vnode.h revision 101711
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 101711 2002-08-12 01:54:10Z rwatson $
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 *	f - freelist mutex
99 *	i - interlock
100 *	m - mntvnodes mutex
101 *	p - pollinfo lock
102 *	v - vnode lock
103 *
104 * XXX Not all fields are locked yet and some fields that are marked are not
105 * locked consistently.  This is a work in progress.
106 */
107
108struct vnode {
109	struct	mtx v_interlock;		/* lock for "i" things */
110	u_long	v_iflag;			/* i vnode flags (see below) */
111	int	v_usecount;			/* i ref count of users */
112	int	v_writecount;			/* i ref count of writers */
113	long	v_numoutput;			/* i writes in progress */
114	struct thread *v_vxproc;		/* i thread owning VXLOCK */
115	int	v_holdcnt;			/* i page & buffer references */
116	u_long	v_vflag;			/* v vnode flags */
117	u_long	v_id;				/* capability identifier */
118	struct	mount *v_mount;			/* ptr to vfs we are in */
119	vop_t	**v_op;				/* vnode operations vector */
120	TAILQ_ENTRY(vnode) v_freelist;		/* f vnode freelist */
121	TAILQ_ENTRY(vnode) v_nmntvnodes;	/* m vnodes for mount point */
122	struct	buflists v_cleanblkhd;		/* SORTED clean blocklist */
123	struct buf	*v_cleanblkroot;	/* clean buf splay tree root */
124	struct	buflists v_dirtyblkhd;		/* SORTED dirty blocklist */
125	struct buf	*v_dirtyblkroot;	/* dirty buf splay tree root */
126	LIST_ENTRY(vnode) v_synclist;		/* vnodes with dirty buffers */
127	enum	vtype v_type;			/* vnode type */
128	union {
129		struct mount	*vu_mountedhere;/* ptr to mounted vfs (VDIR) */
130		struct socket	*vu_socket;	/* unix ipc (VSOCK) */
131		struct {
132			struct specinfo	*vu_specinfo; /* device (VCHR, VBLK) */
133			SLIST_ENTRY(vnode) vu_specnext;
134		} vu_spec;
135		struct fifoinfo	*vu_fifoinfo;	/* fifo (VFIFO) */
136	} v_un;
137	daddr_t	v_lastw;			/* last write (write cluster) */
138	daddr_t	v_cstart;			/* start block of cluster */
139	daddr_t	v_lasta;			/* last allocation (cluster) */
140	int	v_clen;				/* length of current cluster */
141	struct vm_object *v_object;		/* Place to store VM object */
142	struct	lock v_lock;			/* used if fs don't have one */
143	struct	lock *v_vnlock;			/* pointer to vnode lock */
144	enum	vtagtype v_tag;			/* type of underlying data */
145	void	*v_data;			/* private data for fs */
146	LIST_HEAD(, namecache) v_cache_src;	/* Cache entries from us */
147	TAILQ_HEAD(, namecache) v_cache_dst;	/* Cache entries to us */
148	struct	vnode *v_dd;			/* .. vnode */
149	u_long	v_ddid;				/* .. capability identifier */
150	struct vpollinfo *v_pollinfo;		/* p Poll events */
151	struct label v_label;			/* MAC label for vnode */
152#ifdef	DEBUG_LOCKS
153	const char *filename;			/* Source file doing locking */
154	int line;				/* Line number doing locking */
155#endif
156	udev_t	v_cachedfs;			/* cached fs id */
157	ino_t	v_cachedid;			/* cached file id */
158};
159#define	v_mountedhere	v_un.vu_mountedhere
160#define	v_socket	v_un.vu_socket
161#define	v_rdev		v_un.vu_spec.vu_specinfo
162#define	v_specnext	v_un.vu_spec.vu_specnext
163#define	v_fifoinfo	v_un.vu_fifoinfo
164
165/*
166 * Userland version of struct vnode, for sysctl.
