vnode.h revision 101308
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 101308 2002-08-04 10:29:36Z 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 * v_freelist is locked by the global vnode_free_list mutex.
97 * v_mntvnodes is locked by the global mntvnodes mutex.
98 * v_iflag, v_usecount, v_holdcount and v_writecount are
99 *    locked by the v_interlock mutex.
100 * v_pollinfo is locked by the lock contained inside it.
101 * V vnode lock
102 * I inter lock
103 */
104struct vnode {
105	struct	mtx v_interlock;		/* lock on usecount and flag */
106	u_long	v_iflag;			/* I vnode flags (see below) */
107	int	v_usecount;			/* I ref count of users */
108	int	v_writecount;			/* I ref count of writers */
109	long	v_numoutput;			/* I writes in progress */
110	struct thread *v_vxproc;		/* I thread owning VXLOCK */
111	u_long	v_vflag;			/* V vnode flags */
112	int	v_holdcnt;			/* page & buffer references */
113	u_long	v_id;				/* capability identifier */
114	struct	mount *v_mount;			/* ptr to vfs we are in */
115	vop_t	**v_op;				/* vnode operations vector */
116	TAILQ_ENTRY(vnode) v_freelist;		/* vnode freelist */
117	TAILQ_ENTRY(vnode) v_nmntvnodes;	/* vnodes for mount point */
118	struct	buflists v_cleanblkhd;		/* SORTED clean blocklist */
119	struct buf	*v_cleanblkroot;	/* clean buf splay tree root */
120	struct	buflists v_dirtyblkhd;		/* SORTED dirty blocklist */
121	struct buf	*v_dirtyblkroot;	/* dirty buf splay tree root */
122	LIST_ENTRY(vnode) v_synclist;		/* vnodes with dirty buffers */
123	enum	vtype v_type;			/* vnode type */
124	union {
125		struct mount	*vu_mountedhere;/* ptr to mounted vfs (VDIR) */
126		struct socket	*vu_socket;	/* unix ipc (VSOCK) */
127		struct {
128			struct specinfo	*vu_specinfo; /* device (VCHR, VBLK) */
129			SLIST_ENTRY(vnode) vu_specnext;
130		} vu_spec;
131		struct fifoinfo	*vu_fifoinfo;	/* fifo (VFIFO) */
132	} v_un;
133	daddr_t	v_lastw;			/* last write (write cluster) */
134	daddr_t	v_cstart;			/* start block of cluster */
135	daddr_t	v_lasta;			/* last allocation (cluster) */
136	int	v_clen;				/* length of current cluster */
137	struct vm_object *v_object;		/* Place to store VM object */
138	struct	lock v_lock;			/* used if fs don't have one */
139	struct	lock *v_vnlock;			/* pointer to vnode lock */
140	enum	vtagtype v_tag;			/* type of underlying data */
141	void	*v_data;			/* private data for fs */
142	LIST_HEAD(, namecache) v_cache_src;	/* Cache entries from us */
143	TAILQ_HEAD(, namecache) v_cache_dst;	/* Cache entries to us */
144	struct	vnode *v_dd;			/* .. vnode */
145	u_long	v_ddid;				/* .. capability identifier */
146	struct vpollinfo *v_pollinfo;
147	struct label v_label;			/* MAC label for vnode */
148#ifdef	DEBUG_LOCKS
149	const char *filename;			/* Source file doing locking */
150	int line;				/* Line number doing locking */
151#endif
152	udev_t	v_cachedfs;			/* cached fs id */
153	ino_t	v_cachedid;			/* cached file id */
154};
155#define	v_mountedhere	v_un.vu_mountedhere
156#define	v_socket	v_un.vu_socket
157#define	v_rdev		v_un.vu_spec.vu_specinfo
158#define	v_specnext	v_un.vu_spec.vu_specnext
159#define	v_fifoinfo	v_un.vu_fifoinfo
160
161/*
162 * Userland version of struct vnode, for sysctl.
