procfs_vfsops.c revision 1.92 1 /* $NetBSD: procfs_vfsops.c,v 1.92 2014/07/27 16:47:26 hannken Exp $ */
2
3 /*
4 * Copyright (c) 1993
5 * The Regents of the University of California. All rights reserved.
6 *
7 * This code is derived from software contributed to Berkeley by
8 * Jan-Simon Pendry.
9 *
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
12 * are met:
13 * 1. Redistributions of source code must retain the above copyright
14 * notice, this list of conditions and the following disclaimer.
15 * 2. Redistributions in binary form must reproduce the above copyright
16 * notice, this list of conditions and the following disclaimer in the
17 * documentation and/or other materials provided with the distribution.
18 * 3. Neither the name of the University nor the names of its contributors
19 * may be used to endorse or promote products derived from this software
20 * without specific prior written permission.
21 *
22 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32 * SUCH DAMAGE.
33 *
34 * @(#)procfs_vfsops.c 8.7 (Berkeley) 5/10/95
35 */
36
37 /*
38 * Copyright (c) 1993 Jan-Simon Pendry
39 *
40 * This code is derived from software contributed to Berkeley by
41 * Jan-Simon Pendry.
42 *
43 * Redistribution and use in source and binary forms, with or without
44 * modification, are permitted provided that the following conditions
45 * are met:
46 * 1. Redistributions of source code must retain the above copyright
47 * notice, this list of conditions and the following disclaimer.
48 * 2. Redistributions in binary form must reproduce the above copyright
49 * notice, this list of conditions and the following disclaimer in the
50 * documentation and/or other materials provided with the distribution.
51 * 3. All advertising materials mentioning features or use of this software
52 * must display the following acknowledgement:
53 * This product includes software developed by the University of
54 * California, Berkeley and its contributors.
55 * 4. Neither the name of the University nor the names of its contributors
56 * may be used to endorse or promote products derived from this software
57 * without specific prior written permission.
58 *
59 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
60 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
61 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
62 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
63 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
64 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
65 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
66 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
67 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
68 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
69 * SUCH DAMAGE.
70 *
71 * @(#)procfs_vfsops.c 8.7 (Berkeley) 5/10/95
72 */
73
74 /*
75 * procfs VFS interface
76 */
77
78 #include <sys/cdefs.h>
79 __KERNEL_RCSID(0, "$NetBSD: procfs_vfsops.c,v 1.92 2014/07/27 16:47:26 hannken Exp $");
80
81 #if defined(_KERNEL_OPT)
82 #include "opt_compat_netbsd.h"
83 #endif
84
85 #include <sys/param.h>
86 #include <sys/time.h>
87 #include <sys/kernel.h>
88 #include <sys/systm.h>
89 #include <sys/sysctl.h>
90 #include <sys/proc.h>
91 #include <sys/buf.h>
92 #include <sys/syslog.h>
93 #include <sys/mount.h>
94 #include <sys/dirent.h>
95 #include <sys/signalvar.h>
96 #include <sys/vnode.h>
97 #include <sys/malloc.h>
98 #include <sys/file.h>
99 #include <sys/filedesc.h>
100 #include <sys/kauth.h>
101 #include <sys/module.h>
102
103 #include <miscfs/genfs/genfs.h>
104
105 #include <miscfs/procfs/procfs.h>
106
107 #include <uvm/uvm_extern.h> /* for PAGE_SIZE */
108
109 MODULE(MODULE_CLASS_VFS, procfs, NULL);
110
111 VFS_PROTOS(procfs);
112
113 static struct sysctllog *procfs_sysctl_log;
114
115 static kauth_listener_t procfs_listener;
116
117 /*
118 * VFS Operations.