167 */
168struct xvnode {
169	size_t	xv_size;			/* sizeof(struct xvnode) */
170	void	*xv_vnode;			/* address of real vnode */
171	u_long	xv_flag;			/* vnode vflags */
172	int	xv_usecount;			/* reference count of users */
173	int	xv_writecount;			/* reference count of writers */
174	int	xv_holdcnt;			/* page & buffer references */
175	u_long	xv_id;				/* capability identifier */
176	void	*xv_mount;			/* address of parent mount */
177	long	xv_numoutput;			/* num of writes in progress */
178	enum	vtype xv_type;			/* vnode type */
179	union {
180		void	*xvu_socket;		/* socket, if VSOCK */
181		void	*xvu_fifo;		/* fifo, if VFIFO */
182		udev_t	xvu_rdev;		/* maj/min, if VBLK/VCHR */
183		struct {
184			udev_t	xvu_dev;	/* device, if VDIR/VREG/VLNK */
185			ino_t	xvu_ino;	/* id, if VDIR/VREG/VLNK */
186		};
187	} xv_un;
188};
189#define xv_socket	xv_un.xvu_socket
190#define xv_fifo		xv_un.xvu_fifo
191#define xv_rdev		xv_un.xvu_rdev
192#define xv_dev		xv_un.xvu_dev
193#define xv_ino		xv_un.xvu_ino
194
195#define	VN_POLLEVENT(vp, events)				\
196	do {							\
197		if ((vp)->v_pollinfo != NULL && 		\
198		    (vp)->v_pollinfo->vpi_events & (events))	\
199			vn_pollevent((vp), (events));		\
200	} while (0)
201
202#define VN_KNOTE(vp, b)						\
203	do {							\
204		if ((vp)->v_pollinfo != NULL)			\
205			KNOTE(&vp->v_pollinfo->vpi_selinfo.si_note, (b)); \
206	} while (0)
207
208/*
209 * Vnode flags.
210 *	VI flags are protected by interlock and live in v_iflag
211 *	VV flags are protected by the vnode lock and live in v_vflag
212 */
213#define	VI_XLOCK	0x0001	/* vnode is locked to change vtype */
214#define	VI_XWANT	0x0002	/* thread is waiting for vnode */
215#define	VI_BWAIT	0x0004	/* waiting for output to complete */
216#define	VI_OLOCK	0x0008	/* vnode is locked waiting for an object */
217#define	VI_OWANT	0x0010	/* a thread is waiting for VOLOCK */
218#define	VI_MOUNT	0x0020	/* Mount in progress */
219#define	VI_AGE		0x0040	/* Insert vnode at head of free list */
220#define	VI_DOOMED	0x0080	/* This vnode is being recycled */
221#define	VI_FREE		0x0100	/* This vnode is on the freelist */
222#define	VI_OBJDIRTY	0x0400	/* object might be dirty */
223/*
224 * XXX VI_ONWORKLST could be replaced with a check for NULL list elements
225 * in v_synclist.
226 */
227#define	VI_ONWORKLST	0x0200	/* On syncer work-list */
228
229#define	VV_ROOT		0x0001	/* root of its filesystem */
230#define	VV_ISTTY	0x0002	/* vnode represents a tty */
231#define	VV_NOSYNC	0x0004	/* unlinked, stop syncing */
232#define	VV_OBJBUF	0x0008	/* Allocate buffers in VM object */
233#define	VV_CACHEDLABEL	0x0010	/* Vnode has valid cached MAC label */
234#define	VV_TEXT		0x0020	/* vnode is a pure text prototype */
235#define	VV_COPYONWRITE	0x0040	/* vnode is doing copy-on-write */
236#define	VV_SYSTEM	0x0080	/* vnode being used by kernel */
237
238/*
239 * Vnode attributes.  A field value of VNOVAL represents a field whose value
240 * is unavailable (getattr) or which is not to be changed (setattr).
