| MNT_UNION	| MNT_ASYNC	| \
305 			MNT_NOATIME | \
306 			MNT_NOSYMFOLLOW	| MNT_IGNORE	| \
307 			MNT_NOCLUSTERR	| MNT_NOCLUSTERW | MNT_SUIDDIR	| \
308 			MNT_ACLS	| MNT_USER	| MNT_NFS4ACLS	| \
309 			MNT_AUTOMOUNTED)
310 
311 /*
312  * External filesystem command modifier flags.
313  * Unmount can use the MNT_FORCE flag.
314  * XXX: These are not STATES and really should be somewhere else.
315  * XXX: MNT_BYFSID collides with MNT_ACLS, but because MNT_ACLS is only used for
316  *      mount(2) and MNT_BYFSID is only used for unmount(2) it's harmless.
317  */
318 #define	MNT_UPDATE	0x0000000000010000ULL /* not real mount, just update */
319 #define	MNT_DELEXPORT	0x0000000000020000ULL /* delete export host lists */
320 #define	MNT_RELOAD	0x0000000000040000ULL /* reload filesystem data */
321 #define	MNT_FORCE	0x0000000000080000ULL /* force unmount or readonly */
322 #define	MNT_SNAPSHOT	0x0000000001000000ULL /* snapshot the filesystem */
323 #define	MNT_BYFSID	0x0000000008000000ULL /* specify filesystem by ID. */
324 #define MNT_CMDFLAGS   (MNT_UPDATE	| MNT_DELEXPORT	| MNT_RELOAD	| \
325 			MNT_FORCE	| MNT_SNAPSHOT	| MNT_BYFSID)
326 /*
327  * Internal filesystem control flags stored in mnt_kern_flag.
328  *
329  * MNTK_UNMOUNT locks the mount entry so that name lookup cannot proceed
330  * past the mount point.  This keeps the subtree stable during mounts
331  * and unmounts.
332  *
333  * MNTK_UNMOUNTF permits filesystems to detect a forced unmount while
334  * dounmount() is still waiting to lock the mountpoint. This allows
335  * the filesystem to cancel operations that might otherwise deadlock
336  * with the unmount attempt (used by NFS).
337  *
338  * MNTK_NOINSMNTQ is strict subset of MNTK_UNMOUNT. They are separated
339  * to allow for failed unmount attempt to restore the syncer vnode for
340  * the mount.
341  */
342 #define MNTK_UNMOUNTF	0x00000001	/* forced unmount in progress */
343 #define MNTK_ASYNC	0x00000002	/* filtered async flag */
344 #define MNTK_SOFTDEP	0x00000004	/* async disabled by softdep */
345 #define MNTK_NOINSMNTQ	0x00000008	/* insmntque is not allowed */
346 #define	MNTK_DRAINING	0x00000010	/* lock draining is happening */
347 #define	MNTK_REFEXPIRE	0x00000020	/* refcount expiring is happening */
348 #define MNTK_EXTENDED_SHARED	0x00000040 /* Allow shared locking for more ops */
349 #define	MNTK_SHARED_WRITES	0x00000080 /* Allow shared locking for writes */
350 #define	MNTK_NO_IOPF	0x00000100	/* Disallow page faults during reads
351 					   and writes. Filesystem shall properly
352 					   handle i/o state on EFAULT. */
353 #define	MNTK_VGONE_UPPER	0x00000200
354 #define	MNTK_VGONE_WAITER	0x00000400
355 #define	MNTK_LOOKUP_EXCL_DOTDOT	0x00000800
356 #define	MNTK_MARKER		0x00001000
357 #define	MNTK_UNMAPPED_BUFS	0x00002000
358 #define	MNTK_USES_BCACHE	0x00004000 /* FS uses the buffer cache. */
359 #define MNTK_NOASYNC	0x00800000	/* disable async */
360 #define MNTK_UNMOUNT	0x01000000	/* unmount in progress */
361 #define	MNTK_MWAIT	0x02000000	/* waiting for unmount to finish */
362 #define	MNTK_SUSPEND	0x08000000	/* request write suspension */
363 #define	MNTK_SUSPEND2	0x04000000	/* block secondary writes */
364 #define	MNTK_SUSPENDED	0x10000000	/* write operations are suspended */
365 #define	MNTK_UNUSED1	0x20000000
366 #define MNTK_LOOKUP_SHARED	0x40000000 /* FS supports shared lock lookups */
367 #define	MNTK_NOKNOTE	0x80000000	/* Don't send KNOTEs from VOP hooks */
368 
369 #ifdef _KERNEL
370 static inline int
371 MNT_SHARED_WRITES(struct mount *mp)
372 {
373 
374 	return (mp != NULL && (mp->mnt_kern_flag & MNTK_SHARED_WRITES) != 0);
375 }
376 
377 static inline int
378 MNT_EXTENDED_SHARED(struct mount *mp)
379 {
380 
381 	return (mp != NULL && (mp->mnt_kern_flag & MNTK_EXTENDED_SHARED) != 0);
382 }
383 #endif
384 
385 /*
386  * Sysctl CTL_VFS definitions.