119 *
120 * mount system call
121 */
122 /* ARGSUSED */
123 int
124 procfs_mount(
125 struct mount *mp,
126 const char *path,
127 void *data,
128 size_t *data_len)
129 {
130 struct lwp *l = curlwp;
131 struct procfsmount *pmnt;
132 struct procfs_args *args = data;
133 int error;
134
135 if (args == NULL)
136 return EINVAL;
137
138 if (UIO_MX & (UIO_MX-1)) {
139 log(LOG_ERR, "procfs: invalid directory entry size");
140 return (EINVAL);
141 }
142
143 if (mp->mnt_flag & MNT_GETARGS) {
144 if (*data_len < sizeof *args)
145 return EINVAL;
146
147 pmnt = VFSTOPROC(mp);
148 if (pmnt == NULL)
149 return EIO;
150 args->version = PROCFS_ARGSVERSION;
151 args->flags = pmnt->pmnt_flags;
152 *data_len = sizeof *args;
153 return 0;
154 }
155
156 if (mp->mnt_flag & MNT_UPDATE)
157 return (EOPNOTSUPP);
158
159 if (*data_len >= sizeof *args && args->version != PROCFS_ARGSVERSION)
160 return EINVAL;
161
162 pmnt = kmem_zalloc(sizeof(struct procfsmount), KM_SLEEP);
163
164 mp->mnt_stat.f_namemax = PROCFS_MAXNAMLEN;
165 mp->mnt_flag |= MNT_LOCAL;
166 mp->mnt_data = pmnt;
167 vfs_getnewfsid(mp);
168
169 error = set_statvfs_info(path, UIO_USERSPACE, "procfs", UIO_SYSSPACE,
170 mp->mnt_op->vfs_name, mp, l);
171 pmnt->pmnt_exechook = exechook_establish(procfs_revoke_vnodes, mp);
172 if (*data_len >= sizeof *args)
173 pmnt->pmnt_flags = args->flags;
174 else
175 pmnt->pmnt_flags = 0;
176
177 mp->mnt_iflag |= IMNT_MPSAFE;
178 return error;
179 }
180
181 /*
182 * unmount system call
183 */
184 int
185 procfs_unmount(struct mount *mp, int mntflags)
186 {
187 int error;
188 int flags = 0;
189
190 if (mntflags & MNT_FORCE)
191 flags |= FORCECLOSE;
192
193 if ((error = vflush(mp, 0, flags)) != 0)
194 return (error);
195
196 exechook_disestablish(VFSTOPROC(mp)->pmnt_exechook);
197
198 kmem_free(mp->mnt_data, sizeof(struct procfsmount));
199 mp->mnt_data = NULL;
200
201 return 0;
202 }
203
204 int
205 procfs_root(struct mount *mp, struct vnode **vpp)
206 {
207 int error;
208
209 error = procfs_allocvp(mp, vpp, 0, PFSroot, -1);
210 if (error == 0) {
211 error = vn_lock(*vpp, LK_EXCLUSIVE);
212 if (error != 0) {
213 vrele(*vpp);
214 *vpp = NULL;
215 }
216 }
217
218 return error;
219 }
220
221 /* ARGSUSED */
222 int
223 procfs_start(struct mount *mp, int flags)
224 {
225
226 return (0);
227 }
228
229 /*
230 * Get file system statistics.
