1 /* $OpenBSD: vnode.h,v 1.61 2005/05/26 00:33:45 pedro Exp $ */
2 /* $NetBSD: vnode.h,v 1.38 1996/02/29 20:59:05 cgd Exp $ */
3
4 /*
5 * Copyright (c) 1989, 1993
6 * The Regents of the University of California. All rights reserved.
7 *
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
10 * are met:
11 * 1. Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
13 * 2. Redistributions in binary form must reproduce the above copyright
14 * notice, this list of conditions and the following disclaimer in the
15 * documentation and/or other materials provided with the distribution.
16 * 3. Neither the name of the University nor the names of its contributors
17 * may be used to endorse or promote products derived from this software
18 * without specific prior written permission.
19 *
20 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
21 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
24 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
30 * SUCH DAMAGE.
31 *
32 * @(#)vnode.h 8.11 (Berkeley) 11/21/94
33 */
34
35 #include <sys/types.h>
36 #include <sys/queue.h>
37 #include <sys/lock.h>
38 #include <sys/select.h>
39
40 /* XXX: clean up includes later */
41 #include <uvm/uvm_pglist.h> /* XXX */
42 #include <sys/lock.h> /* XXX */
43 #include <uvm/uvm.h> /* XXX */
44 #include <uvm/uvm_vnode.h> /* XXX */
45
46 /*
47 * The vnode is the focus of all file activity in UNIX. There is a
48 * unique vnode allocated for each active file, each current directory,
49 * each mounted-on file, text file, and the root.
50 */
51
52 /*
53 * Vnode types. VNON means no type.
54 */
55 enum vtype { VNON, VREG, VDIR, VBLK, VCHR, VLNK, VSOCK, VFIFO, VBAD };
56
57 /*
58 * Vnode tag types.
59 * These are for the benefit of external programs only (e.g., pstat)
60 * and should NEVER be inspected by the kernel.
61 *
62 * Note that v_tag is actually used to tell MFS from FFS, and EXT2FS from
63 * the rest, so don't believe the above comment!
64 */
65 enum vtagtype {
66 VT_NON, VT_UFS, VT_NFS, VT_MFS, VT_MSDOSFS, VT_LFS, VT_LOFS, VT_FDESC,
67 VT_PORTAL, VT_KERNFS, VT_PROCFS, VT_AFS, VT_ISOFS, VT_ADOSFS, VT_EXT2FS,
68 VT_NCPFS, VT_VFS, VT_XFS, VT_NTFS, VT_UDF
69 };
70
71 /*
72 * Each underlying filesystem allocates its own private area and hangs
73 * it from v_data. If non-null, this area is freed in getnewvnode().
74 */
75 LIST_HEAD(buflists, buf);
76
77 /*
78 * Reading or writing any of these items requires holding the appropriate lock.
79 * v_freelist is locked by the global vnode_free_list simple lock.
80 * v_mntvnodes is locked by the global mntvnodes simple lock.
81 * v_flag, v_usecount, v_holdcount and v_writecount are
82 * locked by the v_interlock simple lock.