387  *
388  * Second level identifier specifies which filesystem. Second level
389  * identifier VFS_VFSCONF returns information about all filesystems.
390  * Second level identifier VFS_GENERIC is non-terminal.
391  */
392 #define	VFS_VFSCONF		0	/* get configured filesystems */
393 #define	VFS_GENERIC		0	/* generic filesystem information */
394 /*
395  * Third level identifiers for VFS_GENERIC are given below; third
396  * level identifiers for specific filesystems are given in their
397  * mount specific header files.
398  */
399 #define VFS_MAXTYPENUM	1	/* int: highest defined filesystem type */
400 #define VFS_CONF	2	/* struct: vfsconf for filesystem given
401 				   as next argument */
402 
403 /*
404  * Flags for various system call interfaces.
405  *
406  * waitfor flags to vfs_sync() and getfsstat()
407  */
408 #define MNT_WAIT	1	/* synchronously wait for I/O to complete */
409 #define MNT_NOWAIT	2	/* start all I/O, but do not wait for it */
410 #define MNT_LAZY	3	/* push data not written by filesystem syncer */
411 #define MNT_SUSPEND	4	/* Suspend file system after sync */
412 
413 /*
414  * Generic file handle
415  */
416 struct fhandle {
417 	fsid_t	fh_fsid;	/* Filesystem id of mount point */
418 	struct	fid fh_fid;	/* Filesys specific id */
419 };
420 typedef struct fhandle	fhandle_t;
421 
422 /*
423  * Old export arguments without security flavor list
424  */
425 struct oexport_args {
426 	int	ex_flags;		/* export related flags */
427 	uid_t	ex_root;		/* mapping for root uid */
428 	struct	xucred ex_anon;		/* mapping for anonymous user */
429 	struct	sockaddr *ex_addr;	/* net address to which exported */
430 	u_char	ex_addrlen;		/* and the net address length */
431 	struct	sockaddr *ex_mask;	/* mask of valid bits in saddr */
432 	u_char	ex_masklen;		/* and the smask length */
433 	char	*ex_indexfile;		/* index file for WebNFS URLs */
434 };
435 
436 /*
437  * Export arguments for local filesystem mount calls.
438  */
439 #define	MAXSECFLAVORS	5
440 struct export_args {
441 	int	ex_flags;		/* export related flags */
442 	uid_t	ex_root;		/* mapping for root uid */
443 	struct	xucred ex_anon;		/* mapping for anonymous user */
444 	struct	sockaddr *ex_addr;	/* net address to which exported */
445 	u_char	ex_addrlen;		/* and the net address length */
446 	struct	sockaddr *ex_mask;	/* mask of valid bits in saddr */
447 	u_char	ex_masklen;		/* and the smask length */
448 	char	*ex_indexfile;		/* index file for WebNFS URLs */
449 	int	ex_numsecflavors;	/* security flavor count */
450 	int	ex_secflavors[MAXSECFLAVORS]; /* list of security flavors */
451 };
452 
453 /*
454  * Structure holding information for a publicly exported filesystem
455  * (WebNFS). Currently the specs allow just for one such filesystem.