241 */
242struct vattr {
243	enum vtype	va_type;	/* vnode type (for create) */
244	u_short		va_mode;	/* files access mode and type */
245	short		va_nlink;	/* number of references to file */
246	uid_t		va_uid;		/* owner user id */
247	gid_t		va_gid;		/* owner group id */
248	udev_t		va_fsid;	/* filesystem id */
249	long		va_fileid;	/* file id */
250	u_quad_t	va_size;	/* file size in bytes */
251	long		va_blocksize;	/* blocksize preferred for i/o */
252	struct timespec	va_atime;	/* time of last access */
253	struct timespec	va_mtime;	/* time of last modification */
254	struct timespec	va_ctime;	/* time file changed */
255	struct timespec	va_birthtime;	/* time file created */
256	u_long		va_gen;		/* generation number of file */
257	u_long		va_flags;	/* flags defined for file */
258	udev_t		va_rdev;	/* device the special file represents */
259	u_quad_t	va_bytes;	/* bytes of disk space held by file */
260	u_quad_t	va_filerev;	/* file modification number */
261	u_int		va_vaflags;	/* operations flags, see below */
262	long		va_spare;	/* remain quad aligned */
263};
264
265/*
266 * Flags for va_vaflags.
267 */
268#define	VA_UTIMES_NULL	0x01		/* utimes argument was NULL */
269#define	VA_EXCLUSIVE	0x02		/* exclusive create request */
270
271/*
272 * Flags for ioflag. (high 16 bits used to ask for read-ahead and
273 * help with write clustering)
274 */
275#define	IO_UNIT		0x0001		/* do I/O as atomic unit */
276#define	IO_APPEND	0x0002		/* append write to end */
277#define	IO_SYNC		0x0004		/* do I/O synchronously */
278#define	IO_NODELOCKED	0x0008		/* underlying node already locked */
279#define	IO_NDELAY	0x0010		/* FNDELAY flag set in file table */
280#define	IO_VMIO		0x0020		/* data already in VMIO space */
281#define	IO_INVAL	0x0040		/* invalidate after I/O */
282#define	IO_ASYNC	0x0080		/* bawrite rather then bdwrite */
283#define	IO_DIRECT	0x0100		/* attempt to bypass buffer cache */
284#define	IO_NOWDRAIN	0x0200		/* do not block on wdrain */
285#define	IO_EXT		0x0400		/* operate on external attributes */
286#define	IO_NORMAL	0x0800		/* operate on regular data */
287#define	IO_NOMACCHECK	0x1000		/* MAC checks unnecessary */
288
289/*
290 *  Modes.  Some values same as Ixxx entries from inode.h for now.
291 */
292#define	VEXEC	000100		/* execute/search permission */
293#define	VWRITE	000200		/* write permission */
294#define	VREAD	000400		/* read permission */
295#define	VSVTX	001000		/* save swapped text even after use */
296#define	VSGID	002000		/* set group id on execution */
297#define	VSUID	004000		/* set user id on execution */
298#define	VADMIN	010000		/* permission to administer */
299#define	VSTAT	020000		/* permission to retrieve attrs */
300#define	VAPPEND	040000		/* permission to write/append */
301#define	VALLPERM	(VEXEC | VWRITE | VREAD | VADMIN | VSTAT | VAPPEND)
302
303/*
304 * Token indicating no attribute value yet assigned.
305 */
306#define	VNOVAL	(-1)
307
308/*
309 * LK_TIMELOCK timeout for vnode locks (used mainly by the pageout daemon)
310 */
311#define VLKTIMEOUT	(hz / 20 + 1)
312
313#ifdef _KERNEL
314
315#ifdef MALLOC_DECLARE
316MALLOC_DECLARE(M_VNODE);
317#endif
318
319/*
320 * Convert between vnode types and inode formats (since POSIX.1
321 * defines mode word of stat structure in terms of inode formats).
322 */
323extern enum vtype	iftovt_tab[];
324extern int		vttoif_tab[];
325#define	IFTOVT(mode)	(iftovt_tab[((mode) & S_IFMT) >> 12])
326#define	VTTOIF(indx)	(vttoif_tab[(int)(indx)])
327#define	MAKEIMODE(indx, mode)	(int)(VTTOIF(indx) | (mode))
328
329/*
330 * Flags to various vnode functions.