83 */
84
85 struct vnode {
86 struct uvm_vnode v_uvm; /* uvm data */
87 int (**v_op)(void *); /* vnode operations vector */
88 enum vtype v_type; /* vnode type */
89 u_int v_flag; /* vnode flags (see below) */
90 u_int v_usecount; /* reference count of users */
91 /* reference count of writers */
92 u_int v_writecount;
93 /* Flags that can be read/written in interrupts */
94 u_int v_bioflag;
95 u_int v_holdcnt; /* buffer references */
96 u_int v_id; /* capability identifier */
97 struct mount *v_mount; /* ptr to vfs we are in */
98 TAILQ_ENTRY(vnode) v_freelist; /* vnode freelist */
99 LIST_ENTRY(vnode) v_mntvnodes; /* vnodes for mount point */
100 struct buflists v_cleanblkhd; /* clean blocklist head */
101 struct buflists v_dirtyblkhd; /* dirty blocklist head */
102 u_int v_numoutput; /* num of writes in progress */
103 LIST_ENTRY(vnode) v_synclist; /* vnode with dirty buffers */
104 union {
105 struct mount *vu_mountedhere;/* ptr to mounted vfs (VDIR) */
106 struct socket *vu_socket; /* unix ipc (VSOCK) */
107 struct specinfo *vu_specinfo; /* device (VCHR, VBLK) */
108 struct fifoinfo *vu_fifoinfo; /* fifo (VFIFO) */
109 } v_un;
110
111 struct simplelock v_interlock; /* lock on usecount and flag */
112 struct lock *v_vnlock; /* used for non-locking fs's */
113 enum vtagtype v_tag; /* type of underlying data */
114 void *v_data; /* private data for fs */
115 struct {
116 struct simplelock vsi_lock; /* lock to protect below */
117 struct selinfo vsi_selinfo; /* identity of poller(s) */
118 } v_selectinfo;
119 };
120 #define v_mountedhere v_un.vu_mountedhere
121 #define v_socket v_un.vu_socket
122 #define v_specinfo v_un.vu_specinfo
123 #define v_fifoinfo v_un.vu_fifoinfo
124
125 /*
126 * Vnode flags.
127 */
128 #define VROOT 0x0001 /* root of its file system */
129 #define VTEXT 0x0002 /* vnode is a pure text prototype */
130 #define VSYSTEM 0x0004 /* vnode being used by kernel */
131 #define VISTTY 0x0008 /* vnode represents a tty */
132 #define VXLOCK 0x0100 /* vnode is locked to change underlying type */
133 #define VXWANT 0x0200 /* process is waiting for vnode */
134 #define VALIASED 0x0800 /* vnode has an alias */
135 #define VLOCKSWORK 0x4000 /* FS supports locking discipline */
136
137 /*
138 * (v_bioflag) Flags that may be manipulated by interrupt handlers
139 */
140 #define VBIOWAIT 0x0001 /* waiting for output to complete */
141 #define VBIOONSYNCLIST 0x0002 /* Vnode is on syncer worklist */
142 #define VBIOONFREELIST 0x0004 /* Vnode is on a free list */
143
144 /*
145 * Vnode attributes. A field value of VNOVAL represents a field whose value
146 * is unavailable (getattr) or which is not to be changed (setattr).
147 */
148 struct vattr {
149 enum vtype va_type; /* vnode type (for create) */
150 mode_t va_mode; /* files access mode and type */
151 nlink_t va_nlink; /* number of references to file */
152 uid_t va_uid; /* owner user id */
153 gid_t va_gid; /* owner group id */
154 long va_fsid; /* file system id (dev for now) */
155 long va_fileid; /* file id */
156 u_quad_t va_size; /* file size in bytes */
157 long va_blocksize; /* blocksize preferred for i/o */
158 struct timespec va_atime; /* time of last access */
159 struct timespec va_mtime; /* time of last modification */
160 struct timespec va_ctime; /* time file changed */
161 u_long va_gen; /* generation number of file */
162 u_long va_flags; /* flags defined for file */
163 dev_t va_rdev; /* device the special file represents */
164 u_quad_t va_bytes; /* bytes of disk space held by file */
165 u_quad_t va_filerev; /* file modification number */
166 u_int va_vaflags; /* operations flags, see below */
167 long va_spare; /* remain quad aligned */
168 };
169
170 /*
171 * Flags for va_cflags.
172 */
173 #define VA_UTIMES_NULL 0x01 /* utimes argument was NULL */
174 #define VA_EXCLUSIVE 0x02 /* exclusive create request */
175 /*
176 * Flags for ioflag.
177 */
178 #define IO_UNIT 0x01 /* do I/O as atomic unit */
179 #define IO_APPEND 0x02 /* append write to end */
180 #define IO_SYNC 0x04 /* do I/O synchronously */
181 #define IO_NODELOCKED 0x08 /* underlying node already locked */
182 #define IO_NDELAY 0x10 /* FNDELAY flag set in file table */
183 #define IO_NOLIMIT 0x20 /* don't enforce limits on i/o */
184
185 /*
186 * Modes. Some values same as Ixxx entries from inode.h for now.