163 */
164struct xvnode {
165	size_t	xv_size;			/* sizeof(struct xvnode) */
166	void	*xv_vnode;			/* address of real vnode */
167	u_long	xv_flag;			/* vnode vflags */
168	int	xv_usecount;			/* reference count of users */
169	int	xv_writecount;			/* reference count of writers */
170	int	xv_holdcnt;			/* page & buffer references */
171	u_long	xv_id;				/* capability identifier */
172	void	*xv_mount;			/* address of parent mount */
173	long	xv_numoutput;			/* num of writes in progress */
174	enum	vtype xv_type;			/* vnode type */
175	union {
176		void	*xvu_socket;		/* socket, if VSOCK */
177		void	*xvu_fifo;		/* fifo, if VFIFO */
178		udev_t	xvu_rdev;		/* maj/min, if VBLK/VCHR */
179		struct {
180			udev_t	xvu_dev;	/* device, if VDIR/VREG/VLNK */
181			ino_t	xvu_ino;	/* id, if VDIR/VREG/VLNK */
182		};
183	} xv_un;
184};
185#define xv_socket	xv_un.xvu_socket
186#define xv_fifo		xv_un.xvu_fifo
187#define xv_rdev		xv_un.xvu_rdev
188#define xv_dev		xv_un.xvu_dev
189#define xv_ino		xv_un.xvu_ino
190
191#define	VN_POLLEVENT(vp, events)				\
192	do {							\
193		if ((vp)->v_pollinfo != NULL && 		\
194		    (vp)->v_pollinfo->vpi_events & (events))	\
195			vn_pollevent((vp), (events));		\
196	} while (0)
197
198#define VN_KNOTE(vp, b)						\
199	do {							\
200		if ((vp)->v_pollinfo != NULL)			\
201			KNOTE(&vp->v_pollinfo->vpi_selinfo.si_note, (b)); \
202	} while (0)
203
204/*
205 * Vnode flags.
206 *	VI flags are protected by interlock and live in v_iflag
207 *	VV flags are protected by the vnode lock and live in v_vflag
208 */
209#define	VI_XLOCK	0x0001	/* vnode is locked to change vtype */
210#define	VI_XWANT	0x0002	/* thread is waiting for vnode */
211#define	VI_BWAIT	0x0004	/* waiting for output to complete */
212#define	VI_OLOCK	0x0008	/* vnode is locked waiting for an object */
213#define	VI_OWANT	0x0010	/* a thread is waiting for VOLOCK */
214#define	VI_MOUNT	0x0020	/* Mount in progress */
215#define	VI_AGE		0x0040	/* Insert vnode at head of free list */
216#define	VI_DOOMED	0x0080	/* This vnode is being recycled */
217#define	VI_FREE		0x0100	/* This vnode is on the freelist */
218#define	VI_OBJDIRTY	0x0400	/* object might be dirty */
219/*
220 * XXX VI_ONWORKLST could be replaced with a check for NULL list elements
221 * in v_synclist.
222 */
223#define	VI_ONWORKLST	0x0200	/* On syncer work-list */
224
225#define	VV_ROOT		0x0001	/* root of its filesystem */
226#define	VV_ISTTY	0x0002	/* vnode represents a tty */
227#define	VV_NOSYNC	0x0004	/* unlinked, stop syncing */
228#define	VV_OBJBUF	0x0008	/* Allocate buffers in VM object */
229#define	VV_CACHEDLABEL	0x0010	/* Vnode has valid cached MAC label */
230#define	VV_TEXT		0x0020	/* vnode is a pure text prototype */
231#define	VV_COPYONWRITE	0x0040	/* vnode is doing copy-on-write */
232#define	VV_SYSTEM	0x0080	/* vnode being used by kernel */
233
234/*
235 * Vnode attributes.  A field value of VNOVAL represents a field whose value
236 * is unavailable (getattr) or which is not to be changed (setattr).