456  */
457 struct nfs_public {
458 	int		np_valid;	/* Do we hold valid information */
459 	fhandle_t	np_handle;	/* Filehandle for pub fs (internal) */
460 	struct mount	*np_mount;	/* Mountpoint of exported fs */
461 	char		*np_index;	/* Index file */
462 };
463 
464 /*
465  * Filesystem configuration information. One of these exists for each
466  * type of filesystem supported by the kernel. These are searched at
467  * mount time to identify the requested filesystem.
468  *
469  * XXX: Never change the first two arguments!
470  */
471 struct vfsconf {
472 	u_int	vfc_version;		/* ABI version number */
473 	char	vfc_name[MFSNAMELEN];	/* filesystem type name */
474 	struct	vfsops *vfc_vfsops;	/* filesystem operations vector */
475 	int	vfc_typenum;		/* historic filesystem type number */
476 	int	vfc_refcount;		/* number mounted of this type */
477 	int	vfc_flags;		/* permanent flags */
478 	struct	vfsoptdecl *vfc_opts;	/* mount options */
479 	TAILQ_ENTRY(vfsconf) vfc_list;	/* list of vfscons */
480 };
481 
482 /* Userland version of the struct vfsconf. */
483 struct xvfsconf {
484 	struct	vfsops *vfc_vfsops;	/* filesystem operations vector */
485 	char	vfc_name[MFSNAMELEN];	/* filesystem type name */
486 	int	vfc_typenum;		/* historic filesystem type number */
487 	int	vfc_refcount;		/* number mounted of this type */
488 	int	vfc_flags;		/* permanent flags */
489 	struct	vfsconf *vfc_next;	/* next in list */
490 };
491 
492 #ifndef BURN_BRIDGES
493 struct ovfsconf {
494 	void	*vfc_vfsops;
495 	char	vfc_name[32];
496 	int	vfc_index;
497 	int	vfc_refcount;
498 	int	vfc_flags;
499 };
500 #endif
501 
502 /*
503  * NB: these flags refer to IMPLEMENTATION properties, not properties of
504  * any actual mounts; i.e., it does not make sense to change the flags.
505  */
506 #define	VFCF_STATIC	0x00010000	/* statically compiled into kernel */
507 #define	VFCF_NETWORK	0x00020000	/* may get data over the network */
508 #define	VFCF_READONLY	0x00040000	/* writes are not implemented */
509 #define	VFCF_SYNTHETIC	0x00080000	/* data does not represent real files */
510 #define	VFCF_LOOPBACK	0x00100000	/* aliases some other mounted FS */
511 #define	VFCF_UNICODE	0x00200000	/* stores file names as Unicode */
512 #define	VFCF_JAIL	0x00400000	/* can be mounted from within a jail */
513 #define	VFCF_DELEGADMIN	0x00800000	/* supports delegated administration */
514 #define	VFCF_SBDRY	0x01000000	/* defer stop requests */
515 
516 typedef uint32_t fsctlop_t;
517 
518 struct vfsidctl {
519 	int		vc_vers;	/* should be VFSIDCTL_VERS1 (below) */
520 	fsid_t		vc_fsid;	/* fsid to operate on */
521 	char		vc_fstypename[MFSNAMELEN];
522 					/* type of fs 'nfs' or '*' */
523 	fsctlop_t	vc_op;		/* operation VFS_CTL_* (below) */
524 	void		*vc_ptr;	/* pointer to data structure */
525 	size_t		vc_len;		/* sizeof said structure */
526 	u_int32_t	vc_spare[12];	/* spare (must be zero) */
527 };
528 
529 /* vfsidctl API version. */
530 #define VFS_CTL_VERS1	0x01
531 
532 /*
533  * New style VFS sysctls, do not reuse/conflict with the namespace for
534  * private sysctls.
535  * All "global" sysctl ops have the 33rd bit set:
536  * 0x...1....
537  * Private sysctl ops should have the 33rd bit unset.