331 */
332#define	SKIPSYSTEM	0x0001	/* vflush: skip vnodes marked VSYSTEM */
333#define	FORCECLOSE	0x0002	/* vflush: force file closure */
334#define	WRITECLOSE	0x0004	/* vflush: only close writable files */
335#define	DOCLOSE		0x0008	/* vclean: close active files */
336#define	V_SAVE		0x0001	/* vinvalbuf: sync file first */
337#define	V_ALT		0x0002	/* vinvalbuf: invalidate only alternate bufs */
338#define	V_NORMAL	0x0004	/* vinvalbuf: invalidate only regular bufs */
339#define	REVOKEALL	0x0001	/* vop_revoke: revoke all aliases */
340#define	V_WAIT		0x0001	/* vn_start_write: sleep for suspend */
341#define	V_NOWAIT	0x0002	/* vn_start_write: don't sleep for suspend */
342#define	V_XSLEEP	0x0004	/* vn_start_write: just return after sleep */
343
344#define	VREF(vp)	vref(vp)
345
346
347#ifdef DIAGNOSTIC
348#define	VATTR_NULL(vap)	vattr_null(vap)
349#else
350#define	VATTR_NULL(vap)	(*(vap) = va_null)	/* initialize a vattr */
351#endif /* DIAGNOSTIC */
352
353#define	NULLVP	((struct vnode *)NULL)
354
355#define	VNODEOP_SET(f) \
356	C_SYSINIT(f##init, SI_SUB_VFS, SI_ORDER_SECOND, vfs_add_vnodeops, &f); \
357	C_SYSUNINIT(f##uninit, SI_SUB_VFS, SI_ORDER_SECOND, vfs_rm_vnodeops, &f);
358
359/*
360 * Global vnode data.
361 */
362extern	struct vnode *rootvnode;	/* root (i.e. "/") vnode */
363extern	int desiredvnodes;		/* number of vnodes desired */
364extern	uma_zone_t namei_zone;
365extern	int prtactive;			/* nonzero to call vprint() */
366extern	struct vattr va_null;		/* predefined null vattr structure */
367extern	int vfs_ioopt;
368
369/*
370 * Macro/function to check for client cache inconsistency w.r.t. leasing.
371 */
372#define	LEASE_READ	0x1		/* Check lease for readers */
373#define	LEASE_WRITE	0x2		/* Check lease for modifiers */
374
375
376extern void	(*lease_updatetime)(int deltat);
377
378/* Requires interlock */
379#define	VSHOULDFREE(vp)	\
380	(!((vp)->v_iflag & (VI_FREE|VI_DOOMED)) && \
381	 !(vp)->v_holdcnt && !(vp)->v_usecount && \
382	 (!(vp)->v_object || \
383	  !((vp)->v_object->ref_count || (vp)->v_object->resident_page_count)))
384
385/* Requires interlock */
386#define VMIGHTFREE(vp) \
387	(!((vp)->v_iflag & (VI_FREE|VI_DOOMED|VI_XLOCK)) &&	\
388	 LIST_EMPTY(&(vp)->v_cache_src) && !(vp)->v_usecount)
389
390/* Requires interlock */
391#define	VSHOULDBUSY(vp)	\
392	(((vp)->v_iflag & VI_FREE) && \
393	 ((vp)->v_holdcnt || (vp)->v_usecount))
394
395#define	VI_LOCK(vp)	mtx_lock(&(vp)->v_interlock)
396#define	VI_TRYLOCK(vp)	mtx_trylock(&(vp)->v_interlock)
397#define	VI_UNLOCK(vp)	mtx_unlock(&(vp)->v_interlock)
398#define	VI_MTX(vp)	(&(vp)->v_interlock)
399
400#endif /* _KERNEL */
401
402
403/*
404 * Mods for extensibility.
405 */
406
407/*
408 * Flags for vdesc_flags:
409 */
410#define	VDESC_MAX_VPS		16
411/* Low order 16 flag bits are reserved for willrele flags for vp arguments. */
412#define	VDESC_VP0_WILLRELE	0x0001
413#define	VDESC_VP1_WILLRELE	0x0002
414#define	VDESC_VP2_WILLRELE	0x0004
415#define	VDESC_VP3_WILLRELE	0x0008
416#define	VDESC_NOMAP_VPP		0x0100
417#define	VDESC_VPP_WILLRELE	0x0200
418
419/*
420 * VDESC_NO_OFFSET is used to identify the end of the offset list
421 * and in places where no such field exists.