237 */
238struct vattr {
239	enum vtype	va_type;	/* vnode type (for create) */
240	u_short		va_mode;	/* files access mode and type */
241	short		va_nlink;	/* number of references to file */
242	uid_t		va_uid;		/* owner user id */
243	gid_t		va_gid;		/* owner group id */
244	udev_t		va_fsid;	/* filesystem id */
245	long		va_fileid;	/* file id */
246	u_quad_t	va_size;	/* file size in bytes */
247	long		va_blocksize;	/* blocksize preferred for i/o */
248	struct timespec	va_atime;	/* time of last access */
249	struct timespec	va_mtime;	/* time of last modification */
250	struct timespec	va_ctime;	/* time file changed */
251	struct timespec	va_birthtime;	/* time file created */
252	u_long		va_gen;		/* generation number of file */
253	u_long		va_flags;	/* flags defined for file */
254	udev_t		va_rdev;	/* device the special file represents */
255	u_quad_t	va_bytes;	/* bytes of disk space held by file */
256	u_quad_t	va_filerev;	/* file modification number */
257	u_int		va_vaflags;	/* operations flags, see below */
258	long		va_spare;	/* remain quad aligned */
259};
260
261/*
262 * Flags for va_vaflags.
263 */
264#define	VA_UTIMES_NULL	0x01		/* utimes argument was NULL */
265#define	VA_EXCLUSIVE	0x02		/* exclusive create request */
266
267/*
268 * Flags for ioflag. (high 16 bits used to ask for read-ahead and
269 * help with write clustering)
270 */
271#define	IO_UNIT		0x0001		/* do I/O as atomic unit */
272#define	IO_APPEND	0x0002		/* append write to end */
273#define	IO_SYNC		0x0004		/* do I/O synchronously */
274#define	IO_NODELOCKED	0x0008		/* underlying node already locked */
275#define	IO_NDELAY	0x0010		/* FNDELAY flag set in file table */
276#define	IO_VMIO		0x0020		/* data already in VMIO space */
277#define	IO_INVAL	0x0040		/* invalidate after I/O */
278#define	IO_ASYNC	0x0080		/* bawrite rather then bdwrite */
279#define	IO_DIRECT	0x0100		/* attempt to bypass buffer cache */
280#define	IO_NOWDRAIN	0x0200		/* do not block on wdrain */
281#define	IO_EXT		0x0400		/* operate on external attributes */
282#define	IO_NORMAL	0x0800		/* operate on regular data */
283
284/*
285 *  Modes.  Some values same as Ixxx entries from inode.h for now.
286 */
287#define	VEXEC	000100		/* execute/search permission */
288#define	VWRITE	000200		/* write permission */
289#define	VREAD	000400		/* read permission */
290#define	VSVTX	001000		/* save swapped text even after use */
291#define	VSGID	002000		/* set group id on execution */
292#define	VSUID	004000		/* set user id on execution */
293#define	VADMIN	010000		/* permission to administer */
294#define	VSTAT	020000		/* permission to retrieve attrs */
295#define	VAPPEND	040000		/* permission to write/append */
296#define	VALLPERM	(VEXEC | VWRITE | VREAD | VADMIN | VSTAT | VAPPEND)
297
298/*
299 * Token indicating no attribute value yet assigned.
300 */
301#define	VNOVAL	(-1)
302
303/*
304 * LK_TIMELOCK timeout for vnode locks (used mainly by the pageout daemon)
305 */
306#define VLKTIMEOUT	(hz / 20 + 1)
307
308#ifdef _KERNEL
309
310#ifdef MALLOC_DECLARE
311MALLOC_DECLARE(M_VNODE);
312#endif
313
314/*
315 * Convert between vnode types and inode formats (since POSIX.1
316 * defines mode word of stat structure in terms of inode formats).
317 */
318extern enum vtype	iftovt_tab[];
319extern int		vttoif_tab[];
320#define	IFTOVT(mode)	(iftovt_tab[((mode) & S_IFMT) >> 12])
321#define	VTTOIF(indx)	(vttoif_tab[(int)(indx)])
322#define	MAKEIMODE(indx, mode)	(int)(VTTOIF(indx) | (mode))
323
324/*
325 * Flags to various vnode functions.
