vnode.h revision 101368
1/*
2 * Copyright (c) 1989, 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 * 3. All advertising materials mentioning features or use of this software
14 *    must display the following acknowledgement:
15 *	This product includes software developed by the University of
16 *	California, Berkeley and its contributors.
17 * 4. Neither the name of the University nor the names of its contributors
18 *    may be used to endorse or promote products derived from this software
19 *    without specific prior written permission.
20 *
21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
31 * SUCH DAMAGE.
32 *
33 *	@(#)vnode.h	8.7 (Berkeley) 2/4/94
34 * $FreeBSD: head/sys/sys/vnode.h 101368 2002-08-05 10:25:56Z jeff $
35 */
36
37#ifndef _SYS_VNODE_H_
38#define	_SYS_VNODE_H_
39
40/*
41 * XXX - compatability until lockmgr() goes away or all the #includes are
42 * updated.
43 */
44#include <sys/lockmgr.h>
45
46#include <sys/queue.h>
47#include <sys/_lock.h>
48#include <sys/_mutex.h>
49#include <sys/selinfo.h>
50#include <sys/uio.h>
51#include <sys/acl.h>
52#include <sys/mac.h>
53#include <vm/uma.h>
54
55/*
56 * The vnode is the focus of all file activity in UNIX.  There is a
57 * unique vnode allocated for each active file, each current directory,
58 * each mounted-on file, text file, and the root.
59 */
60
61/*
62 * Vnode types.  VNON means no type.
63 */
64enum vtype	{ VNON, VREG, VDIR, VBLK, VCHR, VLNK, VSOCK, VFIFO, VBAD };
65
66/*
67 * Vnode tag types.
68 * These are for the benefit of external programs only (e.g., pstat)
69 * and should NEVER be inspected by the kernel.
70 */
71enum vtagtype	{
72	VT_NON, VT_UFS, VT_NFS, VT_UNUSED, VT_PC, VT_LFS, VT_LOFS, VT_FDESC,
73	VT_PORTAL, VT_NULL, VT_UMAP, VT_KERNFS, VT_PROCFS, VT_AFS, VT_ISOFS,
74	VT_UNION, VT_MSDOSFS, VT_DEVFS, VT_TFS, VT_VFS, VT_CODA, VT_NTFS,
75	VT_HPFS, VT_NWFS, VT_PSEUDOFS, VT_SMBFS, VT_UDF
76};
77
78/*
79 * Each underlying filesystem allocates its own private area and hangs
80 * it from v_data.  If non-null, this area is freed in getnewvnode().
81 */
82TAILQ_HEAD(buflists, buf);
83
84typedef	int	vop_t(void *);
85struct namecache;
86
87struct vpollinfo {
88	struct	mtx vpi_lock;		/* lock to protect below */
89	struct	selinfo vpi_selinfo;	/* identity of poller(s) */
90	short	vpi_events;		/* what they are looking for */
91	short	vpi_revents;		/* what has happened */
92};
93
94/*
95 * Reading or writing any of these items requires holding the appropriate lock.
96 *
97 * Lock reference:
98 * V		vnode lock
99 * I		inter lock
100 * F		freelist mutex
101 * M		mntvnodes mutex
102 * P		pollinfo lock
103 */
104
105struct vnode {
106	struct	mtx v_interlock;		/* lock on usecount and flag */
107	u_long	v_iflag;			/* I vnode flags (see below) */
108	int	v_usecount;			/* I ref count of users */
109	int	v_writecount;			/* I ref count of writers */
110	long	v_numoutput;			/* I writes in progress */
111	struct thread *v_vxproc;		/* I thread owning VXLOCK */
112	int	v_holdcnt;			/* I page & buffer references */
113	u_long	v_vflag;			/* V vnode flags */
114	u_long	v_id;				/* capability identifier */
115	struct	mount *v_mount;			/* ptr to vfs we are in */
116	vop_t	**v_op;				/* vnode operations vector */
117	TAILQ_ENTRY(vnode) v_freelist;		/* F vnode freelist */
118	TAILQ_ENTRY(vnode) v_nmntvnodes;	/* M vnodes for mount point */
119	struct	buflists v_cleanblkhd;		/* SORTED clean blocklist */