538  */
539 #define VFS_CTL_QUERY	0x00010001	/* anything wrong? (vfsquery) */
540 #define VFS_CTL_TIMEO	0x00010002	/* set timeout for vfs notification */
541 #define VFS_CTL_NOLOCKS	0x00010003	/* disable file locking */
542 
543 struct vfsquery {
544 	u_int32_t	vq_flags;
545 	u_int32_t	vq_spare[31];
546 };
547 
548 /* vfsquery flags */
549 #define VQ_NOTRESP	0x0001	/* server down */
550 #define VQ_NEEDAUTH	0x0002	/* server bad auth */
551 #define VQ_LOWDISK	0x0004	/* we're low on space */
552 #define VQ_MOUNT	0x0008	/* new filesystem arrived */
553 #define VQ_UNMOUNT	0x0010	/* filesystem has left */
554 #define VQ_DEAD		0x0020	/* filesystem is dead, needs force unmount */
555 #define VQ_ASSIST	0x0040	/* filesystem needs assistance from external
556 				   program */
557 #define VQ_NOTRESPLOCK	0x0080	/* server lockd down */
558 #define VQ_FLAG0100	0x0100	/* placeholder */
559 #define VQ_FLAG0200	0x0200	/* placeholder */
560 #define VQ_FLAG0400	0x0400	/* placeholder */
561 #define VQ_FLAG0800	0x0800	/* placeholder */
562 #define VQ_FLAG1000	0x1000	/* placeholder */
563 #define VQ_FLAG2000	0x2000	/* placeholder */
564 #define VQ_FLAG4000	0x4000	/* placeholder */
565 #define VQ_FLAG8000	0x8000	/* placeholder */
566 
567 #ifdef _KERNEL
568 /* Point a sysctl request at a vfsidctl's data. */
569 #define VCTLTOREQ(vc, req)						\
570 	do {								\
571 		(req)->newptr = (vc)->vc_ptr;				\
572 		(req)->newlen = (vc)->vc_len;				\
573 		(req)->newidx = 0;					\
574 	} while (0)
575 #endif
576 
577 struct iovec;
578 struct uio;
579 
580 #ifdef _KERNEL
581 
582 /*
583  * vfs_busy specific flags and mask.
584  */
585 #define	MBF_NOWAIT	0x01
586 #define	MBF_MNTLSTLOCK	0x02
587 #define	MBF_MASK	(MBF_NOWAIT | MBF_MNTLSTLOCK)
588 
589 #ifdef MALLOC_DECLARE
590 MALLOC_DECLARE(M_MOUNT);
591 #endif
592 extern int maxvfsconf;		/* highest defined filesystem type */
593 
594 TAILQ_HEAD(vfsconfhead, vfsconf);
595 extern struct vfsconfhead vfsconf;
596 
597 /*
598  * Operations supported on mounted filesystem.
599  */
600 struct mount_args;
601 struct nameidata;
602 struct sysctl_req;
603 struct mntarg;
604 
605 typedef int vfs_cmount_t(struct mntarg *ma, void *data, uint64_t flags);
606 typedef int vfs_unmount_t(struct mount *mp, int mntflags);
607 typedef int vfs_root_t(struct mount *mp, int flags, struct vnode **vpp);
608 typedef	int vfs_quotactl_t(struct mount *mp, int cmds, uid_t uid, void *arg);
609 typedef	int vfs_statfs_t(struct mount *mp, struct statfs *sbp);
610 typedef	int vfs_sync_t(struct mount *mp, int waitfor);
611 typedef	int vfs_vget_t(struct mount *mp, ino_t ino, int flags,
612 		    struct vnode **vpp);
613 typedef	int vfs_fhtovp_t(struct mount *mp, struct fid *fhp,
614 		    int flags, struct vnode **vpp);
615 typedef	int vfs_checkexp_t(struct mount *mp, struct sockaddr *nam,
616 		    int *extflagsp, struct ucred **credanonp,