326 */
327#define	SKIPSYSTEM	0x0001	/* vflush: skip vnodes marked VSYSTEM */
328#define	FORCECLOSE	0x0002	/* vflush: force file closure */
329#define	WRITECLOSE	0x0004	/* vflush: only close writable files */
330#define	DOCLOSE		0x0008	/* vclean: close active files */
331#define	V_SAVE		0x0001	/* vinvalbuf: sync file first */
332#define	V_ALT		0x0002	/* vinvalbuf: invalidate only alternate bufs */
333#define	V_NORMAL	0x0004	/* vinvalbuf: invalidate only regular bufs */
334#define	REVOKEALL	0x0001	/* vop_revoke: revoke all aliases */
335#define	V_WAIT		0x0001	/* vn_start_write: sleep for suspend */
336#define	V_NOWAIT	0x0002	/* vn_start_write: don't sleep for suspend */
337#define	V_XSLEEP	0x0004	/* vn_start_write: just return after sleep */
338
339#define	VREF(vp)	vref(vp)
340
341
342#ifdef DIAGNOSTIC
343#define	VATTR_NULL(vap)	vattr_null(vap)
344#else
345#define	VATTR_NULL(vap)	(*(vap) = va_null)	/* initialize a vattr */
346#endif /* DIAGNOSTIC */
347
348#define	NULLVP	((struct vnode *)NULL)
349
350#define	VNODEOP_SET(f) \
351	C_SYSINIT(f##init, SI_SUB_VFS, SI_ORDER_SECOND, vfs_add_vnodeops, &f); \
352	C_SYSUNINIT(f##uninit, SI_SUB_VFS, SI_ORDER_SECOND, vfs_rm_vnodeops, &f);
353
354/*
355 * Global vnode data.
356 */
357extern	struct vnode *rootvnode;	/* root (i.e. "/") vnode */
358extern	int desiredvnodes;		/* number of vnodes desired */
359extern	uma_zone_t namei_zone;
360extern	int prtactive;			/* nonzero to call vprint() */
361extern	struct vattr va_null;		/* predefined null vattr structure */
362extern	int vfs_ioopt;
363
364/*
365 * Macro/function to check for client cache inconsistency w.r.t. leasing.
366 */
367#define	LEASE_READ	0x1		/* Check lease for readers */
368#define	LEASE_WRITE	0x2		/* Check lease for modifiers */
369
370
371extern void	(*lease_updatetime)(int deltat);
372
373/* Requires interlock */
374#define	VSHOULDFREE(vp)	\
375	(!((vp)->v_iflag & (VI_FREE|VI_DOOMED)) && \
376	 !(vp)->v_holdcnt && !(vp)->v_usecount && \
377	 (!(vp)->v_object || \
378	  !((vp)->v_object->ref_count || (vp)->v_object->resident_page_count)))
379
380/* Requires interlock */
381#define VMIGHTFREE(vp) \
382	(!((vp)->v_iflag & (VI_FREE|VI_DOOMED|VI_XLOCK)) &&	\
383	 LIST_EMPTY(&(vp)->v_cache_src) && !(vp)->v_usecount)
384
385/* Requires interlock */
386#define	VSHOULDBUSY(vp)	\
387	(((vp)->v_iflag & VI_FREE) && \
388	 ((vp)->v_holdcnt || (vp)->v_usecount))
389
390#define	VI_LOCK(vp)	mtx_lock(&(vp)->v_interlock)
391#define	VI_TRYLOCK(vp)	mtx_trylock(&(vp)->v_interlock)
392#define	VI_UNLOCK(vp)	mtx_unlock(&(vp)->v_interlock)
393#define	VI_MTX(vp)	(&(vp)->v_interlock)
394
395#endif /* _KERNEL */
396
397
398/*
399 * Mods for extensibility.
400 */
401
402/*
403 * Flags for vdesc_flags:
404 */
405#define	VDESC_MAX_VPS		16
406/* Low order 16 flag bits are reserved for willrele flags for vp arguments. */
407#define	VDESC_VP0_WILLRELE	0x0001
408#define	VDESC_VP1_WILLRELE	0x0002
409#define	VDESC_VP2_WILLRELE	0x0004
410#define	VDESC_VP3_WILLRELE	0x0008
411#define	VDESC_NOMAP_VPP		0x0100
412#define	VDESC_VPP_WILLRELE	0x0200
413
414/*
415 * VDESC_NO_OFFSET is used to identify the end of the offset list
416 * and in places where no such field exists.