120	struct buf	*v_cleanblkroot;	/* clean buf splay tree root */
121	struct	buflists v_dirtyblkhd;		/* SORTED dirty blocklist */
122	struct buf	*v_dirtyblkroot;	/* dirty buf splay tree root */
123	LIST_ENTRY(vnode) v_synclist;		/* vnodes with dirty buffers */
124	enum	vtype v_type;			/* vnode type */
125	union {
126		struct mount	*vu_mountedhere;/* ptr to mounted vfs (VDIR) */
127		struct socket	*vu_socket;	/* unix ipc (VSOCK) */
128		struct {
129			struct specinfo	*vu_specinfo; /* device (VCHR, VBLK) */
130			SLIST_ENTRY(vnode) vu_specnext;
131		} vu_spec;
132		struct fifoinfo	*vu_fifoinfo;	/* fifo (VFIFO) */
133	} v_un;
134	daddr_t	v_lastw;			/* last write (write cluster) */
135	daddr_t	v_cstart;			/* start block of cluster */
136	daddr_t	v_lasta;			/* last allocation (cluster) */
137	int	v_clen;				/* length of current cluster */
138	struct vm_object *v_object;		/* Place to store VM object */
139	struct	lock v_lock;			/* used if fs don't have one */
140	struct	lock *v_vnlock;			/* pointer to vnode lock */
141	enum	vtagtype v_tag;			/* type of underlying data */
142	void	*v_data;			/* private data for fs */
143	LIST_HEAD(, namecache) v_cache_src;	/* Cache entries from us */
144	TAILQ_HEAD(, namecache) v_cache_dst;	/* Cache entries to us */
145	struct	vnode *v_dd;			/* .. vnode */
146	u_long	v_ddid;				/* .. capability identifier */
147	struct vpollinfo *v_pollinfo;		/* P Poll events */
148	struct label v_label;			/* MAC label for vnode */
149#ifdef	DEBUG_LOCKS
150	const char *filename;			/* Source file doing locking */
151	int line;				/* Line number doing locking */
152#endif
153	udev_t	v_cachedfs;			/* cached fs id */
154	ino_t	v_cachedid;			/* cached file id */
155};
156#define	v_mountedhere	v_un.vu_mountedhere
157#define	v_socket	v_un.vu_socket
158#define	v_rdev		v_un.vu_spec.vu_specinfo
159#define	v_specnext	v_un.vu_spec.vu_specnext
160#define	v_fifoinfo	v_un.vu_fifoinfo
161
162/*
163 * Userland version of struct vnode, for sysctl.
164 */
165struct xvnode {
166	size_t	xv_size;			/* sizeof(struct xvnode) */
167	void	*xv_vnode;			/* address of real vnode */
168	u_long	xv_flag;			/* vnode vflags */
169	int	xv_usecount;			/* reference count of users */
170	int	xv_writecount;			/* reference count of writers */
171	int	xv_holdcnt;			/* page & buffer references */
172	u_long	xv_id;				/* capability identifier */
173	void	*xv_mount;			/* address of parent mount */
174	long	xv_numoutput;			/* num of writes in progress */
175	enum	vtype xv_type;			/* vnode type */
176	union {
177		void	*xvu_socket;		/* socket, if VSOCK */
178		void	*xvu_fifo;		/* fifo, if VFIFO */
179		udev_t	xvu_rdev;		/* maj/min, if VBLK/VCHR */
180		struct {
181			udev_t	xvu_dev;	/* device, if VDIR/VREG/VLNK */
182			ino_t	xvu_ino;	/* id, if VDIR/VREG/VLNK */
183		};
184	} xv_un;
185};
186#define xv_socket	xv_un.xvu_socket
187#define xv_fifo		xv_un.xvu_fifo
188#define xv_rdev		xv_un.xvu_rdev
189#define xv_dev		xv_un.xvu_dev
190#define xv_ino		xv_un.xvu_ino
191
192#define	VN_POLLEVENT(vp, events)				\
193	do {							\
194		if ((vp)->v_pollinfo != NULL && 		\
195		    (vp)->v_pollinfo->vpi_events & (events))	\
196			vn_pollevent((vp), (events));		\
197	} while (0)
198
199#define VN_KNOTE(vp, b)						\
200	do {							\
201		if ((vp)->v_pollinfo != NULL)			\
202			KNOTE(&vp->v_pollinfo->vpi_selinfo.si_note, (b)); \
203	} while (0)
204
205/*
206 * Vnode flags.
