1 /* $OpenBSD: vfs_vnops.c,v 1.46 2005/05/28 07:30:25 marius Exp $ */
2 /* $NetBSD: vfs_vnops.c,v 1.20 1996/02/04 02:18:41 christos Exp $ */
3
4 /*
5 * Copyright (c) 1982, 1986, 1989, 1993
6 * The Regents of the University of California. All rights reserved.
7 * (c) UNIX System Laboratories, Inc.
8 * All or some portions of this file are derived from material licensed
9 * to the University of California by American Telephone and Telegraph
10 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
11 * the permission of UNIX System Laboratories, Inc.
12 *
13 * Redistribution and use in source and binary forms, with or without
14 * modification, are permitted provided that the following conditions
15 * are met:
16 * 1. Redistributions of source code must retain the above copyright
17 * notice, this list of conditions and the following disclaimer.
18 * 2. Redistributions in binary form must reproduce the above copyright
19 * notice, this list of conditions and the following disclaimer in the
20 * documentation and/or other materials provided with the distribution.
21 * 3. Neither the name of the University nor the names of its contributors
22 * may be used to endorse or promote products derived from this software
23 * without specific prior written permission.
24 *
25 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
26 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
27 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
28 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
29 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
30 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
31 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
32 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
33 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
34 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
35 * SUCH DAMAGE.
36 *
37 * @(#)vfs_vnops.c 8.5 (Berkeley) 12/8/94
38 */
39
40 #include <sys/param.h>
41 #include <sys/systm.h>
42 #include <sys/kernel.h>
43 #include <sys/file.h>
44 #include <sys/stat.h>
45 #include <sys/buf.h>
46 #include <sys/proc.h>
47 #include <sys/mount.h>
48 #include <sys/namei.h>
49 #include <sys/vnode.h>
50 #include <sys/ioctl.h>
51 #include <sys/tty.h>
52 #include <sys/cdio.h>
53 #include <sys/poll.h>
54
55 #include <uvm/uvm_extern.h>
56
57 int vn_read(struct file *fp, off_t *off, struct uio *uio,
58 struct ucred *cred);
59 int vn_write(struct file *fp, off_t *off, struct uio *uio,
60 struct ucred *cred);
61 int vn_poll(struct file *fp, int events, struct proc *p);
62 int vn_kqfilter(struct file *fp, struct knote *kn);
63 int vn_closefile(struct file *fp, struct proc *p);
64 int vn_ioctl(struct file *fp, u_long com, caddr_t data,
65 struct proc *p);
66
67 struct fileops vnops =
68 { vn_read, vn_write, vn_ioctl, vn_poll, vn_kqfilter, vn_statfile,
69 vn_closefile };
70
71 /*
72 * Common code for vnode open operations.
73 * Check permissions, and call the VOP_OPEN or VOP_CREATE routine.
74 */
75 int
vn_open(ndp,fmode,cmode)76 vn_open(ndp, fmode, cmode)
77 register struct nameidata *ndp;
78 int fmode, cmode;
79 {
80 register struct vnode *vp;
81 register struct proc *p = ndp->ni_cnd.cn_proc;
82 register struct ucred *cred = p->p_ucred;
83 struct vattr va;
84 int error;
85
86 if ((fmode & (FREAD|FWRITE)) == 0)
87 return (EINVAL);
88 if ((fmode & (O_TRUNC | FWRITE)) == O_TRUNC)
89 return (EINVAL);
90 if (fmode & O_CREAT) {
91 ndp->ni_cnd.cn_nameiop = CREATE;
92 ndp->ni_cnd.cn_flags = LOCKPARENT | LOCKLEAF;