417 */
418#define VDESC_NO_OFFSET -1
419
420/*
421 * This structure describes the vnode operation taking place.
422 */
423struct vnodeop_desc {
424	int	 vdesc_offset;		/* offset in vector,first for speed */
425	char	*vdesc_name;		/* a readable name for debugging */
426	int	 vdesc_flags;		/* VDESC_* flags */
427
428	/*
429	 * These ops are used by bypass routines to map and locate arguments.
430	 * Creds and procs are not needed in bypass routines, but sometimes
431	 * they are useful to (for example) transport layers.
432	 * Nameidata is useful because it has a cred in it.
433	 */
434	int	*vdesc_vp_offsets;	/* list ended by VDESC_NO_OFFSET */
435	int	vdesc_vpp_offset;	/* return vpp location */
436	int	vdesc_cred_offset;	/* cred location, if any */
437	int	vdesc_thread_offset;	/* thread location, if any */
438	int	vdesc_componentname_offset; /* if any */
439	/*
440	 * Finally, we've got a list of private data (about each operation)
441	 * for each transport layer.  (Support to manage this list is not
442	 * yet part of BSD.)
443	 */
444	caddr_t	*vdesc_transports;
445};
446
447#ifdef _KERNEL
448/*
449 * A list of all the operation descs.
450 */
451extern struct vnodeop_desc *vnodeop_descs[];
452
453/*
454 * Interlock for scanning list of vnodes attached to a mountpoint
455 */
456extern struct mtx mntvnode_mtx;
457
458/*
459 * This macro is very helpful in defining those offsets in the vdesc struct.
460 *
461 * This is stolen from X11R4.  I ignored all the fancy stuff for
462 * Crays, so if you decide to port this to such a serious machine,
463 * you might want to consult Intrinsic.h's XtOffset{,Of,To}.
464 */
465#define	VOPARG_OFFSET(p_type,field) \
466	((int) (((char *) (&(((p_type)NULL)->field))) - ((char *) NULL)))
467#define	VOPARG_OFFSETOF(s_type,field) \
468	VOPARG_OFFSET(s_type*,field)
469#define	VOPARG_OFFSETTO(S_TYPE,S_OFFSET,STRUCT_P) \
470	((S_TYPE)(((char*)(STRUCT_P))+(S_OFFSET)))
471
472
473/*
474 * This structure is used to configure the new vnodeops vector.
475 */
476struct vnodeopv_entry_desc {
477	struct vnodeop_desc *opve_op;   /* which operation this is */
478	vop_t *opve_impl;		/* code implementing this operation */
479};
480struct vnodeopv_desc {
481			/* ptr to the ptr to the vector where op should go */
482	vop_t ***opv_desc_vector_p;
483	struct vnodeopv_entry_desc *opv_desc_ops;   /* null terminated list */
484};
485
486/*
487 * A generic structure.
488 * This can be used by bypass routines to identify generic arguments.
489 */
490struct vop_generic_args {
491	struct vnodeop_desc *a_desc;
492	/* other random data follows, presumably */
493};
494
495
496#ifdef DEBUG_VFS_LOCKS
497/*
498 * Macros to aid in tracing VFS locking problems.  Not totally
499 * reliable since if the thread sleeps between changing the lock
500 * state and checking it with the assert, some other thread could
501 * change the state.  They are good enough for debugging a single
502 * filesystem using a single-threaded test.  I find that 'cvs co src'
503 * is a pretty good test.
504 */
505
506extern int vfs_badlock_panic;
507extern int vfs_badlock_print;
508
509/*
510 * This only exists to supress warnings from unlocked specfs accesses.  It is
511 * no longer ok to have an unlocked VFS.