207 *	VI flags are protected by interlock and live in v_iflag
208 *	VV flags are protected by the vnode lock and live in v_vflag
209 */
210#define	VI_XLOCK	0x0001	/* vnode is locked to change vtype */
211#define	VI_XWANT	0x0002	/* thread is waiting for vnode */
212#define	VI_BWAIT	0x0004	/* waiting for output to complete */
213#define	VI_OLOCK	0x0008	/* vnode is locked waiting for an object */
214#define	VI_OWANT	0x0010	/* a thread is waiting for VOLOCK */
215#define	VI_MOUNT	0x0020	/* Mount in progress */
216#define	VI_AGE		0x0040	/* Insert vnode at head of free list */
217#define	VI_DOOMED	0x0080	/* This vnode is being recycled */
218#define	VI_FREE		0x0100	/* This vnode is on the freelist */
219#define	VI_OBJDIRTY	0x0400	/* object might be dirty */
220/*
221 * XXX VI_ONWORKLST could be replaced with a check for NULL list elements
222 * in v_synclist.
223 */
224#define	VI_ONWORKLST	0x0200	/* On syncer work-list */
225
226#define	VV_ROOT		0x0001	/* root of its filesystem */
227#define	VV_ISTTY	0x0002	/* vnode represents a tty */
228#define	VV_NOSYNC	0x0004	/* unlinked, stop syncing */
229#define	VV_OBJBUF	0x0008	/* Allocate buffers in VM object */
230#define	VV_CACHEDLABEL	0x0010	/* Vnode has valid cached MAC label */
231#define	VV_TEXT		0x0020	/* vnode is a pure text prototype */
232#define	VV_COPYONWRITE	0x0040	/* vnode is doing copy-on-write */
233#define	VV_SYSTEM	0x0080	/* vnode being used by kernel */
234
235/*
236 * Vnode attributes.  A field value of VNOVAL represents a field whose value
237 * is unavailable (getattr) or which is not to be changed (setattr).
238 */
239struct vattr {
240	enum vtype	va_type;	/* vnode type (for create) */
241	u_short		va_mode;	/* files access mode and type */
242	short		va_nlink;	/* number of references to file */
243	uid_t		va_uid;		/* owner user id */
244	gid_t		va_gid;		/* owner group id */
245	udev_t		va_fsid;	/* filesystem id */
246	long		va_fileid;	/* file id */
247	u_quad_t	va_size;	/* file size in bytes */
248	long		va_blocksize;	/* blocksize preferred for i/o */
249	struct timespec	va_atime;	/* time of last access */
250	struct timespec	va_mtime;	/* time of last modification */
251	struct timespec	va_ctime;	/* time file changed */
252	struct timespec	va_birthtime;	/* time file created */
253	u_long		va_gen;		/* generation number of file */
254	u_long		va_flags;	/* flags defined for file */
255	udev_t		va_rdev;	/* device the special file represents */
256	u_quad_t	va_bytes;	/* bytes of disk space held by file */
257	u_quad_t	va_filerev;	/* file modification number */
258	u_int		va_vaflags;	/* operations flags, see below */
259	long		va_spare;	/* remain quad aligned */
260};
261
262/*
263 * Flags for va_vaflags.
264 */
265#define	VA_UTIMES_NULL	0x01		/* utimes argument was NULL */
266#define	VA_EXCLUSIVE	0x02		/* exclusive create request */
267
268/*
269 * Flags for ioflag. (high 16 bits used to ask for read-ahead and
270 * help with write clustering)
271 */
272#define	IO_UNIT		0x0001		/* do I/O as atomic unit */
273#define	IO_APPEND	0x0002		/* append write to end */
274#define	IO_SYNC		0x0004		/* do I/O synchronously */
275#define	IO_NODELOCKED	0x0008		/* underlying node already locked */
276#define	IO_NDELAY	0x0010		/* FNDELAY flag set in file table */
277#define	IO_VMIO		0x0020		/* data already in VMIO space */
278#define	IO_INVAL	0x0040		/* invalidate after I/O */
279#define	IO_ASYNC	0x0080		/* bawrite rather then bdwrite */
280#define	IO_DIRECT	0x0100		/* attempt to bypass buffer cache */
281#define	IO_NOWDRAIN	0x0200		/* do not block on wdrain */
282#define	IO_EXT		0x0400		/* operate on external attributes */
283#define	IO_NORMAL	0x0800		/* operate on regular data */
284
285/*
286 *  Modes.  Some values same as Ixxx entries from inode.h for now.