187 */
188 #define VSUID 04000 /* set user id on execution */
189 #define VSGID 02000 /* set group id on execution */
190 #define VSVTX 01000 /* save swapped text even after use */
191 #define VREAD 00400 /* read, write, execute permissions */
192 #define VWRITE 00200
193 #define VEXEC 00100
194
195 /*
196 * Token indicating no attribute value yet assigned.
197 */
198 #define VNOVAL (-1)
199
200 /*
201 * Structure returned by the KERN_VNODE sysctl
202 */
203 struct e_vnode {
204 struct vnode *vptr;
205 struct vnode vnode;
206 };
207
208 #ifdef _KERNEL
209 /*
210 * Convert between vnode types and inode formats (since POSIX.1
211 * defines mode word of stat structure in terms of inode formats).
212 */
213 extern enum vtype iftovt_tab[];
214 extern int vttoif_tab[];
215 #define IFTOVT(mode) (iftovt_tab[((mode) & S_IFMT) >> 12])
216 #define VTTOIF(indx) (vttoif_tab[(int)(indx)])
217 #define MAKEIMODE(indx, mode) (int)(VTTOIF(indx) | (mode))
218
219 /*
220 * Flags to various vnode functions.
221 */
222 #define SKIPSYSTEM 0x0001 /* vflush: skip vnodes marked VSYSTEM */
223 #define FORCECLOSE 0x0002 /* vflush: force file closeure */
224 #define WRITECLOSE 0x0004 /* vflush: only close writeable files */
225 #define DOCLOSE 0x0008 /* vclean: close active files */
226 #define V_SAVE 0x0001 /* vinvalbuf: sync file first */
227 #define V_SAVEMETA 0x0002 /* vinvalbuf: leave indirect blocks */
228
229 #define REVOKEALL 0x0001 /* vop_reovke: revoke all aliases */
230
231
232 TAILQ_HEAD(freelst, vnode);
233 extern struct freelst vnode_hold_list; /* free vnodes referencing buffers */
234 extern struct freelst vnode_free_list; /* vnode free list */
235 extern struct simplelock vnode_free_list_slock;
236
237 #ifdef DIAGNOSTIC
238 #define VATTR_NULL(vap) vattr_null(vap)
239
240 #define VREF(vp) vref(vp)
241 void vref(struct vnode *);
242 #else
243 #define VATTR_NULL(vap) (*(vap) = va_null) /* initialize a vattr */
244
245 static __inline void vref(struct vnode *);
246 #define VREF(vp) vref(vp) /* increase reference */
247 static __inline void
vref(vp)248 vref(vp)
249 struct vnode *vp;
250 {
251 simple_lock(&vp->v_interlock);
252 vp->v_usecount++;
253 simple_unlock(&vp->v_interlock);
254 }
255 #endif /* DIAGNOSTIC */
256
257 #define NULLVP ((struct vnode *)NULL)
258
259 /*
260 * Global vnode data.
261 */
262 extern struct vnode *rootvnode; /* root (i.e. "/") vnode */
263 extern int desiredvnodes; /* number of vnodes desired */
264 extern time_t syncdelay; /* time to delay syncing vnodes */
265 extern int rushjob; /* # of slots syncer should run ASAP */
266 extern struct vattr va_null; /* predefined null vattr structure */
267
268 /*
269 * Macro/function to check for client cache inconsistency w.r.t. leasing.
270 */
271 #define LEASE_READ 0x1 /* Check lease for readers */
272 #define LEASE_WRITE 0x2 /* Check lease for modifiers */
273
274 #endif /* _KERNEL */
275
276
277 /*
278 * Mods for exensibility.