512 */
513
514#define IGNORE_LOCK(vp)	((vp)->v_type == VCHR || (vp)->v_type == VBAD)
515
516#define ASSERT_VOP_LOCKED(vp, str)					\
517do {									\
518	struct vnode *_vp = (vp);					\
519									\
520	if (_vp && !IGNORE_LOCK(_vp) && !VOP_ISLOCKED(_vp, NULL)) {	\
521		if (vfs_badlock_print)					\
522			printf("%s: %p is not locked but should be\n",	\
523			    str, _vp);					\
524		if (vfs_badlock_panic)					\
525			Debugger("Lock violation.\n");			\
526	}								\
527} while (0)
528
529#define ASSERT_VOP_UNLOCKED(vp, str)					\
530do {									\
531	struct vnode *_vp = (vp);					\
532	int lockstate;							\
533									\
534	if (_vp && !IGNORE_LOCK(_vp)) {					\
535		lockstate = VOP_ISLOCKED(_vp, curthread);		\
536		if (lockstate == LK_EXCLUSIVE) {			\
537			if (vfs_badlock_print)				\
538				printf("%s: %p is locked but should not be\n",	\
539				    str, _vp);				\
540			if (vfs_badlock_panic)				\
541				Debugger("Lock Violation.\n");		\
542		}							\
543	}								\
544} while (0)
545
546#define ASSERT_VOP_ELOCKED(vp, str)					\
547do {									\
548	struct vnode *_vp = (vp);					\
549									\
550	if (_vp && !IGNORE_LOCK(_vp) &&					\
551	    VOP_ISLOCKED(_vp, curthread) != LK_EXCLUSIVE) {		\
552		if (vfs_badlock_print)					\
553			printf("%s: %p is not exclusive locked but should be\n",\
554			    str, _vp);					\
555		if (vfs_badlock_panic)					\
556			Debugger("Lock violation.\n");			\
557	}								\
558} while (0)
559
560#define ASSERT_VOP_ELOCKED_OTHER(vp, str)				\
561do {									\
562	struct vnode *_vp = (vp);					\
563									\
564	if (_vp && !IGNORE_LOCK(_vp) &&					\
565	    VOP_ISLOCKED(_vp, curthread) != LK_EXCLOTHER) {		\
566		if (vfs_badlock_print)					\
567			printf("%s: %p is not exclusive locked by another thread\n",\
568			    str, _vp);					\
569		if (vfs_badlock_panic)					\
570			Debugger("Lock violation.\n");			\
571	}								\
572} while (0)
573
574#define ASSERT_VOP_SLOCKED(vp, str)					\
575do {									\
576	struct vnode *_vp = (vp);					\
577									\
578	if (_vp && !IGNORE_LOCK(_vp) &&					\
579	    VOP_ISLOCKED(_vp, NULL) != LK_SHARED) {			\
580		if (vfs_badlock_print)					\
581			printf("%s: %p is not locked shared but should be",\
582		    str, _vp);						\
583		if (vfs_badlock_panic)					\
584			Debugger("Lock violation.\n");			\
585	}								\
586} while (0)
587
588void vop_rename_pre(void *a);
589void vop_strategy_pre(void *a);
590void vop_lookup_pre(void *a);
591void vop_lookup_post(void *a, int rc);
592
593#else
594
595#define ASSERT_VOP_LOCKED(vp, str)
596#define ASSERT_VOP_UNLOCKED(vp, str)
597
598#endif
599
600/*
601 * VOCALL calls an op given an ops vector.  We break it out because BSD's
602 * vclean changes the ops vector and then wants to call ops with the old
603 * vector.
604 */
605#define VOCALL(OPSV,OFF,AP) (( *((OPSV)[(OFF)])) (AP))
606
607/*
608 * This call works for vnodes in the kernel.
609 */
610#define VCALL(VP,OFF,AP) VOCALL((VP)->v_op,(OFF),(AP))
611#define VDESC(OP) (& __CONCAT(OP,_desc))
612#define VOFFSET(OP) (VDESC(OP)->vdesc_offset)
613
614/*
615 * VMIO support inline
616 */
617
618extern int vmiodirenable;
619
620static __inline int
621vn_canvmio(struct vnode *vp)
622{
623      if (vp && (vp->v_type == VREG || (vmiodirenable && vp->v_type == VDIR)))
624		return(TRUE);
625	return(FALSE);
626}
627
628/*
629 * Finally, include the default set of vnode operations.