287 */
288#define	VEXEC	000100		/* execute/search permission */
289#define	VWRITE	000200		/* write permission */
290#define	VREAD	000400		/* read permission */
291#define	VSVTX	001000		/* save swapped text even after use */
292#define	VSGID	002000		/* set group id on execution */
293#define	VSUID	004000		/* set user id on execution */
294#define	VADMIN	010000		/* permission to administer */
295#define	VSTAT	020000		/* permission to retrieve attrs */
296#define	VAPPEND	040000		/* permission to write/append */
297#define	VALLPERM	(VEXEC | VWRITE | VREAD | VADMIN | VSTAT | VAPPEND)
298
299/*
300 * Token indicating no attribute value yet assigned.
301 */
302#define	VNOVAL	(-1)
303
304/*
305 * LK_TIMELOCK timeout for vnode locks (used mainly by the pageout daemon)
306 */
307#define VLKTIMEOUT	(hz / 20 + 1)
308
309#ifdef _KERNEL
310
311#ifdef MALLOC_DECLARE
312MALLOC_DECLARE(M_VNODE);
313#endif
314
315/*
316 * Convert between vnode types and inode formats (since POSIX.1
317 * defines mode word of stat structure in terms of inode formats).
318 */
319extern enum vtype	iftovt_tab[];
320extern int		vttoif_tab[];
321#define	IFTOVT(mode)	(iftovt_tab[((mode) & S_IFMT) >> 12])
322#define	VTTOIF(indx)	(vttoif_tab[(int)(indx)])
323#define	MAKEIMODE(indx, mode)	(int)(VTTOIF(indx) | (mode))
324
325/*
326 * Flags to various vnode functions.
327 */
328#define	SKIPSYSTEM	0x0001	/* vflush: skip vnodes marked VSYSTEM */
329#define	FORCECLOSE	0x0002	/* vflush: force file closure */
330#define	WRITECLOSE	0x0004	/* vflush: only close writable files */
331#define	DOCLOSE		0x0008	/* vclean: close active files */
332#define	V_SAVE		0x0001	/* vinvalbuf: sync file first */
333#define	V_ALT		0x0002	/* vinvalbuf: invalidate only alternate bufs */
334#define	V_NORMAL	0x0004	/* vinvalbuf: invalidate only regular bufs */
335#define	REVOKEALL	0x0001	/* vop_revoke: revoke all aliases */
336#define	V_WAIT		0x0001	/* vn_start_write: sleep for suspend */
337#define	V_NOWAIT	0x0002	/* vn_start_write: don't sleep for suspend */
338#define	V_XSLEEP	0x0004	/* vn_start_write: just return after sleep */
339
340#define	VREF(vp)	vref(vp)
341
342
343#ifdef DIAGNOSTIC
344#define	VATTR_NULL(vap)	vattr_null(vap)
345#else
346#define	VATTR_NULL(vap)	(*(vap) = va_null)	/* initialize a vattr */
347#endif /* DIAGNOSTIC */
348
349#define	NULLVP	((struct vnode *)NULL)
350
351#define	VNODEOP_SET(f) \
352	C_SYSINIT(f##init, SI_SUB_VFS, SI_ORDER_SECOND, vfs_add_vnodeops, &f); \
353	C_SYSUNINIT(f##uninit, SI_SUB_VFS, SI_ORDER_SECOND, vfs_rm_vnodeops, &f);
354
355/*
356 * Global vnode data.
357 */
358extern	struct vnode *rootvnode;	/* root (i.e. "/") vnode */
359extern	int desiredvnodes;		/* number of vnodes desired */
360extern	uma_zone_t namei_zone;
361extern	int prtactive;			/* nonzero to call vprint() */
362extern	struct vattr va_null;		/* predefined null vattr structure */
363extern	int vfs_ioopt;
364
365/*
366 * Macro/function to check for client cache inconsistency w.r.t. leasing.