231 */
232 int
233 procfs_statvfs(struct mount *mp, struct statvfs *sbp)
234 {
235
236 genfs_statvfs(mp, sbp);
237
238 sbp->f_bsize = PAGE_SIZE;
239 sbp->f_frsize = PAGE_SIZE;
240 sbp->f_iosize = PAGE_SIZE;
241 sbp->f_blocks = 1;
242 sbp->f_files = maxproc; /* approx */
243 sbp->f_ffree = maxproc - nprocs; /* approx */
244 sbp->f_favail = maxproc - nprocs; /* approx */
245
246 return (0);
247 }
248
249 /*ARGSUSED*/
250 int
251 procfs_sync(
252 struct mount *mp,
253 int waitfor,
254 kauth_cred_t uc)
255 {
256
257 return (0);
258 }
259
260 /*ARGSUSED*/
261 int
262 procfs_vget(struct mount *mp, ino_t ino,
263 struct vnode **vpp)
264 {
265 return (EOPNOTSUPP);
266 }
267
268 int
269 procfs_loadvnode(struct mount *mp, struct vnode *vp,
270 const void *key, size_t key_len, const void **new_key)
271 {
272 int error;
273 struct pfskey pfskey;
274 struct pfsnode *pfs;
275
276 KASSERT(key_len == sizeof(pfskey));
277 memcpy(&pfskey, key, key_len);
278
279 pfs = kmem_alloc(sizeof(*pfs), KM_SLEEP);
280 pfs->pfs_pid = pfskey.pk_pid;
281 pfs->pfs_type = pfskey.pk_type;
282 pfs->pfs_fd = pfskey.pk_fd;
283 pfs->pfs_vnode = vp;
284 pfs->pfs_flags = 0;
285 pfs->pfs_fileno =
286 PROCFS_FILENO(pfs->pfs_pid, pfs->pfs_type, pfs->pfs_fd);
287 vp->v_tag = VT_PROCFS;
288 vp->v_op = procfs_vnodeop_p;
289 vp->v_data = pfs;
290
291 switch (pfs->pfs_type) {
292 case PFSroot: /* /proc = dr-xr-xr-x */
293 vp->v_vflag |= VV_ROOT;
294 /*FALLTHROUGH*/
295 case PFSproc: /* /proc/N = dr-xr-xr-x */
296 pfs->pfs_mode = S_IRUSR|S_IXUSR|S_IRGRP|S_IXGRP|S_IROTH|S_IXOTH;
297 vp->v_type = VDIR;
298 break;
299
300 case PFStask: /* /proc/N/task = dr-xr-xr-x */
301 if (pfs->pfs_fd == -1) {
302 pfs->pfs_mode = S_IRUSR|S_IXUSR|S_IRGRP|S_IXGRP|
303 S_IROTH|S_IXOTH;
304 vp->v_type = VDIR;
305 break;
306 }
307 /*FALLTHROUGH*/
308 case PFScurproc: /* /proc/curproc = lr-xr-xr-x */
309 case PFSself: /* /proc/self = lr-xr-xr-x */
310 case PFScwd: /* /proc/N/cwd = lr-xr-xr-x */
311 case PFSchroot: /* /proc/N/chroot = lr-xr-xr-x */
312 case PFSexe: /* /proc/N/exe = lr-xr-xr-x */
313 pfs->pfs_mode = S_IRUSR|S_IXUSR|S_IRGRP|S_IXGRP|S_IROTH|S_IXOTH;
314 vp->v_type = VLNK;
315 break;
316
317 case PFSfd:
318 if (pfs->pfs_fd == -1) { /* /proc/N/fd = dr-x------ */
319 pfs->pfs_mode = S_IRUSR|S_IXUSR;
320 vp->v_type = VDIR;
321 } else { /* /proc/N/fd/M = [ps-]rw------- */
322 file_t *fp;
323 vnode_t *vxp;
324 struct proc *p;
325
326 mutex_enter(proc_lock);
327 p = proc_find(pfs->pfs_pid);
328 mutex_exit(proc_lock);
329 if (p == NULL) {
330 error = ENOENT;
331 goto bad;
332 }
333 KASSERT(rw_read_held(&p->p_reflock));
334 if ((fp = fd_getfile2(p, pfs->pfs_fd)) == NULL) {
335 error = EBADF;
336 goto bad;
337 }
338
339 pfs->pfs_mode = S_IRUSR|S_IWUSR;
340 switch (fp->f_type) {
341 case DTYPE_VNODE:
342 vxp = fp->f_data;
343
344 /*
345 * We make symlinks for directories
346 * to avoid cycles.