279 */
280
281 /*
282 * Flags for vdesc_flags:
283 */
284 #define VDESC_MAX_VPS 16
285 /* Low order 16 flag bits are reserved for willrele flags for vp arguments. */
286 #define VDESC_VP0_WILLRELE 0x00000001
287 #define VDESC_VP1_WILLRELE 0x00000002
288 #define VDESC_VP2_WILLRELE 0x00000004
289 #define VDESC_VP3_WILLRELE 0x00000008
290 #define VDESC_VP0_WILLUNLOCK 0x00000100
291 #define VDESC_VP1_WILLUNLOCK 0x00000200
292 #define VDESC_VP2_WILLUNLOCK 0x00000400
293 #define VDESC_VP3_WILLUNLOCK 0x00000800
294 #define VDESC_VP0_WILLPUT 0x00000101
295 #define VDESC_VP1_WILLPUT 0x00000202
296 #define VDESC_VP2_WILLPUT 0x00000404
297 #define VDESC_VP3_WILLPUT 0x00000808
298 #define VDESC_NOMAP_VPP 0x00010000
299 #define VDESC_VPP_WILLRELE 0x00020000
300
301 /*
302 * VDESC_NO_OFFSET is used to identify the end of the offset list
303 * and in places where no such field exists.
304 */
305 #define VDESC_NO_OFFSET -1
306
307 /*
308 * This structure describes the vnode operation taking place.
309 */
310 struct vnodeop_desc {
311 int vdesc_offset; /* offset in vector--first for speed */
312 char *vdesc_name; /* a readable name for debugging */
313 int vdesc_flags; /* VDESC_* flags */
314
315 /*
316 * These ops are used by bypass routines to map and locate arguments.
317 * Creds and procs are not needed in bypass routines, but sometimes
318 * they are useful to (for example) transport layers.
319 * Nameidata is useful because it has a cred in it.
320 */
321 int *vdesc_vp_offsets; /* list ended by VDESC_NO_OFFSET */
322 int vdesc_vpp_offset; /* return vpp location */
323 int vdesc_cred_offset; /* cred location, if any */
324 int vdesc_proc_offset; /* proc location, if any */
325 int vdesc_componentname_offset; /* if any */
326 /*
327 * Finally, we've got a list of private data (about each operation)
328 * for each transport layer. (Support to manage this list is not
329 * yet part of BSD.)
330 */
331 caddr_t *vdesc_transports;
332 };
333
334 #ifdef _KERNEL
335 /*
336 * A list of all the operation descs.
337 */
338 extern struct vnodeop_desc *vnodeop_descs[];
339
340
341 /*
342 * Interlock for scanning list of vnodes attached to a mountpoint
343 */
344 extern struct simplelock mntvnode_slock;
345
346 /*
347 * This macro is very helpful in defining those offsets in the vdesc struct.
348 *
349 * This is stolen from X11R4. I ingored all the fancy stuff for
350 * Crays, so if you decide to port this to such a serious machine,
351 * you might want to consult Intrisics.h's XtOffset{,Of,To}.
352 */
353 #define VOPARG_OFFSET(p_type,field) \
354 ((int) (((char *) (&(((p_type)NULL)->field))) - ((char *) NULL)))
355 #define VOPARG_OFFSETOF(s_type,field) \
356 VOPARG_OFFSET(s_type*,field)
357 #define VOPARG_OFFSETTO(S_TYPE,S_OFFSET,STRUCT_P) \
358 ((S_TYPE)(((char *)(STRUCT_P))+(S_OFFSET)))
359
360
361 /*
362 * This structure is used to configure the new vnodeops vector.
363 */
364 struct vnodeopv_entry_desc {
365 struct vnodeop_desc *opve_op; /* which operation this is */
366 int (*opve_impl)(void *); /* code implementing this operation */
367 };
368 struct vnodeopv_desc {
369 /* ptr to the ptr to the vector where op should go */
370 int (***opv_desc_vector_p)(void *);
371 struct vnodeopv_entry_desc *opv_desc_ops; /* null terminated list */
372 };
373
374 /*
375 * A default routine which just returns an error.
376 */
377 int vn_default_error(void *);
378
379 /*
380 * A generic structure.
381 * This can be used by bypass routines to identify generic arguments.
382 */
383 struct vop_generic_args {
384 struct vnodeop_desc *a_desc;
385 /* other random data follows, presumably */
386 };
387
388 /*
389 * VOCALL calls an op given an ops vector. We break it out because BSD's
390 * vclean changes the ops vector and then wants to call ops with the old
391 * vector.