367 */
368#define	LEASE_READ	0x1		/* Check lease for readers */
369#define	LEASE_WRITE	0x2		/* Check lease for modifiers */
370
371
372extern void	(*lease_updatetime)(int deltat);
373
374/* Requires interlock */
375#define	VSHOULDFREE(vp)	\
376	(!((vp)->v_iflag & (VI_FREE|VI_DOOMED)) && \
377	 !(vp)->v_holdcnt && !(vp)->v_usecount && \
378	 (!(vp)->v_object || \
379	  !((vp)->v_object->ref_count || (vp)->v_object->resident_page_count)))
380
381/* Requires interlock */
382#define VMIGHTFREE(vp) \
383	(!((vp)->v_iflag & (VI_FREE|VI_DOOMED|VI_XLOCK)) &&	\
384	 LIST_EMPTY(&(vp)->v_cache_src) && !(vp)->v_usecount)
385
386/* Requires interlock */
387#define	VSHOULDBUSY(vp)	\
388	(((vp)->v_iflag & VI_FREE) && \
389	 ((vp)->v_holdcnt || (vp)->v_usecount))
390
391#define	VI_LOCK(vp)	mtx_lock(&(vp)->v_interlock)
392#define	VI_TRYLOCK(vp)	mtx_trylock(&(vp)->v_interlock)
393#define	VI_UNLOCK(vp)	mtx_unlock(&(vp)->v_interlock)
394#define	VI_MTX(vp)	(&(vp)->v_interlock)
395
396#endif /* _KERNEL */
397
398
399/*
400 * Mods for extensibility.
401 */
402
403/*
404 * Flags for vdesc_flags:
405 */
406#define	VDESC_MAX_VPS		16
407/* Low order 16 flag bits are reserved for willrele flags for vp arguments. */
408#define	VDESC_VP0_WILLRELE	0x0001
409#define	VDESC_VP1_WILLRELE	0x0002
410#define	VDESC_VP2_WILLRELE	0x0004
411#define	VDESC_VP3_WILLRELE	0x0008
412#define	VDESC_NOMAP_VPP		0x0100
413#define	VDESC_VPP_WILLRELE	0x0200
414
415/*
416 * VDESC_NO_OFFSET is used to identify the end of the offset list
417 * and in places where no such field exists.
418 */
419#define VDESC_NO_OFFSET -1
420
421/*
422 * This structure describes the vnode operation taking place.
423 */
424struct vnodeop_desc {
425	int	 vdesc_offset;		/* offset in vector,first for speed */
426	char	*vdesc_name;		/* a readable name for debugging */
427	int	 vdesc_flags;		/* VDESC_* flags */
428
429	/*
430	 * These ops are used by bypass routines to map and locate arguments.
431	 * Creds and procs are not needed in bypass routines, but sometimes
432	 * they are useful to (for example) transport layers.
433	 * Nameidata is useful because it has a cred in it.
434	 */
435	int	*vdesc_vp_offsets;	/* list ended by VDESC_NO_OFFSET */
436	int	vdesc_vpp_offset;	/* return vpp location */
437	int	vdesc_cred_offset;	/* cred location, if any */
438	int	vdesc_thread_offset;	/* thread location, if any */
439	int	vdesc_componentname_offset; /* if any */
440	/*
441	 * Finally, we've got a list of private data (about each operation)
442	 * for each transport layer.  (Support to manage this list is not
443	 * yet part of BSD.)
444	 */
445	caddr_t	*vdesc_transports;
446};
447
448#ifdef _KERNEL
449/*
450 * A list of all the operation descs.
451 */
452extern struct vnodeop_desc *vnodeop_descs[];
453
454/*
455 * Interlock for scanning list of vnodes attached to a mountpoint
456 */
457extern struct mtx mntvnode_mtx;
458
459/*
460 * This macro is very helpful in defining those offsets in the vdesc struct.
461 *
462 * This is stolen from X11R4.  I ignored all the fancy stuff for
463 * Crays, so if you decide to port this to such a serious machine,
464 * you might want to consult Intrinsic.h's XtOffset{,Of,To}.
465 */
466#define	VOPARG_OFFSET(p_type,field) \
467	((int) (((char *) (&(((p_type)NULL)->field))) - ((char *) NULL)))
468#define	VOPARG_OFFSETOF(s_type,field) \
469	VOPARG_OFFSET(s_type*,field)
470#define	VOPARG_OFFSETTO(S_TYPE,S_OFFSET,STRUCT_P) \
471	((S_TYPE)(((char*)(STRUCT_P))+(S_OFFSET)))
472
473
474/*
475 * This structure is used to configure the new vnodeops vector.
476 */
477struct vnodeopv_entry_desc {
478	struct vnodeop_desc *opve_op;   /* which operation this is */
479	vop_t *opve_impl;		/* code implementing this operation */
480};
481struct vnodeopv_desc {
482			/* ptr to the ptr to the vector where op should go */
483	vop_t ***opv_desc_vector_p;
484	struct vnodeopv_entry_desc *opv_desc_ops;   /* null terminated list */
485};
486
487/*
488 * A generic structure.
489 * This can be used by bypass routines to identify generic arguments.