422 */
423#define VDESC_NO_OFFSET -1
424
425/*
426 * This structure describes the vnode operation taking place.
427 */
428struct vnodeop_desc {
429	int	 vdesc_offset;		/* offset in vector,first for speed */
430	char	*vdesc_name;		/* a readable name for debugging */
431	int	 vdesc_flags;		/* VDESC_* flags */
432
433	/*
434	 * These ops are used by bypass routines to map and locate arguments.
435	 * Creds and procs are not needed in bypass routines, but sometimes
436	 * they are useful to (for example) transport layers.
437	 * Nameidata is useful because it has a cred in it.
438	 */
439	int	*vdesc_vp_offsets;	/* list ended by VDESC_NO_OFFSET */
440	int	vdesc_vpp_offset;	/* return vpp location */
441	int	vdesc_cred_offset;	/* cred location, if any */
442	int	vdesc_thread_offset;	/* thread location, if any */
443	int	vdesc_componentname_offset; /* if any */
444	/*
445	 * Finally, we've got a list of private data (about each operation)
446	 * for each transport layer.  (Support to manage this list is not
447	 * yet part of BSD.)
448	 */
449	caddr_t	*vdesc_transports;
450};
451
452#ifdef _KERNEL
453/*
454 * A list of all the operation descs.
455 */
456extern struct vnodeop_desc *vnodeop_descs[];
457
458/*
459 * Interlock for scanning list of vnodes attached to a mountpoint
460 */
461extern struct mtx mntvnode_mtx;
462
463/*
464 * This macro is very helpful in defining those offsets in the vdesc struct.
465 *
466 * This is stolen from X11R4.  I ignored all the fancy stuff for
467 * Crays, so if you decide to port this to such a serious machine,
468 * you might want to consult Intrinsic.h's XtOffset{,Of,To}.
469 */
470#define	VOPARG_OFFSET(p_type,field) \
471	((int) (((char *) (&(((p_type)NULL)->field))) - ((char *) NULL)))
472#define	VOPARG_OFFSETOF(s_type,field) \
473	VOPARG_OFFSET(s_type*,field)
474#define	VOPARG_OFFSETTO(S_TYPE,S_OFFSET,STRUCT_P) \
475	((S_TYPE)(((char*)(STRUCT_P))+(S_OFFSET)))
476
477
478/*
479 * This structure is used to configure the new vnodeops vector.
480 */
481struct vnodeopv_entry_desc {
482	struct vnodeop_desc *opve_op;   /* which operation this is */
483	vop_t *opve_impl;		/* code implementing this operation */
484};
485struct vnodeopv_desc {
486			/* ptr to the ptr to the vector where op should go */
487	vop_t ***opv_desc_vector_p;
488	struct vnodeopv_entry_desc *opv_desc_ops;   /* null terminated list */
489};
490
491/*
492 * A generic structure.
493 * This can be used by bypass routines to identify generic arguments.
494 */
495struct vop_generic_args {
496	struct vnodeop_desc *a_desc;
497	/* other random data follows, presumably */
498};
499
500
501#ifdef DEBUG_VFS_LOCKS
502/*
503 * Macros to aid in tracing VFS locking problems.  Not totally
504 * reliable since if the thread sleeps between changing the lock
505 * state and checking it with the assert, some other thread could
506 * change the state.  They are good enough for debugging a single
507 * filesystem using a single-threaded test.  I find that 'cvs co src'
508 * is a pretty good test.
509 */
510
511extern int vfs_badlock_panic;
512extern int vfs_badlock_print;
513
514/*
515 * This only exists to supress warnings from unlocked specfs accesses.  It is
516 * no longer ok to have an unlocked VFS.
