mount.h revision 282270
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: stable/10/sys/sys/mount.h 282270 2015-04-30 12:39:24Z rmacklem $
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
50typedef 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
58struct 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 */
70struct 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. */
101struct 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
131TAILQ_HEAD(vnodelst, vnode);
132
133/* Mount options list */
134TAILQ_HEAD(vfsoptlist, vfsopt);
135struct 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 */
157struct 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 */
199struct vnode *__mnt_vnode_next_all(struct vnode **mvp, struct mount *mp);
200struct vnode *__mnt_vnode_first_all(struct vnode **mvp, struct mount *mp);
201void          __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 */
218struct vnode *__mnt_vnode_next_active(struct vnode **mvp, struct mount *mp);
219struct vnode *__mnt_vnode_first_active(struct vnode **mvp, struct mount *mp);
220void          __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_UNUSED25	0x20000000	/*  --available-- */
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
370static inline int
371MNT_SHARED_WRITES(struct mount *mp)
372{
373
374	return (mp != NULL && (mp->mnt_kern_flag & MNTK_SHARED_WRITES) != 0);
375}
376
377static inline int
378MNT_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 */
416struct fhandle {
417	fsid_t	fh_fsid;	/* Filesystem id of mount point */
418	struct	fid fh_fid;	/* Filesys specific id */
419};
420typedef struct fhandle	fhandle_t;
421
422/*
423 * Old export arguments without security flavor list
424 */
425struct 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
440struct 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 */
457struct 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 */
471struct 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. */
483struct 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
493struct 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
516typedef uint32_t fsctlop_t;
517
518struct 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
543struct 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
577struct iovec;
578struct 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
590MALLOC_DECLARE(M_MOUNT);
591#endif
592extern int maxvfsconf;		/* highest defined filesystem type */
593extern int nfs_mount_type;	/* vfc_typenum for nfs, or -1 */
594
595TAILQ_HEAD(vfsconfhead, vfsconf);
596extern struct vfsconfhead vfsconf;
597
598/*
599 * Operations supported on mounted filesystem.
600 */
601struct mount_args;
602struct nameidata;
603struct sysctl_req;
604struct mntarg;
605
606typedef int vfs_cmount_t(struct mntarg *ma, void *data, uint64_t flags);
607typedef int vfs_unmount_t(struct mount *mp, int mntflags);
608typedef int vfs_root_t(struct mount *mp, int flags, struct vnode **vpp);
609typedef	int vfs_quotactl_t(struct mount *mp, int cmds, uid_t uid, void *arg);
610typedef	int vfs_statfs_t(struct mount *mp, struct statfs *sbp);
611typedef	int vfs_sync_t(struct mount *mp, int waitfor);
612typedef	int vfs_vget_t(struct mount *mp, ino_t ino, int flags,
613		    struct vnode **vpp);
614typedef	int vfs_fhtovp_t(struct mount *mp, struct fid *fhp,
615		    int flags, struct vnode **vpp);
616typedef	int vfs_checkexp_t(struct mount *mp, struct sockaddr *nam,
617		    int *extflagsp, struct ucred **credanonp,
618		    int *numsecflavors, int **secflavors);
619typedef	int vfs_init_t(struct vfsconf *);
620typedef	int vfs_uninit_t(struct vfsconf *);
621typedef	int vfs_extattrctl_t(struct mount *mp, int cmd,