490 */
491struct vop_generic_args {
492	struct vnodeop_desc *a_desc;
493	/* other random data follows, presumably */
494};
495
496
497#ifdef DEBUG_VFS_LOCKS
498/*
499 * Macros to aid in tracing VFS locking problems.  Not totally
500 * reliable since if the thread sleeps between changing the lock
501 * state and checking it with the assert, some other thread could
502 * change the state.  They are good enough for debugging a single
503 * filesystem using a single-threaded test.  I find that 'cvs co src'
504 * is a pretty good test.
505 */
506
507extern int vfs_badlock_panic;
508extern int vfs_badlock_print;
509
510/*
511 * This only exists to supress warnings from unlocked specfs accesses.  It is
512 * no longer ok to have an unlocked VFS.
513 */
514
515#define IGNORE_LOCK(vp)	((vp)->v_type == VCHR || (vp)->v_type == VBAD)
516
517#define ASSERT_VOP_LOCKED(vp, str)					\
518do {									\
519	struct vnode *_vp = (vp);					\
520									\
521	if (_vp && !IGNORE_LOCK(_vp) && !VOP_ISLOCKED(_vp, NULL)) {	\
522		if (vfs_badlock_print)					\
523			printf("%s: %p is not locked but should be\n",	\
524			    str, _vp);					\
525		if (vfs_badlock_panic)					\
526			Debugger("Lock violation.\n");			\
527	}								\
528} while (0)
529
530#define ASSERT_VOP_UNLOCKED(vp, str)					\
531do {									\
532	struct vnode *_vp = (vp);					\
533	int lockstate;							\
534									\
535	if (_vp && !IGNORE_LOCK(_vp)) {					\
536		lockstate = VOP_ISLOCKED(_vp, curthread);		\
537		if (lockstate == LK_EXCLUSIVE) {			\
538			if (vfs_badlock_print)				\
539				printf("%s: %p is locked but should not be\n",	\
540				    str, _vp);				\
541			if (vfs_badlock_panic)				\
542				Debugger("Lock Violation.\n");		\
543		}							\
544	}								\
545} while (0)
546
547#define ASSERT_VOP_ELOCKED(vp, str)					\
548do {									\
549	struct vnode *_vp = (vp);					\
550									\
551	if (_vp && !IGNORE_LOCK(_vp) &&					\
552	    VOP_ISLOCKED(_vp, curthread) != LK_EXCLUSIVE) {		\
553		if (vfs_badlock_print)					\
554			printf("%s: %p is not exclusive locked but should be\n",\
555			    str, _vp);					\
556		if (vfs_badlock_panic)					\
557			Debugger("Lock violation.\n");			\
558	}								\
559} while (0)
560
561#define ASSERT_VOP_ELOCKED_OTHER(vp, str)				\
562do {									\
563	struct vnode *_vp = (vp);					\
564									\
565	if (_vp && !IGNORE_LOCK(_vp) &&					\
566	    VOP_ISLOCKED(_vp, curthread) != LK_EXCLOTHER) {		\
567		if (vfs_badlock_print)					\
568			printf("%s: %p is not exclusive locked by another thread\n",\
569			    str, _vp);					\
570		if (vfs_badlock_panic)					\
571			Debugger("Lock violation.\n");			\
572	}								\
573} while (0)
574
575#define ASSERT_VOP_SLOCKED(vp, str)					\
576do {									\
577	struct vnode *_vp = (vp);					\
578									\
579	if (_vp && !IGNORE_LOCK(_vp) &&					\
580	    VOP_ISLOCKED(_vp, NULL) != LK_SHARED) {			\
581		if (vfs_badlock_print)					\
582			printf("%s: %p is not locked shared but should be",\
583		    str, _vp);						\
584		if (vfs_badlock_panic)					\
585			Debugger("Lock violation.\n");			\
586	}								\
587} while (0)
588
589void vop_rename_pre(void *a);
590void vop_strategy_pre(void *a);
591void vop_lookup_pre(void *a);
592void vop_lookup_post(void *a, int rc);
593
594#else
595
596#define ASSERT_VOP_LOCKED(vp, str)
597#define ASSERT_VOP_UNLOCKED(vp, str)
598
599#endif
600
601/*
602 * VOCALL calls an op given an ops vector.  We break it out because BSD's
603 * vclean changes the ops vector and then wants to call ops with the old
604 * vector.
605 */
606#define VOCALL(OPSV,OFF,AP) (( *((OPSV)[(OFF)])) (AP))
607
608/*
609 * This call works for vnodes in the kernel.