630 */
631#include "vnode_if.h"
632
633/*
634 * Public vnode manipulation functions.
635 */
636struct componentname;
637struct file;
638struct mount;
639struct nameidata;
640struct ostat;
641struct thread;
642struct proc;
643struct stat;
644struct nstat;
645struct ucred;
646struct uio;
647struct vattr;
648struct vnode;
649
650extern int	(*lease_check_hook)(struct vop_lease_args *);
651extern int	(*softdep_fsync_hook)(struct vnode *);
652extern int	(*softdep_process_worklist_hook)(struct mount *);
653
654struct	vnode *addaliasu(struct vnode *vp, udev_t nvp_rdev);
655int	bdevvp(dev_t dev, struct vnode **vpp);
656/* cache_* may belong in namei.h. */
657void	cache_enter(struct vnode *dvp, struct vnode *vp,
658	    struct componentname *cnp);
659int	cache_lookup(struct vnode *dvp, struct vnode **vpp,
660	    struct componentname *cnp);
661void	cache_purge(struct vnode *vp);
662void	cache_purgevfs(struct mount *mp);
663int	cache_leaf_test(struct vnode *vp);
664void	cvtstat(struct stat *st, struct ostat *ost);
665void	cvtnstat(struct stat *sb, struct nstat *nsb);
666int	getnewvnode(enum vtagtype tag,
667	    struct mount *mp, vop_t **vops, struct vnode **vpp);
668int	lease_check(struct vop_lease_args *ap);
669int	spec_vnoperate(struct vop_generic_args *);
670int	speedup_syncer(void);
671#define textvp_fullpath(p, rb, rfb) \
672	vn_fullpath(FIRST_THREAD_IN_PROC(p), (p)->p_textvp, rb, rfb)
673int	vn_fullpath(struct thread *td, struct vnode *vn,
674	    char **retbuf, char **freebuf);
675int	vaccess(enum vtype type, mode_t file_mode, uid_t uid, gid_t gid,
676	    mode_t acc_mode, struct ucred *cred, int *privused);
677int	vaccess_acl_posix1e(enum vtype type, uid_t file_uid,
678	    gid_t file_gid, struct acl *acl, mode_t acc_mode,
679	    struct ucred *cred, int *privused);
680void	vattr_null(struct vattr *vap);
681int	vcount(struct vnode *vp);
682void	vdrop(struct vnode *);
683int	vfinddev(dev_t dev, enum vtype type, struct vnode **vpp);
684void	vfs_add_vnodeops(const void *);
685void	vfs_rm_vnodeops(const void *);
686int	vflush(struct mount *mp, int rootrefs, int flags);
687int	vget(struct vnode *vp, int lockflag, struct thread *td);
688void	vgone(struct vnode *vp);
689void	vgonel(struct vnode *vp, struct thread *td);
690void	vhold(struct vnode *);
691int	vinvalbuf(struct vnode *vp, int save, struct ucred *cred,
692	    struct thread *td, int slpflag, int slptimeo);
693int	vtruncbuf(struct vnode *vp, struct ucred *cred, struct thread *td,
694	    off_t length, int blksize);
695void	vprint(char *label, struct vnode *vp);
696int	vrecycle(struct vnode *vp, struct mtx *inter_lkp,
697	    struct thread *td);
698int	vn_close(struct vnode *vp,
699	    int flags, struct ucred *cred, struct thread *td);
700void	vn_finished_write(struct mount *mp);
701int	vn_isdisk(struct vnode *vp, int *errp);
702int	vn_lock(struct vnode *vp, int flags, struct thread *td);
703#ifdef	DEBUG_LOCKS
704int	debug_vn_lock(struct vnode *vp, int flags, struct thread *p,
705	    const char *filename, int line);
706#define vn_lock(vp,flags,p) debug_vn_lock(vp,flags,p,__FILE__,__LINE__)
707#endif
708int	vn_mkdir(char *path, int mode, enum uio_seg segflg, struct thread *td);
709int	vn_open(struct nameidata *ndp, int *flagp, int cmode);