622		    struct vnode *filename_vp, int attrnamespace,
623		    const char *attrname);
624typedef	int vfs_mount_t(struct mount *mp);
625typedef int vfs_sysctl_t(struct mount *mp, fsctlop_t op,
626		    struct sysctl_req *req);
627typedef void vfs_susp_clean_t(struct mount *mp);
628typedef void vfs_notify_lowervp_t(struct mount *mp, struct vnode *lowervp);
629typedef void vfs_purge_t(struct mount *mp);
630
631struct vfsops {
632	vfs_mount_t		*vfs_mount;
633	vfs_cmount_t		*vfs_cmount;
634	vfs_unmount_t		*vfs_unmount;
635	vfs_root_t		*vfs_root;
636	vfs_quotactl_t		*vfs_quotactl;
637	vfs_statfs_t		*vfs_statfs;
638	vfs_sync_t		*vfs_sync;
639	vfs_vget_t		*vfs_vget;
640	vfs_fhtovp_t		*vfs_fhtovp;
641	vfs_checkexp_t		*vfs_checkexp;
642	vfs_init_t		*vfs_init;
643	vfs_uninit_t		*vfs_uninit;
644	vfs_extattrctl_t	*vfs_extattrctl;
645	vfs_sysctl_t		*vfs_sysctl;
646	vfs_susp_clean_t	*vfs_susp_clean;
647	vfs_notify_lowervp_t	*vfs_reclaim_lowervp;
648	vfs_notify_lowervp_t	*vfs_unlink_lowervp;
649	vfs_purge_t		*vfs_purge;
650	vfs_mount_t		*vfs_spare[6];	/* spares for ABI compat */
651};
652
653vfs_statfs_t	__vfs_statfs;
654
655#define	VFS_PROLOGUE(MP)	do {					\
656	struct mount *mp__;						\
657	int _enable_stops;						\
658									\
659	mp__ = (MP);							\
660	_enable_stops = (mp__ != NULL &&				\
661	    (mp__->mnt_vfc->vfc_flags & VFCF_SBDRY) && sigdeferstop())
662
663#define	VFS_EPILOGUE(MP)						\
664	if (_enable_stops)						\
665		sigallowstop();						\
666} while (0)
667
668#define	VFS_MOUNT(MP) ({						\
669	int _rc;							\
670									\
671	VFS_PROLOGUE(MP);						\
672	_rc = (*(MP)->mnt_op->vfs_mount)(MP);				\
673	VFS_EPILOGUE(MP);						\
674	_rc; })
675
676#define	VFS_UNMOUNT(MP, FORCE) ({					\
677	int _rc;							\
678									\
679	VFS_PROLOGUE(MP);						\
680	_rc = (*(MP)->mnt_op->vfs_unmount)(MP, FORCE);			\
681	VFS_EPILOGUE(MP);						\
682	_rc; })
683
684#define	VFS_ROOT(MP, FLAGS, VPP) ({					\
685	int _rc;							\
686									\
687	VFS_PROLOGUE(MP);						\
688	_rc = (*(MP)->mnt_op->vfs_root)(MP, FLAGS, VPP);		\
689	VFS_EPILOGUE(MP);						\
690	_rc; })
691
692#define	VFS_QUOTACTL(MP, C, U, A) ({					\
693	int _rc;							\
694									\
695	VFS_PROLOGUE(MP);						\
696	_rc = (*(MP)->mnt_op->vfs_quotactl)(MP, C, U, A);		\
697	VFS_EPILOGUE(MP);						\
698	_rc; })
699
700#define	VFS_STATFS(MP, SBP) ({						\
701	int _rc;							\
702									\
703	VFS_PROLOGUE(MP);						\
704	_rc = __vfs_statfs((MP), (SBP));				\
705	VFS_EPILOGUE(MP);						\
706	_rc; })
707
708#define	VFS_SYNC(MP, WAIT) ({						\
709	int _rc;							\
710									\
711	VFS_PROLOGUE(MP);						\
712	_rc = (*(MP)->mnt_op->vfs_sync)(MP, WAIT);			\
713	VFS_EPILOGUE(MP);						\
714	_rc; })
715
716#define	VFS_VGET(MP, INO, FLAGS, VPP) ({				\
717	int _rc;							\
718									\
719	VFS_PROLOGUE(MP);						\
720	_rc = (*(MP)->mnt_op->vfs_vget)(MP, INO, FLAGS, VPP);		\
721	VFS_EPILOGUE(MP);						\
722	_rc; })
723
724#define	VFS_FHTOVP(MP, FIDP, FLAGS, VPP) ({				\
725	int _rc;							\
726									\
727	VFS_PROLOGUE(MP);						\
728	_rc = (*(MP)->mnt_op->vfs_fhtovp)(MP, FIDP, FLAGS, VPP);	\
729	VFS_EPILOGUE(MP);						\
730	_rc; })
731
732#define	VFS_CHECKEXP(MP, NAM, EXFLG, CRED, NUMSEC, SEC) ({		\
733	int _rc;							\
734									\
735	VFS_PROLOGUE(MP);						\
736	_rc = (*(MP)->mnt_op->vfs_checkexp)(MP, NAM, EXFLG, CRED, NUMSEC,\
737	    SEC);							\
738	VFS_EPILOGUE(MP);						\
739	_rc; })
740
741#define	VFS_EXTATTRCTL(MP, C, FN, NS, N) ({				\
742	int _rc;							\
743									\
744	VFS_PROLOGUE(MP);						\
745	_rc = (*(MP)->mnt_op->vfs_extattrctl)(MP, C, FN, NS, N);	\
746	VFS_EPILOGUE(MP);						\
747	_rc; })
748
749#define	VFS_SYSCTL(MP, OP, REQ) ({					\
750	int _rc;							\
751									\