93 if ((fmode & O_EXCL) == 0 && (fmode & O_NOFOLLOW) == 0)
94 ndp->ni_cnd.cn_flags |= FOLLOW;
95 if ((error = namei(ndp)) != 0)
96 return (error);
97
98 if (ndp->ni_vp == NULL) {
99 VATTR_NULL(&va);
100 va.va_type = VREG;
101 va.va_mode = cmode;
102 VOP_LEASE(ndp->ni_dvp, p, cred, LEASE_WRITE);
103 error = VOP_CREATE(ndp->ni_dvp, &ndp->ni_vp,
104 &ndp->ni_cnd, &va);
105 if (error)
106 return (error);
107 fmode &= ~O_TRUNC;
108 vp = ndp->ni_vp;
109 } else {
110 VOP_ABORTOP(ndp->ni_dvp, &ndp->ni_cnd);
111 if (ndp->ni_dvp == ndp->ni_vp)
112 vrele(ndp->ni_dvp);
113 else
114 vput(ndp->ni_dvp);
115 ndp->ni_dvp = NULL;
116 vp = ndp->ni_vp;
117 if (fmode & O_EXCL) {
118 error = EEXIST;
119 goto bad;
120 }
121 fmode &= ~O_CREAT;
122 }
123 } else {
124 ndp->ni_cnd.cn_nameiop = LOOKUP;
125 ndp->ni_cnd.cn_flags =
126 ((fmode & O_NOFOLLOW) ? NOFOLLOW : FOLLOW) | LOCKLEAF;
127 if ((error = namei(ndp)) != 0)
128 return (error);
129 vp = ndp->ni_vp;
130 }
131 if (vp->v_type == VSOCK) {
132 error = EOPNOTSUPP;
133 goto bad;
134 }
135 if (vp->v_type == VLNK) {
136 error = EMLINK;
137 goto bad;
138 }
139 if ((fmode & O_CREAT) == 0) {
140 if (fmode & FREAD) {
141 if ((error = VOP_ACCESS(vp, VREAD, cred, p)) != 0)
142 goto bad;
143 }
144 if (fmode & FWRITE) {
145 if (vp->v_type == VDIR) {
146 error = EISDIR;
147 goto bad;
148 }
149 if ((error = vn_writechk(vp)) != 0 ||
150 (error = VOP_ACCESS(vp, VWRITE, cred, p)) != 0)
151 goto bad;
152 }
153 }
154 if ((fmode & O_TRUNC) && vp->v_type == VREG) {
155 VOP_UNLOCK(vp, 0, p); /* XXX */
156 VOP_LEASE(vp, p, cred, LEASE_WRITE);
157 vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p); /* XXX */
158 VATTR_NULL(&va);
159 va.va_size = 0;
160 if ((error = VOP_SETATTR(vp, &va, cred, p)) != 0)
161 goto bad;
162 }
163 if ((error = VOP_OPEN(vp, fmode, cred, p)) != 0)
164 goto bad;
165 if (fmode & FWRITE)
166 vp->v_writecount++;
167 return (0);
168 bad:
169 vput(vp);
170 return (error);
171 }
172
173 /*
174 * Check for write permissions on the specified vnode.
175 * Prototype text segments cannot be written.
176 */
177 int
vn_writechk(vp)178 vn_writechk(vp)
179 register struct vnode *vp;
180 {
181
182 /*
183 * Disallow write attempts on read-only file systems;
184 * unless the file is a socket or a block or character
185 * device resident on the file system.
186 */
187 if (vp->v_mount->mnt_flag & MNT_RDONLY) {
188 switch (vp->v_type) {
189 case VREG: case VDIR: case VLNK:
190 return (EROFS);
191 case VNON: case VCHR: case VSOCK:
192 case VFIFO: case VBAD: case VBLK:
193 break;
194 }
195 }
196 /*
197 * If there's shared text associated with
198 * the vnode, try to free it up once. If
199 * we fail, we can't allow writing.
200 */
201 if ((vp->v_flag & VTEXT) && !uvm_vnp_uncache(vp))
202 return (ETXTBSY);
203
204 return (0);
205 }
206
207 /*
208 * Mark a vnode as being the text image of a running process.
209 */
210 void
vn_marktext(vp)211 vn_marktext(vp)
212 struct vnode *vp;
213 {
214 vp->v_flag |= VTEXT;
215 }
216
217 /*
218 * Vnode close call
219 */
220 int
vn_close(vp,flags,cred,p)221 vn_close(vp, flags, cred, p)
222 register struct vnode *vp;
223 int flags;
224 struct ucred *cred;
225 struct proc *p;
226 {
227 int error;
228
229 if (flags & FWRITE)
230 vp->v_writecount--;
231 vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
232 error = VOP_CLOSE(vp, flags, cred, p);
233 vput(vp);
234 return (error);
235 }
236
237 /*
238 * Package up an I/O request on a vnode into a uio and do it.