392 */
393 #define VOCALL(OPSV,OFF,AP) (( *((OPSV)[(OFF)])) (AP))
394
395 /*
396 * This call works for vnodes in the kernel.
397 */
398 #define VCALL(VP,OFF,AP) VOCALL((VP)->v_op,(OFF),(AP))
399 #define VDESC(OP) (& __CONCAT(OP,_desc))
400 #define VOFFSET(OP) (VDESC(OP)->vdesc_offset)
401
402 /*
403 * Finally, include the default set of vnode operations.
404 */
405 #include <sys/vnode_if.h>
406
407 /*
408 * Public vnode manipulation functions.
409 */
410 struct file;
411 struct filedesc;
412 struct mount;
413 struct nameidata;
414 struct proc;
415 struct stat;
416 struct ucred;
417 struct uio;
418 struct vattr;
419 struct vnode;
420
421 int bdevvp(dev_t dev, struct vnode **vpp);
422 int cdevvp(dev_t dev, struct vnode **vpp);
423 int getnewvnode(enum vtagtype tag, struct mount *mp,
424 int (**vops)(void *), struct vnode **vpp);
425 int getvnode(struct filedesc *fdp, int fd, struct file **fpp);
426 void getnewfsid(struct mount *, int);
427 void vattr_null(struct vattr *vap);
428 int vcount(struct vnode *vp);
429 int vfinddev(dev_t, enum vtype, struct vnode **);
430 void vflushbuf(struct vnode *vp, int sync);
431 int vflush(struct mount *mp, struct vnode *vp, int flags);
432 void vntblinit(void);
433 void vn_initialize_syncerd(void);
434 int vwaitforio(struct vnode *, int, char *, int);
435 void vwakeup(struct vnode *);
436 void vdevgone(int, int, int, enum vtype);
437 int vget(struct vnode *vp, int lockflag, struct proc *p);
438 void vgone(struct vnode *vp);
439 void vgonel(struct vnode *, struct proc *);
440 int vinvalbuf(struct vnode *vp, int save, struct ucred *cred,
441 struct proc *p, int slpflag, int slptimeo);
442 void vprint(char *label, struct vnode *vp);
443 int vop_generic_bwrite(void *ap);
444 void vn_update(void);
445 int vn_close(struct vnode *vp,
446 int flags, struct ucred *cred, struct proc *p);
447 int vn_open(struct nameidata *ndp, int fmode, int cmode);
448 int vrecycle(struct vnode *vp, struct simplelock *inter_lkp,
449 struct proc *p);
450 int vn_rdwr(enum uio_rw rw, struct vnode *vp, caddr_t base,
451 int len, off_t offset, enum uio_seg segflg, int ioflg,
452 struct ucred *cred, size_t *aresid, struct proc *p);
453 int vn_lock(struct vnode *vp, int flags, struct proc *p);
454
455 int vop_generic_abortop(void *);
456 int vop_generic_islocked(void *);
457 int vop_generic_lock(void *);
458 int vop_generic_unlock(void *);
459 int vop_generic_revoke(void *);
460 int vop_generic_kqfilter(void *);
461
462 int vn_stat(struct vnode *vp, struct stat *sb, struct proc *p);
463 int vn_statfile(struct file *fp, struct stat *sb, struct proc *p);
464 int vn_writechk(struct vnode *vp);
465 void vn_marktext(struct vnode *vp);
466 void vn_syncer_add_to_worklist(struct vnode *vp, int delay);
467 void sched_sync(struct proc *);
468
469 struct vnode *
470 checkalias(struct vnode *vp, dev_t nvp_rdev, struct mount *mp);
471 void vput(struct vnode *vp);
472 void vrele(struct vnode *vp);
473 int vaccess(mode_t file_mode, uid_t uid, gid_t gid,
474 mode_t acc_mode, struct ucred *cred);
475
476 int vn_isdisk(struct vnode *vp, int *errp);
477
478 int softdep_fsync(struct vnode *vp);
479
480 #endif /* _KERNEL */
481