752	VFS_PROLOGUE(MP);						\
753	_rc = (*(MP)->mnt_op->vfs_sysctl)(MP, OP, REQ);			\
754	VFS_EPILOGUE(MP);						\
755	_rc; })
756
757#define	VFS_SUSP_CLEAN(MP) do {						\
758	if (*(MP)->mnt_op->vfs_susp_clean != NULL) {			\
759		VFS_PROLOGUE(MP);					\
760		(*(MP)->mnt_op->vfs_susp_clean)(MP);			\
761		VFS_EPILOGUE(MP);					\
762	}								\
763} while (0)
764
765#define	VFS_RECLAIM_LOWERVP(MP, VP) do {				\
766	if (*(MP)->mnt_op->vfs_reclaim_lowervp != NULL) {		\
767		VFS_PROLOGUE(MP);					\
768		(*(MP)->mnt_op->vfs_reclaim_lowervp)((MP), (VP));	\
769		VFS_EPILOGUE(MP);					\
770	}								\
771} while (0)
772
773#define	VFS_UNLINK_LOWERVP(MP, VP) do {					\
774	if (*(MP)->mnt_op->vfs_unlink_lowervp != NULL) {		\
775		VFS_PROLOGUE(MP);					\
776		(*(MP)->mnt_op->vfs_unlink_lowervp)((MP), (VP));	\
777		VFS_EPILOGUE(MP);					\
778	}								\
779} while (0)
780
781#define	VFS_PURGE(MP) do {						\
782	if (*(MP)->mnt_op->vfs_purge != NULL) {				\
783		VFS_PROLOGUE(MP);					\
784		(*(MP)->mnt_op->vfs_purge)(MP);				\
785		VFS_EPILOGUE(MP);					\
786	}								\
787} while (0)
788
789#define VFS_KNOTE_LOCKED(vp, hint) do					\
790{									\
791	if (((vp)->v_vflag & VV_NOKNOTE) == 0)				\
792		VN_KNOTE((vp), (hint), KNF_LISTLOCKED);			\
793} while (0)
794
795#define VFS_KNOTE_UNLOCKED(vp, hint) do					\
796{									\
797	if (((vp)->v_vflag & VV_NOKNOTE) == 0)				\
798		VN_KNOTE((vp), (hint), 0);				\
799} while (0)
800
801#define	VFS_NOTIFY_UPPER_RECLAIM	1
802#define	VFS_NOTIFY_UPPER_UNLINK		2
803
804#include <sys/module.h>
805
806/*
807 * Version numbers.
808 */
809#define VFS_VERSION_00	0x19660120
810#define VFS_VERSION_01	0x20121030
811#define VFS_VERSION	VFS_VERSION_01
812
813#define VFS_SET(vfsops, fsname, flags) \
814	static struct vfsconf fsname ## _vfsconf = {		\
815		.vfc_version = VFS_VERSION,			\
816		.vfc_name = #fsname,				\
817		.vfc_vfsops = &vfsops,				\
818		.vfc_typenum = -1,				\
819		.vfc_flags = flags,				\
820	};							\
821	static moduledata_t fsname ## _mod = {			\
822		#fsname,					\
823		vfs_modevent,					\
824		& fsname ## _vfsconf				\
825	};							\
826	DECLARE_MODULE(fsname, fsname ## _mod, SI_SUB_VFS, SI_ORDER_MIDDLE)
827
828extern	char *mountrootfsname;
829
830/*
831 * exported vnode operations
832 */
833
834int	dounmount(struct mount *, int, struct thread *);
835
836int	kernel_mount(struct mntarg *ma, uint64_t flags);
837int	kernel_vmount(int flags, ...);
838struct mntarg *mount_arg(struct mntarg *ma, const char *name, const void *val, int len);
839struct mntarg *mount_argb(struct mntarg *ma, int flag, const char *name);
840struct mntarg *mount_argf(struct mntarg *ma, const char *name, const char *fmt, ...);
841struct mntarg *mount_argsu(struct mntarg *ma, const char *name, const void *val, int len);
842void	statfs_scale_blocks(struct statfs *sf, long max_size);
843struct vfsconf *vfs_byname(const char *);
844struct vfsconf *vfs_byname_kld(const char *, struct thread *td, int *);
845void	vfs_mount_destroy(struct mount *);
846void	vfs_event_signal(fsid_t *, u_int32_t, intptr_t);
847void	vfs_freeopts(struct vfsoptlist *opts);
848void	vfs_deleteopt(struct vfsoptlist *opts, const char *name);
849int	vfs_buildopts(struct uio *auio, struct vfsoptlist **options);
850int	vfs_flagopt(struct vfsoptlist *opts, const char *name, uint64_t *w,
851	    uint64_t val);
852int	vfs_getopt(struct vfsoptlist *, const char *, void **, int *);
853int	vfs_getopt_pos(struct vfsoptlist *opts, const char *name);
854int	vfs_getopt_size(struct vfsoptlist *opts, const char *name,
855	    off_t *value);
856char	*vfs_getopts(struct vfsoptlist *, const char *, int *error);
857int	vfs_copyopt(struct vfsoptlist *, const char *, void *, int);