710int	vn_open_cred(struct nameidata *ndp, int *flagp, int cmode,
711	    struct ucred *cred);
712void	vn_pollevent(struct vnode *vp, int events);
713void	vn_pollgone(struct vnode *vp);
714int	vn_pollrecord(struct vnode *vp, struct thread *p, int events);
715int	vn_rdwr(enum uio_rw rw, struct vnode *vp, caddr_t base,
716	    int len, off_t offset, enum uio_seg segflg, int ioflg,
717	    struct ucred *cred, int *aresid, struct thread *td);
718int	vn_rdwr_inchunks(enum uio_rw rw, struct vnode *vp, caddr_t base,
719	    int len, off_t offset, enum uio_seg segflg, int ioflg,
720	    struct ucred *cred, int *aresid, struct thread *td);
721int	vn_stat(struct vnode *vp, struct stat *sb, struct thread *td);
722int	vn_start_write(struct vnode *vp, struct mount **mpp, int flags);
723dev_t	vn_todev(struct vnode *vp);
724int	vn_write_suspend_wait(struct vnode *vp, struct mount *mp,
725	    int flags);
726int	vn_writechk(struct vnode *vp);
727int	vn_extattr_get(struct vnode *vp, int ioflg, int attrnamespace,
728	    const char *attrname, int *buflen, char *buf, struct thread *td);
729int	vn_extattr_set(struct vnode *vp, int ioflg, int attrnamespace,
730	    const char *attrname, int buflen, char *buf, struct thread *td);
731int	vn_extattr_rm(struct vnode *vp, int ioflg, int attrnamespace,
732	    const char *attrname, struct thread *td);
733int	vfs_cache_lookup(struct vop_lookup_args *ap);
734int	vfs_object_create(struct vnode *vp, struct thread *td,
735	    struct ucred *cred);
736void	vfs_timestamp(struct timespec *);
737void	vfs_write_resume(struct mount *mp);
738void	vfs_write_suspend(struct mount *mp);
739int	vop_stdbmap(struct vop_bmap_args *);
740int	vop_stdgetwritemount(struct vop_getwritemount_args *);
741int	vop_stdgetpages(struct vop_getpages_args *);
742int	vop_stdinactive(struct vop_inactive_args *);
743int	vop_stdislocked(struct vop_islocked_args *);
744int	vop_stdlock(struct vop_lock_args *);
745int	vop_stdputpages(struct vop_putpages_args *);
746int	vop_stdunlock(struct vop_unlock_args *);
747int	vop_noislocked(struct vop_islocked_args *);
748int	vop_nolock(struct vop_lock_args *);
749int	vop_nopoll(struct vop_poll_args *);
750int	vop_nounlock(struct vop_unlock_args *);
751int	vop_stdpathconf(struct vop_pathconf_args *);
752int	vop_stdpoll(struct vop_poll_args *);
753int	vop_revoke(struct vop_revoke_args *);
754int	vop_sharedlock(struct vop_lock_args *);
755int	vop_eopnotsupp(struct vop_generic_args *ap);
756int	vop_ebadf(struct vop_generic_args *ap);
757int	vop_einval(struct vop_generic_args *ap);
758int	vop_enotty(struct vop_generic_args *ap);
759int	vop_defaultop(struct vop_generic_args *ap);
760int	vop_null(struct vop_generic_args *ap);
761int	vop_panic(struct vop_generic_args *ap);
762int	vop_stdcreatevobject(struct vop_createvobject_args *ap);
763int	vop_stddestroyvobject(struct vop_destroyvobject_args *ap);
764int	vop_stdgetvobject(struct vop_getvobject_args *ap);
765
766void	vfree(struct vnode *);
767void	vput(struct vnode *vp);
768void	vrele(struct vnode *vp);
769void	vref(struct vnode *vp);
770void	vbusy(struct vnode *vp);
771void 	v_addpollinfo(struct vnode *vp);
772
773extern	vop_t **default_vnodeop_p;
774extern	vop_t **spec_vnodeop_p;
775extern	vop_t **dead_vnodeop_p;
776
777#endif /* _KERNEL */
778
779#endif /* !_SYS_VNODE_H_ */
780