347 */
348 if (vxp->v_type == VDIR)
349 goto symlink;
350 vp->v_type = vxp->v_type;
351 break;
352 case DTYPE_PIPE:
353 vp->v_type = VFIFO;
354 break;
355 case DTYPE_SOCKET:
356 vp->v_type = VSOCK;
357 break;
358 case DTYPE_KQUEUE:
359 case DTYPE_MISC:
360 case DTYPE_SEM:
361 symlink:
362 pfs->pfs_mode = S_IRUSR|S_IXUSR|S_IRGRP|
363 S_IXGRP|S_IROTH|S_IXOTH;
364 vp->v_type = VLNK;
365 break;
366 default:
367 error = EOPNOTSUPP;
368 closef(fp);
369 goto bad;
370 }
371 closef(fp);
372 }
373 break;
374
375 case PFSfile: /* /proc/N/file = -rw------- */
376 case PFSmem: /* /proc/N/mem = -rw------- */
377 case PFSregs: /* /proc/N/regs = -rw------- */
378 case PFSfpregs: /* /proc/N/fpregs = -rw------- */
379 pfs->pfs_mode = S_IRUSR|S_IWUSR;
380 vp->v_type = VREG;
381 break;
382
383 case PFSctl: /* /proc/N/ctl = --w------ */
384 case PFSnote: /* /proc/N/note = --w------ */
385 case PFSnotepg: /* /proc/N/notepg = --w------ */
386 pfs->pfs_mode = S_IWUSR;
387 vp->v_type = VREG;
388 break;
389
390 case PFSmap: /* /proc/N/map = -r--r--r-- */
391 case PFSmaps: /* /proc/N/maps = -r--r--r-- */
392 case PFSstatus: /* /proc/N/status = -r--r--r-- */
393 case PFSstat: /* /proc/N/stat = -r--r--r-- */
394 case PFScmdline: /* /proc/N/cmdline = -r--r--r-- */
395 case PFSemul: /* /proc/N/emul = -r--r--r-- */
396 case PFSmeminfo: /* /proc/meminfo = -r--r--r-- */
397 case PFScpustat: /* /proc/stat = -r--r--r-- */
398 case PFSdevices: /* /proc/devices = -r--r--r-- */
399 case PFScpuinfo: /* /proc/cpuinfo = -r--r--r-- */
400 case PFSuptime: /* /proc/uptime = -r--r--r-- */
401 case PFSmounts: /* /proc/mounts = -r--r--r-- */
402 case PFSloadavg: /* /proc/loadavg = -r--r--r-- */
403 case PFSstatm: /* /proc/N/statm = -r--r--r-- */
404 case PFSversion: /* /proc/version = -r--r--r-- */
405 pfs->pfs_mode = S_IRUSR|S_IRGRP|S_IROTH;
406 vp->v_type = VREG;
407 break;
408
409 #ifdef __HAVE_PROCFS_MACHDEP
410 PROCFS_MACHDEP_NODETYPE_CASES
411 procfs_machdep_allocvp(vp);
412 break;
413 #endif
414
415 default:
416 panic("procfs_allocvp");
417 }
418
419 uvm_vnp_setsize(vp, 0);
420 *new_key = &pfs->pfs_key;
421
422 return 0;
423
424 bad:
425 vp->v_tag =VT_NON;
426 vp->v_type = VNON;
427 vp->v_op = NULL;
428 vp->v_data = NULL;
429 kmem_free(pfs, sizeof(*pfs));
430 return error;
431 }
432
433 void
434 procfs_init(void)
435 {
436
437 }
438
439 void
440 procfs_reinit(void)
441 {
442
443 }
444
445 void
446 procfs_done(void)
447 {
448
449 }
450
451 extern const struct vnodeopv_desc procfs_vnodeop_opv_desc;
452
453 const struct vnodeopv_desc * const procfs_vnodeopv_descs[] = {
454 &procfs_vnodeop_opv_desc,
455 NULL,
456 };
457
458 struct vfsops procfs_vfsops = {
459 .vfs_name = MOUNT_PROCFS,