617 		    int *numsecflavors, int **secflavors);
618 typedef	int vfs_init_t(struct vfsconf *);
619 typedef	int vfs_uninit_t(struct vfsconf *);
620 typedef	int vfs_extattrctl_t(struct mount *mp, int cmd,
621 		    struct vnode *filename_vp, int attrnamespace,
622 		    const char *attrname);
623 typedef	int vfs_mount_t(struct mount *mp);
624 typedef int vfs_sysctl_t(struct mount *mp, fsctlop_t op,
625 		    struct sysctl_req *req);
626 typedef void vfs_susp_clean_t(struct mount *mp);
627 typedef void vfs_notify_lowervp_t(struct mount *mp, struct vnode *lowervp);
628 typedef void vfs_purge_t(struct mount *mp);
629 
630 struct vfsops {
631 	vfs_mount_t		*vfs_mount;
632 	vfs_cmount_t		*vfs_cmount;
633 	vfs_unmount_t		*vfs_unmount;
634 	vfs_root_t		*vfs_root;
635 	vfs_quotactl_t		*vfs_quotactl;
636 	vfs_statfs_t		*vfs_statfs;
637 	vfs_sync_t		*vfs_sync;
638 	vfs_vget_t		*vfs_vget;
639 	vfs_fhtovp_t		*vfs_fhtovp;
640 	vfs_checkexp_t		*vfs_checkexp;
641 	vfs_init_t		*vfs_init;
642 	vfs_uninit_t		*vfs_uninit;
643 	vfs_extattrctl_t	*vfs_extattrctl;
644 	vfs_sysctl_t		*vfs_sysctl;
645 	vfs_susp_clean_t	*vfs_susp_clean;
646 	vfs_notify_lowervp_t	*vfs_reclaim_lowervp;
647 	vfs_notify_lowervp_t	*vfs_unlink_lowervp;
648 	vfs_purge_t		*vfs_purge;
649 	vfs_mount_t		*vfs_spare[6];	/* spares for ABI compat */
650 };
651 
652 vfs_statfs_t	__vfs_statfs;
653 
654 #define	VFS_PROLOGUE(MP)	do {					\
655 	struct mount *mp__;						\
656 	int _enable_stops;						\
657 									\
658 	mp__ = (MP);							\
659 	_enable_stops = (mp__ != NULL &&				\
660 	    (mp__->mnt_vfc->vfc_flags & VFCF_SBDRY) && sigdeferstop())
661 
662 #define	VFS_EPILOGUE(MP)						\
663 	if (_enable_stops)						\
664 		sigallowstop();						\
665 } while (0)
666 
667 #define	VFS_MOUNT(MP) ({						\
668 	int _rc;							\
669 									\
670 	VFS_PROLOGUE(MP);						\
671 	_rc = (*(MP)->mnt_op->vfs_mount)(MP);				\
672 	VFS_EPILOGUE(MP);						\
673 	_rc; })
674 
675 #define	VFS_UNMOUNT(MP, FORCE) ({					\
676 	int _rc;							\
677 									\
678 	VFS_PROLOGUE(MP);						\
679 	_rc = (*(MP)->mnt_op->vfs_unmount)(MP, FORCE);			\
680 	VFS_EPILOGUE(MP);						\
681 	_rc; })
682 
683 #define	VFS_ROOT(MP, FLAGS, VPP) ({					\
684 	int _rc;							\
685 									\
686 	VFS_PROLOGUE(MP);						\
687 	_rc = (*(MP)->mnt_op->vfs_root)(MP, FLAGS, VPP);		\
688 	VFS_EPILOGUE(MP);						\
689 	_rc; })
690 
691 #define	VFS_QUOTACTL(MP, C, U, A) ({					\
692 	int _rc;							\
693 									\
694 	VFS_PROLOGUE(MP);						\
695 	_rc = (*(MP)->mnt_op->vfs_quotactl)(MP, C, U, A);		\
696 	VFS_EPILOGUE(MP);						\
697 	_rc; })
698 
699 #define	VFS_STATFS(MP, SBP) ({						\
700 	int _rc;							\
701 									\
702 	VFS_PROLOGUE(MP);						\
703 	_rc = __vfs_statfs((MP), (SBP));				\
704 	VFS_EPILOGUE(MP);						\
705 	_rc; })
706 
707 #define	VFS_SYNC(MP, WAIT) ({						\