610 */
611#define VCALL(VP,OFF,AP) VOCALL((VP)->v_op,(OFF),(AP))
612#define VDESC(OP) (& __CONCAT(OP,_desc))
613#define VOFFSET(OP) (VDESC(OP)->vdesc_offset)
614
615/*
616 * VMIO support inline
617 */
618
619extern int vmiodirenable;
620
621static __inline int
622vn_canvmio(struct vnode *vp)
623{
624      if (vp && (vp->v_type == VREG || (vmiodirenable && vp->v_type == VDIR)))
625		return(TRUE);
626	return(FALSE);
627}
628
629/*
630 * Finally, include the default set of vnode operations.
631 */
632#include "vnode_if.h"
633
634/*
635 * Public vnode manipulation functions.
636 */
637struct componentname;
638struct file;
639struct mount;
640struct nameidata;
641struct ostat;
642struct thread;
643struct proc;
644struct stat;
645struct nstat;
646struct ucred;
647struct uio;
648struct vattr;
649struct vnode;
650
651extern int	(*lease_check_hook)(struct vop_lease_args *);
652extern int	(*softdep_fsync_hook)(struct vnode *);
653extern int	(*softdep_process_worklist_hook)(struct mount *);
654
655struct	vnode *addaliasu(struct vnode *vp, udev_t nvp_rdev);
656int	bdevvp(dev_t dev, struct vnode **vpp);
657/* cache_* may belong in namei.h. */
658void	cache_enter(struct vnode *dvp, struct vnode *vp,
659	    struct componentname *cnp);
660int	cache_lookup(struct vnode *dvp, struct vnode **vpp,
661	    struct componentname *cnp);
662void	cache_purge(struct vnode *vp);
663void	cache_purgevfs(struct mount *mp);
664int	cache_leaf_test(struct vnode *vp);
665void	cvtstat(struct stat *st, struct ostat *ost);
666void	cvtnstat(struct stat *sb, struct nstat *nsb);
667int	getnewvnode(enum vtagtype tag,
668	    struct mount *mp, vop_t **vops, struct vnode **vpp);
669int	lease_check(struct vop_lease_args *ap);
670int	spec_vnoperate(struct vop_generic_args *);
671int	speedup_syncer(void);
672#define textvp_fullpath(p, rb, rfb) \
673	vn_fullpath(FIRST_THREAD_IN_PROC(p), (p)->p_textvp, rb, rfb)
674int	vn_fullpath(struct thread *td, struct vnode *vn,
675	    char **retbuf, char **freebuf);
676int	vaccess(enum vtype type, mode_t file_mode, uid_t uid, gid_t gid,
677	    mode_t acc_mode, struct ucred *cred, int *privused);
678int	vaccess_acl_posix1e(enum vtype type, uid_t file_uid,
679	    gid_t file_gid, struct acl *acl, mode_t acc_mode,
680	    struct ucred *cred, int *privused);
681void	vattr_null(struct vattr *vap);
682int	vcount(struct vnode *vp);
683void	vdrop(struct vnode *);
684int	vfinddev(dev_t dev, enum vtype type, struct vnode **vpp);
685void	vfs_add_vnodeops(const void *);
686void	vfs_rm_vnodeops(const void *);
687int	vflush(struct mount *mp, int rootrefs, int flags);
688int	vget(struct vnode *vp, int lockflag, struct thread *td);
689void	vgone(struct vnode *vp);
690void	vgonel(struct vnode *vp, struct thread *td);
691void	vhold(struct vnode *);
692int	vinvalbuf(struct vnode *vp, int save, struct ucred *cred,
693	    struct thread *td, int slpflag, int slptimeo);
694int	vtruncbuf(struct vnode *vp, struct ucred *cred, struct thread *td,
695	    off_t length, int blksize);
696void	vprint(char *label, struct vnode *vp);
697int	vrecycle(struct vnode *vp, struct mtx *inter_lkp,
698	    struct thread *td);
699int	vn_close(struct vnode *vp,
700	    int flags, struct ucred *cred, struct thread *td);
701void	vn_finished_write(struct mount *mp);
702int	vn_isdisk(struct vnode *vp, int *errp);
703int	vn_lock(struct vnode *vp, int flags, struct thread *td);
704#ifdef	DEBUG_LOCKS
705int	debug_vn_lock(struct vnode *vp, int flags, struct thread *p,
706	    const char *filename, int line);