517 */
518
519#define IGNORE_LOCK(vp)	((vp)->v_type == VCHR || (vp)->v_type == VBAD)
520
521#define ASSERT_VOP_LOCKED(vp, str)					\
522do {									\
523	struct vnode *_vp = (vp);					\
524									\
525	if (_vp && !IGNORE_LOCK(_vp) && !VOP_ISLOCKED(_vp, NULL)) {	\
526		if (vfs_badlock_print)					\
527			printf("%s: %p is not locked but should be\n",	\
528			    str, _vp);					\
529		if (vfs_badlock_panic)					\
530			Debugger("Lock violation.\n");			\
531	}								\
532} while (0)
533
534#define ASSERT_VOP_UNLOCKED(vp, str)					\
535do {									\
536	struct vnode *_vp = (vp);					\
537	int lockstate;							\
538									\
539	if (_vp && !IGNORE_LOCK(_vp)) {					\
540		lockstate = VOP_ISLOCKED(_vp, curthread);		\
541		if (lockstate == LK_EXCLUSIVE) {			\
542			if (vfs_badlock_print)				\
543				printf("%s: %p is locked but should not be\n",	\
544				    str, _vp);				\
545			if (vfs_badlock_panic)				\
546				Debugger("Lock Violation.\n");		\
547		}							\
548	}								\
549} while (0)
550
551#define ASSERT_VOP_ELOCKED(vp, str)					\
552do {									\
553	struct vnode *_vp = (vp);					\
554									\
555	if (_vp && !IGNORE_LOCK(_vp) &&					\
556	    VOP_ISLOCKED(_vp, curthread) != LK_EXCLUSIVE) {		\
557		if (vfs_badlock_print)					\
558			printf("%s: %p is not exclusive locked but should be\n",\
559			    str, _vp);					\
560		if (vfs_badlock_panic)					\
561			Debugger("Lock violation.\n");			\
562	}								\
563} while (0)
564
565#define ASSERT_VOP_ELOCKED_OTHER(vp, str)				\
566do {									\
567	struct vnode *_vp = (vp);					\
568									\
569	if (_vp && !IGNORE_LOCK(_vp) &&					\
570	    VOP_ISLOCKED(_vp, curthread) != LK_EXCLOTHER) {		\
571		if (vfs_badlock_print)					\
572			printf("%s: %p is not exclusive locked by another thread\n",\
573			    str, _vp);					\
574		if (vfs_badlock_panic)					\
575			Debugger("Lock violation.\n");			\
576	}								\
577} while (0)
578
579#define ASSERT_VOP_SLOCKED(vp, str)					\
580do {									\
581	struct vnode *_vp = (vp);					\
582									\
583	if (_vp && !IGNORE_LOCK(_vp) &&					\
584	    VOP_ISLOCKED(_vp, NULL) != LK_SHARED) {			\
585		if (vfs_badlock_print)					\
586			printf("%s: %p is not locked shared but should be",\
587		    str, _vp);						\
588		if (vfs_badlock_panic)					\
589			Debugger("Lock violation.\n");			\
590	}								\
591} while (0)
592
593void vop_rename_pre(void *a);
594void vop_strategy_pre(void *a);
595void vop_lookup_pre(void *a);
596void vop_lookup_post(void *a, int rc);
597
598#else
599
600#define ASSERT_VOP_LOCKED(vp, str)
601#define ASSERT_VOP_UNLOCKED(vp, str)
602
603#endif
604
605/*
606 * VOCALL calls an op given an ops vector.  We break it out because BSD's
607 * vclean changes the ops vector and then wants to call ops with the old
608 * vector.
609 */
610#define VOCALL(OPSV,OFF,AP) (( *((OPSV)[(OFF)])) (AP))
611
612/*
613 * This call works for vnodes in the kernel.
614 */
615#define VCALL(VP,OFF,AP) VOCALL((VP)->v_op,(OFF),(AP))
616#define VDESC(OP) (& __CONCAT(OP,_desc))
617#define VOFFSET(OP) (VDESC(OP)->vdesc_offset)
618
619/*
620 * VMIO support inline
621 */
622
623extern int vmiodirenable;
624
625static __inline int
626vn_canvmio(struct vnode *vp)
627{
628      if (vp && (vp->v_type == VREG || (vmiodirenable && vp->v_type == VDIR)))
629		return(TRUE);
630	return(FALSE);
631}
632
633/*
634 * Finally, include the default set of vnode operations.