460 .vfs_min_mount_data = sizeof (struct procfs_args),
461 .vfs_mount = procfs_mount,
462 .vfs_start = procfs_start,
463 .vfs_unmount = procfs_unmount,
464 .vfs_root = procfs_root,
465 .vfs_quotactl = (void *)eopnotsupp,
466 .vfs_statvfs = procfs_statvfs,
467 .vfs_sync = procfs_sync,
468 .vfs_vget = procfs_vget,
469 .vfs_loadvnode = procfs_loadvnode,
470 .vfs_fhtovp = (void *)eopnotsupp,
471 .vfs_vptofh = (void *)eopnotsupp,
472 .vfs_init = procfs_init,
473 .vfs_reinit = procfs_reinit,
474 .vfs_done = procfs_done,
475 .vfs_snapshot = (void *)eopnotsupp,
476 .vfs_extattrctl = vfs_stdextattrctl,
477 .vfs_suspendctl = (void *)eopnotsupp,
478 .vfs_renamelock_enter = genfs_renamelock_enter,
479 .vfs_renamelock_exit = genfs_renamelock_exit,
480 .vfs_fsync = (void *)eopnotsupp,
481 .vfs_opv_descs = procfs_vnodeopv_descs
482 };
483
484 static int
485 procfs_listener_cb(kauth_cred_t cred, kauth_action_t action, void *cookie,
486 void *arg0, void *arg1, void *arg2, void *arg3)
487 {
488 struct proc *p;
489 struct pfsnode *pfs;
490 enum kauth_process_req req;
491 int result;
492
493 result = KAUTH_RESULT_DEFER;
494 p = arg0;
495 pfs = arg1;
496 req = (enum kauth_process_req)(unsigned long)arg2;
497
498 if (action != KAUTH_PROCESS_PROCFS)
499 return result;
500
501 /* Privileged; let secmodel handle that. */
502 if (req == KAUTH_REQ_PROCESS_PROCFS_CTL)
503 return result;
504
505 switch (pfs->pfs_type) {
506 case PFSregs:
507 case PFSfpregs:
508 case PFSmem:
509 if (kauth_cred_getuid(cred) != kauth_cred_getuid(p->p_cred) ||
510 ISSET(p->p_flag, PK_SUGID))
511 break;
512
513 /*FALLTHROUGH*/
514 default:
515 result = KAUTH_RESULT_ALLOW;
516 break;
517 }
518
519 return result;
520 }
521
522
523 static int
524 procfs_modcmd(modcmd_t cmd, void *arg)
525 {
526 int error;
527
528 switch (cmd) {
529 case MODULE_CMD_INIT:
530 error = vfs_attach(&procfs_vfsops);
531 if (error != 0)
532 break;
533 sysctl_createv(&procfs_sysctl_log, 0, NULL, NULL,
534 CTLFLAG_PERMANENT,
535 CTLTYPE_NODE, "procfs",
536 SYSCTL_DESCR("Process file system"),
537 NULL, 0, NULL, 0,
538 CTL_VFS, 12, CTL_EOL);
539 /*
540 * XXX the "12" above could be dynamic, thereby eliminating
541 * one more instance of the "number to vfs" mapping problem,
542 * but "12" is the order as taken from sys/mount.h
543 */
544
545 procfs_listener = kauth_listen_scope(KAUTH_SCOPE_PROCESS,
546 procfs_listener_cb, NULL);
547
548 break;
549 case MODULE_CMD_FINI:
550 error = vfs_detach(&procfs_vfsops);
551 if (error != 0)
552 break;
553 sysctl_teardown(&procfs_sysctl_log);
554 kauth_unlisten_scope(procfs_listener);
555 break;
556 default:
557 error = ENOTTY;
558 break;
559 }
560
561 return (error);
562 }
563