707#define vn_lock(vp,flags,p) debug_vn_lock(vp,flags,p,__FILE__,__LINE__)
708#endif
709int	vn_mkdir(char *path, int mode, enum uio_seg segflg, struct thread *td);
710int	vn_open(struct nameidata *ndp, int *flagp, int cmode);
711int	vn_open_cred(struct nameidata *ndp, int *flagp, int cmode,
712	    struct ucred *cred);
713void	vn_pollevent(struct vnode *vp, int events);
714void	vn_pollgone(struct vnode *vp);
715int	vn_pollrecord(struct vnode *vp, struct thread *p, int events);
716int	vn_rdwr(enum uio_rw rw, struct vnode *vp, caddr_t base,
717	    int len, off_t offset, enum uio_seg segflg, int ioflg,
718	    struct ucred *cred, int *aresid, struct thread *td);
719int	vn_rdwr_inchunks(enum uio_rw rw, struct vnode *vp, caddr_t base,
720	    int len, off_t offset, enum uio_seg segflg, int ioflg,
721	    struct ucred *cred, int *aresid, struct thread *td);
722int	vn_stat(struct vnode *vp, struct stat *sb, struct thread *td);
723int	vn_start_write(struct vnode *vp, struct mount **mpp, int flags);
724dev_t	vn_todev(struct vnode *vp);
725int	vn_write_suspend_wait(struct vnode *vp, struct mount *mp,
726	    int flags);
727int	vn_writechk(struct vnode *vp);
728int	vn_extattr_get(struct vnode *vp, int ioflg, int attrnamespace,
729	    const char *attrname, int *buflen, char *buf, struct thread *td);
730int	vn_extattr_set(struct vnode *vp, int ioflg, int attrnamespace,
731	    const char *attrname, int buflen, char *buf, struct thread *td);
732int	vn_extattr_rm(struct vnode *vp, int ioflg, int attrnamespace,
733	    const char *attrname, struct thread *td);
734int	vfs_cache_lookup(struct vop_lookup_args *ap);
735int	vfs_object_create(struct vnode *vp, struct thread *td,
736	    struct ucred *cred);
737void	vfs_timestamp(struct timespec *);
738void	vfs_write_resume(struct mount *mp);
739void	vfs_write_suspend(struct mount *mp);
740int	vop_stdbmap(struct vop_bmap_args *);
741int	vop_stdgetwritemount(struct vop_getwritemount_args *);
742int	vop_stdgetpages(struct vop_getpages_args *);
743int	vop_stdinactive(struct vop_inactive_args *);
744int	vop_stdislocked(struct vop_islocked_args *);
745int	vop_stdlock(struct vop_lock_args *);
746int	vop_stdputpages(struct vop_putpages_args *);
747int	vop_stdunlock(struct vop_unlock_args *);
748int	vop_noislocked(struct vop_islocked_args *);
749int	vop_nolock(struct vop_lock_args *);
750int	vop_nopoll(struct vop_poll_args *);
751int	vop_nounlock(struct vop_unlock_args *);
752int	vop_stdpathconf(struct vop_pathconf_args *);
753int	vop_stdpoll(struct vop_poll_args *);
754int	vop_revoke(struct vop_revoke_args *);
755int	vop_sharedlock(struct vop_lock_args *);
756int	vop_eopnotsupp(struct vop_generic_args *ap);
757int	vop_ebadf(struct vop_generic_args *ap);
758int	vop_einval(struct vop_generic_args *ap);
759int	vop_enotty(struct vop_generic_args *ap);
760int	vop_defaultop(struct vop_generic_args *ap);
761int	vop_null(struct vop_generic_args *ap);
762int	vop_panic(struct vop_generic_args *ap);
763int	vop_stdcreatevobject(struct vop_createvobject_args *ap);
764int	vop_stddestroyvobject(struct vop_destroyvobject_args *ap);
765int	vop_stdgetvobject(struct vop_getvobject_args *ap);
766
767void	vfree(struct vnode *);
768void	vput(struct vnode *vp);
769void	vrele(struct vnode *vp);
770void	vref(struct vnode *vp);
771void	vbusy(struct vnode *vp);
772void 	v_addpollinfo(struct vnode *vp);
773
774extern	vop_t **default_vnodeop_p;
775extern	vop_t **spec_vnodeop_p;
776extern	vop_t **dead_vnodeop_p;
777
778#endif /* _KERNEL */
779
780#endif /* !_SYS_VNODE_H_ */
781