708 	int _rc;							\
709 									\
710 	VFS_PROLOGUE(MP);						\
711 	_rc = (*(MP)->mnt_op->vfs_sync)(MP, WAIT);			\
712 	VFS_EPILOGUE(MP);						\
713 	_rc; })
714 
715 #define	VFS_VGET(MP, INO, FLAGS, VPP) ({				\
716 	int _rc;							\
717 									\
718 	VFS_PROLOGUE(MP);						\
719 	_rc = (*(MP)->mnt_op->vfs_vget)(MP, INO, FLAGS, VPP);		\
720 	VFS_EPILOGUE(MP);						\
721 	_rc; })
722 
723 #define	VFS_FHTOVP(MP, FIDP, FLAGS, VPP) ({				\
724 	int _rc;							\
725 									\
726 	VFS_PROLOGUE(MP);						\
727 	_rc = (*(MP)->mnt_op->vfs_fhtovp)(MP, FIDP, FLAGS, VPP);	\
728 	VFS_EPILOGUE(MP);						\
729 	_rc; })
730 
731 #define	VFS_CHECKEXP(MP, NAM, EXFLG, CRED, NUMSEC, SEC) ({		\
732 	int _rc;							\
733 									\
734 	VFS_PROLOGUE(MP);						\
735 	_rc = (*(MP)->mnt_op->vfs_checkexp)(MP, NAM, EXFLG, CRED, NUMSEC,\
736 	    SEC);							\
737 	VFS_EPILOGUE(MP);						\
738 	_rc; })
739 
740 #define	VFS_EXTATTRCTL(MP, C, FN, NS, N) ({				\
741 	int _rc;							\
742 									\
743 	VFS_PROLOGUE(MP);						\
744 	_rc = (*(MP)->mnt_op->vfs_extattrctl)(MP, C, FN, NS, N);	\
745 	VFS_EPILOGUE(MP);						\
746 	_rc; })
747 
748 #define	VFS_SYSCTL(MP, OP, REQ) ({					\
749 	int _rc;							\
750 									\
751 	VFS_PROLOGUE(MP);						\
752 	_rc = (*(MP)->mnt_op->vfs_sysctl)(MP, OP, REQ);			\
753 	VFS_EPILOGUE(MP);						\
754 	_rc; })
755 
756 #define	VFS_SUSP_CLEAN(MP) do {						\
757 	if (*(MP)->mnt_op->vfs_susp_clean != NULL) {			\
758 		VFS_PROLOGUE(MP);					\
759 		(*(MP)->mnt_op->vfs_susp_clean)(MP);			\
760 		VFS_EPILOGUE(MP);					\
761 	}								\
762 } while (0)
763 
764 #define	VFS_RECLAIM_LOWERVP(MP, VP) do {				\
765 	if (*(MP)->mnt_op->vfs_reclaim_lowervp != NULL) {		\
766 		VFS_PROLOGUE(MP);					\
767 		(*(MP)->mnt_op->vfs_reclaim_lowervp)((MP), (VP));	\
768 		VFS_EPILOGUE(MP);					\
769 	}								\
770 } while (0)
771 
772 #define	VFS_UNLINK_LOWERVP(MP, VP) do {					\
773 	if (*(MP)->mnt_op->vfs_unlink_lowervp != NULL) {		\
774 		VFS_PROLOGUE(MP);					\
775 		(*(MP)->mnt_op->vfs_unlink_lowervp)((MP), (VP));	\
776 		VFS_EPILOGUE(MP);					\
777 	}								\
778 } while (0)
779 
780 #define	VFS_PURGE(MP) do {						\
781 	if (*(MP)->mnt_op->vfs_purge != NULL) {				\
782 		VFS_PROLOGUE(MP);					\
783 		(*(MP)->mnt_op->vfs_purge)(MP);				\
784 		VFS_EPILOGUE(MP);					\
785 	}								\
786 } while (0)
787 
788 #define VFS_KNOTE_LOCKED(vp, hint) do					\
789 {									\
790 	if (((vp)->v_vflag & VV_NOKNOTE) == 0)				\
791 		VN_KNOTE((vp), (hint), KNF_LISTLOCKED);			\
792 } while (0)
793 
794 #define VFS_KNOTE_UNLOCKED(vp, hint) do					\
795 {									\
796 	if (((vp)->v_vflag & VV_NOKNOTE) == 0)				\
797 		VN_KNOTE((vp), (hint), 0);				\
798 } while (0)
799 
800 #define	VFS_NOTIFY_UPPER_RECLAIM	1
801 #define	VFS_NOTIFY_UPPER_UNLINK		2
802 
803 #include <sys/module.h>
804 
805 /*
806  * Version numbers.