635 */
636#include "vnode_if.h"
637
638/*
639 * Public vnode manipulation functions.
640 */
641struct componentname;
642struct file;
643struct mount;
644struct nameidata;
645struct ostat;
646struct thread;
647struct proc;
648struct stat;
649struct nstat;
650struct ucred;
651struct uio;
652struct vattr;
653struct vnode;
654
655extern int	(*lease_check_hook)(struct vop_lease_args *);
656extern int	(*softdep_fsync_hook)(struct vnode *);
657extern int	(*softdep_process_worklist_hook)(struct mount *);
658
659struct	vnode *addaliasu(struct vnode *vp, udev_t nvp_rdev);
660int	bdevvp(dev_t dev, struct vnode **vpp);
661/* cache_* may belong in namei.h. */
662void	cache_enter(struct vnode *dvp, struct vnode *vp,
663	    struct componentname *cnp);
664int	cache_lookup(struct vnode *dvp, struct vnode **vpp,
665	    struct componentname *cnp);
666void	cache_purge(struct vnode *vp);
667void	cache_purgevfs(struct mount *mp);
668int	cache_leaf_test(struct vnode *vp);
669void	cvtstat(struct stat *st, struct ostat *ost);
670void	cvtnstat(struct stat *sb, struct nstat *nsb);
671int	getnewvnode(enum vtagtype tag,
672	    struct mount *mp, vop_t **vops, struct vnode **vpp);
673int	lease_check(struct vop_lease_args *ap);
674int	spec_vnoperate(struct vop_generic_args *);
675int	speedup_syncer(void);
676#define textvp_fullpath(p, rb, rfb) \
677	vn_fullpath(FIRST_THREAD_IN_PROC(p), (p)->p_textvp, rb, rfb)
678int	vn_fullpath(struct thread *td, struct vnode *vn,
679	    char **retbuf, char **freebuf);
680int	vaccess(enum vtype type, mode_t file_mode, uid_t uid, gid_t gid,
681	    mode_t acc_mode, struct ucred *cred, int *privused);
682int	vaccess_acl_posix1e(enum vtype type, uid_t file_uid,
683	    gid_t file_gid, struct acl *acl, mode_t acc_mode,
684	    struct ucred *cred, int *privused);
685void	vattr_null(struct vattr *vap);
686int	vcount(struct vnode *vp);
687void	vdrop(struct vnode *);
688int	vfinddev(dev_t dev, enum vtype type, struct vnode **vpp);
689void	vfs_add_vnodeops(const void *);
690void	vfs_rm_vnodeops(const void *);
691int	vflush(struct mount *mp, int rootrefs, int flags);
692int	vget(struct vnode *vp, int lockflag, struct thread *td);
693void	vgone(struct vnode *vp);
694void	vgonel(struct vnode *vp, struct thread *td);
695void	vhold(struct vnode *);
696int	vinvalbuf(struct vnode *vp, int save, struct ucred *cred,
697	    struct thread *td, int slpflag, int slptimeo);
698int	vtruncbuf(struct vnode *vp, struct ucred *cred, struct thread *td,
699	    off_t length, int blksize);
700void	vprint(char *label, struct vnode *vp);
701int	vrecycle(struct vnode *vp, struct mtx *inter_lkp,
702	    struct thread *td);
703int	vn_close(struct vnode *vp,
704	    int flags, struct ucred *cred, struct thread *td);
705void	vn_finished_write(struct mount *mp);
706int	vn_isdisk(struct vnode *vp, int *errp);
707int	vn_lock(struct vnode *vp, int flags, struct thread *td);
708#ifdef	DEBUG_LOCKS
709int	debug_vn_lock(struct vnode *vp, int flags, struct thread *p,
710	    const char *filename, int line);
711#define vn_lock(vp,flags,p) debug_vn_lock(vp,flags,p,__FILE__,__LINE__)
712#endif
713int	vn_mkdir(char *path, int mode, enum uio_seg segflg, struct thread *td);
714int	vn_open(struct nameidata *ndp, int *flagp, int cmode);