239 */
240 int
vn_rdwr(rw,vp,base,len,offset,segflg,ioflg,cred,aresid,p)241 vn_rdwr(rw, vp, base, len, offset, segflg, ioflg, cred, aresid, p)
242 enum uio_rw rw;
243 struct vnode *vp;
244 caddr_t base;
245 int len;
246 off_t offset;
247 enum uio_seg segflg;
248 int ioflg;
249 struct ucred *cred;
250 size_t *aresid;
251 struct proc *p;
252 {
253 struct uio auio;
254 struct iovec aiov;
255 int error;
256
257 if ((ioflg & IO_NODELOCKED) == 0)
258 vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
259 auio.uio_iov = &aiov;
260 auio.uio_iovcnt = 1;
261 aiov.iov_base = base;
262 aiov.iov_len = len;
263 auio.uio_resid = len;
264 auio.uio_offset = offset;
265 auio.uio_segflg = segflg;
266 auio.uio_rw = rw;
267 auio.uio_procp = p;
268 if (rw == UIO_READ) {
269 error = VOP_READ(vp, &auio, ioflg, cred);
270 } else {
271 error = VOP_WRITE(vp, &auio, ioflg, cred);
272 }
273 if (aresid)
274 *aresid = auio.uio_resid;
275 else
276 if (auio.uio_resid && error == 0)
277 error = EIO;
278 if ((ioflg & IO_NODELOCKED) == 0)
279 VOP_UNLOCK(vp, 0, p);
280 return (error);
281 }
282
283 /*
284 * File table vnode read routine.
285 */
286 int
vn_read(fp,poff,uio,cred)287 vn_read(fp, poff, uio, cred)
288 struct file *fp;
289 off_t *poff;
290 struct uio *uio;
291 struct ucred *cred;
292 {
293 register struct vnode *vp = (struct vnode *)fp->f_data;
294 int error = 0;
295 size_t count;
296 struct proc *p = uio->uio_procp;
297
298 VOP_LEASE(vp, uio->uio_procp, cred, LEASE_READ);
299 vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
300 uio->uio_offset = *poff;
301 count = uio->uio_resid;
302 if (vp->v_type != VDIR)
303 error = VOP_READ(vp, uio,
304 (fp->f_flag & FNONBLOCK) ? IO_NDELAY : 0, cred);
305 *poff += count - uio->uio_resid;
306 VOP_UNLOCK(vp, 0, p);
307 return (error);
308 }
309
310 /*
311 * File table vnode write routine.
312 */
313 int
vn_write(fp,poff,uio,cred)314 vn_write(fp, poff, uio, cred)
315 struct file *fp;
316 off_t *poff;
317 struct uio *uio;
318 struct ucred *cred;
319 {
320 register struct vnode *vp = (struct vnode *)fp->f_data;
321 struct proc *p = uio->uio_procp;
322 int error, ioflag = IO_UNIT;
323 size_t count;
324
325 if (vp->v_type == VREG && (fp->f_flag & O_APPEND))
326 ioflag |= IO_APPEND;
327 if (fp->f_flag & FNONBLOCK)
328 ioflag |= IO_NDELAY;
329 if ((fp->f_flag & FFSYNC) ||
330 (vp->v_mount && (vp->v_mount->mnt_flag & MNT_SYNCHRONOUS)))
331 ioflag |= IO_SYNC;
332 VOP_LEASE(vp, uio->uio_procp, cred, LEASE_WRITE);
333 vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
334 uio->uio_offset = *poff;
335 count = uio->uio_resid;
336 error = VOP_WRITE(vp, uio, ioflag, cred);
337 if (ioflag & IO_APPEND)
338 *poff = uio->uio_offset;
339 else
340 *poff += count - uio->uio_resid;
341 VOP_UNLOCK(vp, 0, p);
342 return (error);
343 }
344
345 /*
346 * File table wrapper for vn_stat
347 */
348 int
vn_statfile(fp,sb,p)349 vn_statfile(fp, sb, p)
350 struct file *fp;
351 struct stat *sb;
352 struct proc *p;
353 {
354 struct vnode *vp = (struct vnode *)fp->f_data;
355
356 return vn_stat(vp, sb, p);
357 }
358
359 /*
360 * vnode stat routine.