807  */
808 #define VFS_VERSION_00	0x19660120
809 #define VFS_VERSION_01	0x20121030
810 #define VFS_VERSION	VFS_VERSION_01
811 
812 #define VFS_SET(vfsops, fsname, flags) \
813 	static struct vfsconf fsname ## _vfsconf = {		\
814 		.vfc_version = VFS_VERSION,			\
815 		.vfc_name = #fsname,				\
816 		.vfc_vfsops = &vfsops,				\
817 		.vfc_typenum = -1,				\
818 		.vfc_flags = flags,				\
819 	};							\
820 	static moduledata_t fsname ## _mod = {			\
821 		#fsname,					\
822 		vfs_modevent,					\
823 		& fsname ## _vfsconf				\
824 	};							\
825 	DECLARE_MODULE(fsname, fsname ## _mod, SI_SUB_VFS, SI_ORDER_MIDDLE)
826 
827 /*
828  * exported vnode operations
829  */
830 
831 int	dounmount(struct mount *, int, struct thread *);
832 
833 int	kernel_mount(struct mntarg *ma, uint64_t flags);
834 int	kernel_vmount(int flags, ...);
835 struct mntarg *mount_arg(struct mntarg *ma, const char *name, const void *val, int len);
836 struct mntarg *mount_argb(struct mntarg *ma, int flag, const char *name);
837 struct mntarg *mount_argf(struct mntarg *ma, const char *name, const char *fmt, ...);
838 struct mntarg *mount_argsu(struct mntarg *ma, const char *name, const void *val, int len);
839 void	statfs_scale_blocks(struct statfs *sf, long max_size);
840 struct vfsconf *vfs_byname(const char *);
841 struct vfsconf *vfs_byname_kld(const char *, struct thread *td, int *);
842 void	vfs_mount_destroy(struct mount *);
843 void	vfs_event_signal(fsid_t *, u_int32_t, intptr_t);
844 void	vfs_freeopts(struct vfsoptlist *opts);
845 void	vfs_deleteopt(struct vfsoptlist *opts, const char *name);
846 int	vfs_buildopts(struct uio *auio, struct vfsoptlist **options);
847 int	vfs_flagopt(struct vfsoptlist *opts, const char *name, uint64_t *w,
848 	    uint64_t val);
849 int	vfs_getopt(struct vfsoptlist *, const char *, void **, int *);
850 int	vfs_getopt_pos(struct vfsoptlist *opts, const char *name);
851 int	vfs_getopt_size(struct vfsoptlist *opts, const char *name,
852 	    off_t *value);
853 char	*vfs_getopts(struct vfsoptlist *, const char *, int *error);
854 int	vfs_copyopt(struct vfsoptlist *, const char *, void *, int);
855 int	vfs_filteropt(struct vfsoptlist *, const char **legal);
856 void	vfs_opterror(struct vfsoptlist *opts, const char *fmt, ...);
857 int	vfs_scanopt(struct vfsoptlist *opts, const char *name, const char *fmt, ...);
858 int	vfs_setopt(struct vfsoptlist *opts, const char *name, void *value,
859 	    int len);
860 int	vfs_setopt_part(struct vfsoptlist *opts, const char *name, void *value,
861 	    int len);
862 int	vfs_setopts(struct vfsoptlist *opts, const char *name,
863 	    const char *value);
864 int	vfs_setpublicfs			    /* set publicly exported fs */
865 	    (struct mount *, struct netexport *, struct export_args *);
866 void	vfs_msync(struct mount *, int);
867 int	vfs_busy(struct mount *, int);
868 int	vfs_export			 /* process mount export info */
869 	    (struct mount *, struct export_args *);
870 void	vfs_allocate_syncvnode(struct mount *);