715int	vn_open_cred(struct nameidata *ndp, int *flagp, int cmode,
716	    struct ucred *cred);
717void	vn_pollevent(struct vnode *vp, int events);
718void	vn_pollgone(struct vnode *vp);
719int	vn_pollrecord(struct vnode *vp, struct thread *p, int events);
720int	vn_rdwr(enum uio_rw rw, struct vnode *vp, caddr_t base,
721	    int len, off_t offset, enum uio_seg segflg, int ioflg,
722	    struct ucred *cred, int *aresid, struct thread *td);
723int	vn_rdwr_inchunks(enum uio_rw rw, struct vnode *vp, caddr_t base,
724	    int len, off_t offset, enum uio_seg segflg, int ioflg,
725	    struct ucred *cred, int *aresid, struct thread *td);
726int	vn_stat(struct vnode *vp, struct stat *sb, struct thread *td);
727int	vn_start_write(struct vnode *vp, struct mount **mpp, int flags);
728dev_t	vn_todev(struct vnode *vp);
729int	vn_write_suspend_wait(struct vnode *vp, struct mount *mp,
730	    int flags);
731int	vn_writechk(struct vnode *vp);
732int	vn_extattr_get(struct vnode *vp, int ioflg, int attrnamespace,
733	    const char *attrname, int *buflen, char *buf, struct thread *td);
734int	vn_extattr_set(struct vnode *vp, int ioflg, int attrnamespace,
735	    const char *attrname, int buflen, char *buf, struct thread *td);
736int	vn_extattr_rm(struct vnode *vp, int ioflg, int attrnamespace,
737	    const char *attrname, struct thread *td);
738int	vfs_cache_lookup(struct vop_lookup_args *ap);
739int	vfs_object_create(struct vnode *vp, struct thread *td,
740	    struct ucred *cred);
741void	vfs_timestamp(struct timespec *);
742void	vfs_write_resume(struct mount *mp);
743void	vfs_write_suspend(struct mount *mp);
744int	vop_stdbmap(struct vop_bmap_args *);
745int	vop_stdgetwritemount(struct vop_getwritemount_args *);
746int	vop_stdgetpages(struct vop_getpages_args *);
747int	vop_stdinactive(struct vop_inactive_args *);
748int	vop_stdislocked(struct vop_islocked_args *);
749int	vop_stdlock(struct vop_lock_args *);
750int	vop_stdputpages(struct vop_putpages_args *);
751int	vop_stdunlock(struct vop_unlock_args *);
752int	vop_noislocked(struct vop_islocked_args *);
753int	vop_nolock(struct vop_lock_args *);
754int	vop_nopoll(struct vop_poll_args *);
755int	vop_nounlock(struct vop_unlock_args *);
756int	vop_stdpathconf(struct vop_pathconf_args *);
757int	vop_stdpoll(struct vop_poll_args *);
758int	vop_revoke(struct vop_revoke_args *);
759int	vop_sharedlock(struct vop_lock_args *);
760int	vop_eopnotsupp(struct vop_generic_args *ap);
761int	vop_ebadf(struct vop_generic_args *ap);
762int	vop_einval(struct vop_generic_args *ap);
763int	vop_enotty(struct vop_generic_args *ap);
764int	vop_defaultop(struct vop_generic_args *ap);
765int	vop_null(struct vop_generic_args *ap);
766int	vop_panic(struct vop_generic_args *ap);
767int	vop_stdcreatevobject(struct vop_createvobject_args *ap);
768int	vop_stddestroyvobject(struct vop_destroyvobject_args *ap);
769int	vop_stdgetvobject(struct vop_getvobject_args *ap);
770
771void	vfree(struct vnode *);
772void	vput(struct vnode *vp);
773void	vrele(struct vnode *vp);
774void	vref(struct vnode *vp);
775void	vbusy(struct vnode *vp);
776void 	v_addpollinfo(struct vnode *vp);
777
778extern	vop_t **default_vnodeop_p;
779extern	vop_t **spec_vnodeop_p;
780extern	vop_t **dead_vnodeop_p;
781
782#endif /* _KERNEL */
783
784#endif /* !_SYS_VNODE_H_ */
785