858int	vfs_filteropt(struct vfsoptlist *, const char **legal);
859void	vfs_opterror(struct vfsoptlist *opts, const char *fmt, ...);
860int	vfs_scanopt(struct vfsoptlist *opts, const char *name, const char *fmt, ...);
861int	vfs_setopt(struct vfsoptlist *opts, const char *name, void *value,
862	    int len);
863int	vfs_setopt_part(struct vfsoptlist *opts, const char *name, void *value,
864	    int len);
865int	vfs_setopts(struct vfsoptlist *opts, const char *name,
866	    const char *value);
867int	vfs_setpublicfs			    /* set publicly exported fs */
868	    (struct mount *, struct netexport *, struct export_args *);
869void	vfs_msync(struct mount *, int);
870int	vfs_busy(struct mount *, int);
871int	vfs_export			 /* process mount export info */
872	    (struct mount *, struct export_args *);
873void	vfs_allocate_syncvnode(struct mount *);
874void	vfs_deallocate_syncvnode(struct mount *);
875int	vfs_donmount(struct thread *td, uint64_t fsflags,
876	    struct uio *fsoptions);
877void	vfs_getnewfsid(struct mount *);
878struct cdev *vfs_getrootfsid(struct mount *);
879struct	mount *vfs_getvfs(fsid_t *);      /* return vfs given fsid */
880struct	mount *vfs_busyfs(fsid_t *);
881int	vfs_modevent(module_t, int, void *);
882void	vfs_mount_error(struct mount *, const char *, ...);
883void	vfs_mountroot(void);			/* mount our root filesystem */
884void	vfs_mountedfrom(struct mount *, const char *from);
885void	vfs_notify_upper(struct vnode *, int);
886void	vfs_oexport_conv(const struct oexport_args *oexp,
887	    struct export_args *exp);
888void	vfs_ref(struct mount *);
889void	vfs_rel(struct mount *);
890struct mount *vfs_mount_alloc(struct vnode *, struct vfsconf *, const char *,
891	    struct ucred *);
892int	vfs_suser(struct mount *, struct thread *);
893void	vfs_unbusy(struct mount *);
894void	vfs_unmountall(void);
895extern	TAILQ_HEAD(mntlist, mount) mountlist;	/* mounted filesystem list */
896extern	struct mtx mountlist_mtx;
897extern	struct nfs_public nfs_pub;
898extern	struct sx vfsconf_sx;
899#define	vfsconf_lock()		sx_xlock(&vfsconf_sx)
900#define	vfsconf_unlock()	sx_xunlock(&vfsconf_sx)
901#define	vfsconf_slock()		sx_slock(&vfsconf_sx)
902#define	vfsconf_sunlock()	sx_sunlock(&vfsconf_sx)
903
904/*
905 * Declarations for these vfs default operations are located in
906 * kern/vfs_default.c.  They will be automatically used to replace
907 * null entries in VFS ops tables when registering a new filesystem
908 * type in the global table.
909 */
910vfs_root_t		vfs_stdroot;
911vfs_quotactl_t		vfs_stdquotactl;
912vfs_statfs_t		vfs_stdstatfs;
913vfs_sync_t		vfs_stdsync;
914vfs_sync_t		vfs_stdnosync;
915vfs_vget_t		vfs_stdvget;
916vfs_fhtovp_t		vfs_stdfhtovp;
917vfs_checkexp_t		vfs_stdcheckexp;
918vfs_init_t		vfs_stdinit;
919vfs_uninit_t		vfs_stduninit;
920vfs_extattrctl_t	vfs_stdextattrctl;
921vfs_sysctl_t		vfs_stdsysctl;
922
923void	syncer_suspend(void);
924void	syncer_resume(void);
925
926#else /* !_KERNEL */
927
928#include <sys/cdefs.h>
929
930struct stat;
931
932__BEGIN_DECLS
933int	fhopen(const struct fhandle *, int);
934int	fhstat(const struct fhandle *, struct stat *);
935int	fhstatfs(const struct fhandle *, struct statfs *);
936int	fstatfs(int, struct statfs *);
937int	getfh(const char *, fhandle_t *);
938int	getfsstat(struct statfs *, long, int);
939int	getmntinfo(struct statfs **, int);
940int	lgetfh(const char *, fhandle_t *);
941int	mount(const char *, const char *, int, void *);
942int	nmount(struct iovec *, unsigned int, int);
943int	statfs(const char *, struct statfs *);
944int	unmount(const char *, int);
945
946/* C library stuff */
947int	getvfsbyname(const char *, struct xvfsconf *);
948__END_DECLS
949
950#endif /* _KERNEL */
951
952#endif /* !_SYS_MOUNT_H_ */
953