361 */
362 int
vn_stat(vp,sb,p)363 vn_stat(vp, sb, p)
364 struct vnode *vp;
365 register struct stat *sb;
366 struct proc *p;
367 {
368 struct vattr va;
369 int error;
370 mode_t mode;
371
372 error = VOP_GETATTR(vp, &va, p->p_ucred, p);
373 if (error)
374 return (error);
375 /*
376 * Copy from vattr table
377 */
378 sb->st_dev = va.va_fsid;
379 sb->st_ino = va.va_fileid;
380 mode = va.va_mode;
381 switch (vp->v_type) {
382 case VREG:
383 mode |= S_IFREG;
384 break;
385 case VDIR:
386 mode |= S_IFDIR;
387 break;
388 case VBLK:
389 mode |= S_IFBLK;
390 break;
391 case VCHR:
392 mode |= S_IFCHR;
393 break;
394 case VLNK:
395 mode |= S_IFLNK;
396 break;
397 case VSOCK:
398 mode |= S_IFSOCK;
399 break;
400 case VFIFO:
401 mode |= S_IFIFO;
402 break;
403 default:
404 return (EBADF);
405 }
406 sb->st_mode = mode;
407 sb->st_nlink = va.va_nlink;
408 sb->st_uid = va.va_uid;
409 sb->st_gid = va.va_gid;
410 sb->st_rdev = va.va_rdev;
411 sb->st_size = va.va_size;
412 sb->st_atimespec = va.va_atime;
413 sb->st_mtimespec = va.va_mtime;
414 sb->st_ctimespec = va.va_ctime;
415 sb->st_blksize = va.va_blocksize;
416 sb->st_flags = va.va_flags;
417 sb->st_gen = va.va_gen;
418 sb->st_blocks = va.va_bytes / S_BLKSIZE;
419 return (0);
420 }
421
422 /*
423 * File table vnode ioctl routine.
424 */
425 int
vn_ioctl(fp,com,data,p)426 vn_ioctl(fp, com, data, p)
427 struct file *fp;
428 u_long com;
429 caddr_t data;
430 struct proc *p;
431 {
432 register struct vnode *vp = ((struct vnode *)fp->f_data);
433 struct vattr vattr;
434 int error;
435
436 switch (vp->v_type) {
437
438 case VREG:
439 case VDIR:
440 if (com == FIONREAD) {
441 error = VOP_GETATTR(vp, &vattr, p->p_ucred, p);
442 if (error)
443 return (error);
444 *(int *)data = vattr.va_size - fp->f_offset;
445 return (0);
446 }
447 if (com == FIBMAP)
448 return VOP_IOCTL(vp, com, data, fp->f_flag,
449 p->p_ucred, p);
450 if (com == FIONBIO || com == FIOASYNC) /* XXX */
451 return (0); /* XXX */
452 /* fall into... */
453
454 default:
455 return (ENOTTY);
456
457 case VFIFO:
458 case VCHR:
459 case VBLK:
460 error = VOP_IOCTL(vp, com, data, fp->f_flag, p->p_ucred, p);
461 if (error == 0 && com == TIOCSCTTY) {
462 if (p->p_session->s_ttyvp)
463 vrele(p->p_session->s_ttyvp);
464 p->p_session->s_ttyvp = vp;
465 VREF(vp);
466 }
467 return (error);
468 }
469 }
470
471 /*
472 * File table vnode poll routine.
473 */
474 int
vn_poll(fp,events,p)475 vn_poll(fp, events, p)
476 struct file *fp;
477 int events;
478 struct proc *p;
479 {
480
481 return (VOP_POLL(((struct vnode *)fp->f_data), events, p));
482 }
483
484 /*
485 * Check that the vnode is still valid, and if so
486 * acquire requested lock.
487 */
488 int
vn_lock(struct vnode * vp,int flags,struct proc * p)489 vn_lock(struct vnode *vp, int flags, struct proc *p)
490 {
491 int error;
492
493 if ((flags & LK_RECURSEFAIL) == 0)
494 flags |= LK_CANRECURSE;
495
496 do {
497 if ((flags & LK_INTERLOCK) == 0)
498 simple_lock(&vp->v_interlock);
499 if (vp->v_flag & VXLOCK) {
500 vp->v_flag |= VXWANT;
501 simple_unlock(&vp->v_interlock);
502 tsleep(vp, PINOD, "vn_lock", 0);
503 error = ENOENT;
504 } else {
505 error = VOP_LOCK(vp, flags | LK_INTERLOCK, p);
506 if (error == 0)
507 return (error);
508 }
509 flags &= ~LK_INTERLOCK;
510 } while (flags & LK_RETRY);
511 return (error);
512 }
513
514 /*
515 * File table vnode close routine.
516 */
517 int
vn_closefile(fp,p)518 vn_closefile(fp, p)
519 struct file *fp;
520 struct proc *p;
521 {
522
523 return (vn_close(((struct vnode *)fp->f_data), fp->f_flag,
524 fp->f_cred, p));
525 }
526
527 /*ARGSUSED*/
528 int
vn_kqfilter(struct file * fp,struct knote * kn)529 vn_kqfilter(struct file *fp, struct knote *kn)
530 {
531 return (VOP_KQFILTER(((struct vnode *)fp->f_data), kn));
532 }
533