871 void	vfs_deallocate_syncvnode(struct mount *);
872 int	vfs_donmount(struct thread *td, uint64_t fsflags,
873 	    struct uio *fsoptions);
874 void	vfs_getnewfsid(struct mount *);
875 struct cdev *vfs_getrootfsid(struct mount *);
876 struct	mount *vfs_getvfs(fsid_t *);      /* return vfs given fsid */
877 struct	mount *vfs_busyfs(fsid_t *);
878 int	vfs_modevent(module_t, int, void *);
879 void	vfs_mount_error(struct mount *, const char *, ...);
880 void	vfs_mountroot(void);			/* mount our root filesystem */
881 void	vfs_mountedfrom(struct mount *, const char *from);
882 void	vfs_notify_upper(struct vnode *, int);
883 void	vfs_oexport_conv(const struct oexport_args *oexp,
884 	    struct export_args *exp);
885 void	vfs_ref(struct mount *);
886 void	vfs_rel(struct mount *);
887 struct mount *vfs_mount_alloc(struct vnode *, struct vfsconf *, const char *,
888 	    struct ucred *);
889 int	vfs_suser(struct mount *, struct thread *);
890 void	vfs_unbusy(struct mount *);
891 void	vfs_unmountall(void);
892 extern	TAILQ_HEAD(mntlist, mount) mountlist;	/* mounted filesystem list */
893 extern	struct mtx mountlist_mtx;
894 extern	struct nfs_public nfs_pub;
895 extern	struct sx vfsconf_sx;
896 #define	vfsconf_lock()		sx_xlock(&vfsconf_sx)
897 #define	vfsconf_unlock()	sx_xunlock(&vfsconf_sx)
898 #define	vfsconf_slock()		sx_slock(&vfsconf_sx)
899 #define	vfsconf_sunlock()	sx_sunlock(&vfsconf_sx)
900 
901 /*
902  * Declarations for these vfs default operations are located in
903  * kern/vfs_default.c.  They will be automatically used to replace
904  * null entries in VFS ops tables when registering a new filesystem
905  * type in the global table.
906  */
907 vfs_root_t		vfs_stdroot;
908 vfs_quotactl_t		vfs_stdquotactl;
909 vfs_statfs_t		vfs_stdstatfs;
910 vfs_sync_t		vfs_stdsync;
911 vfs_sync_t		vfs_stdnosync;
912 vfs_vget_t		vfs_stdvget;
913 vfs_fhtovp_t		vfs_stdfhtovp;
914 vfs_checkexp_t		vfs_stdcheckexp;
915 vfs_init_t		vfs_stdinit;
916 vfs_uninit_t		vfs_stduninit;
917 vfs_extattrctl_t	vfs_stdextattrctl;
918 vfs_sysctl_t		vfs_stdsysctl;
919 
920 void	syncer_suspend(void);
921 void	syncer_resume(void);
922 
923 #else /* !_KERNEL */
924 
925 #include <sys/cdefs.h>
926 
927 struct stat;
928 
929 __BEGIN_DECLS
930 int	fhopen(const struct fhandle *, int);
931 int	fhstat(const struct fhandle *, struct stat *);
932 int	fhstatfs(const struct fhandle *, struct statfs *);
933 int	fstatfs(int, struct statfs *);
934 int	getfh(const char *, fhandle_t *);
935 int	getfsstat(struct statfs *, long, int);
936 int	getmntinfo(struct statfs **, int);
937 int	lgetfh(const char *, fhandle_t *);
938 int	mount(const char *, const char *, int, void *);
939 int	nmount(struct iovec *, unsigned int, int);
940 int	statfs(const char *, struct statfs *);
941 int	unmount(const char *, int);
942 
943 /* C library stuff */
944 int	getvfsbyname(const char *, struct xvfsconf *);
945 __END_DECLS
946 
947 #endif /* _KERNEL */
948 
949 #endif /* !_SYS_MOUNT_H_ */
950