rtsock.c revision 1.46 1 /* $NetBSD: rtsock.c,v 1.46 2001/06/04 01:30:11 itojun Exp $ */
2
3 /*
4 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
5 * All rights reserved.
6 *
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
9 * are met:
10 * 1. Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
15 * 3. Neither the name of the project nor the names of its contributors
16 * may be used to endorse or promote products derived from this software
17 * without specific prior written permission.
18 *
19 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 * SUCH DAMAGE.
30 */
31
32 /*
33 * Copyright (c) 1988, 1991, 1993
34 * The Regents of the University of California. All rights reserved.
35 *
36 * Redistribution and use in source and binary forms, with or without
37 * modification, are permitted provided that the following conditions
38 * are met:
39 * 1. Redistributions of source code must retain the above copyright
40 * notice, this list of conditions and the following disclaimer.
41 * 2. Redistributions in binary form must reproduce the above copyright
42 * notice, this list of conditions and the following disclaimer in the
43 * documentation and/or other materials provided with the distribution.
44 * 3. All advertising materials mentioning features or use of this software
45 * must display the following acknowledgement:
46 * This product includes software developed by the University of
47 * California, Berkeley and its contributors.
48 * 4. Neither the name of the University nor the names of its contributors
49 * may be used to endorse or promote products derived from this software
50 * without specific prior written permission.
51 *
52 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
53 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
54 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
55 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
56 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
57 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
58 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
59 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
60 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
61 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
62 * SUCH DAMAGE.
63 *
64 * @(#)rtsock.c 8.7 (Berkeley) 10/12/95
65 */
66
67 #include "opt_inet.h"
68
69 #include <sys/param.h>
70 #include <sys/systm.h>
71 #include <sys/proc.h>
72 #include <sys/mbuf.h>
73 #include <sys/socket.h>
74 #include <sys/socketvar.h>
75 #include <sys/domain.h>
76 #include <sys/protosw.h>
77
78 #include <uvm/uvm_extern.h>
79
80 #include <sys/sysctl.h>
81
82 #include <net/if.h>
83 #include <net/route.h>
84 #include <net/raw_cb.h>
85
86 #include <machine/stdarg.h>
87
88 struct sockaddr route_dst = { 2, PF_ROUTE, };
89 struct sockaddr route_src = { 2, PF_ROUTE, };
90 struct sockproto route_proto = { PF_ROUTE, };
91
92 struct walkarg {
93 int w_op;
94 int w_arg;
95 int w_given;
96 int w_needed;
97 caddr_t w_where;
98 int w_tmemsize;
99 int w_tmemneeded;
100 caddr_t w_tmem;
101 };
102
103 static struct mbuf *rt_msg1 __P((int, struct rt_addrinfo *, caddr_t, int));
104 static int rt_msg2 __P((int, struct rt_addrinfo *, caddr_t, struct walkarg *,
105 int *));
106 static int rt_xaddrs __P((caddr_t, caddr_t, struct rt_addrinfo *));
107 static int sysctl_dumpentry __P((struct radix_node *, void *));
108 static int sysctl_iflist __P((int, struct walkarg *, int));
109 static int sysctl_rtable __P((int *, u_int, void *, size_t *, void *, size_t));
110 static __inline void rt_adjustcount __P((int, int));
111
112 /* Sleazy use of local variables throughout file, warning!!!! */
113 #define dst info.rti_info[RTAX_DST]
114 #define gate info.rti_info[RTAX_GATEWAY]
115 #define netmask info.rti_info[RTAX_NETMASK]
116 #define genmask info.rti_info[RTAX_GENMASK]
117 #define ifpaddr info.rti_info[RTAX_IFP]
118 #define ifaaddr info.rti_info[RTAX_IFA]
119 #define brdaddr info.rti_info[RTAX_BRD]
120
121 static __inline void
122 rt_adjustcount(af, cnt)
123 int af, cnt;
124 {
125 route_cb.any_count += cnt;
126 switch (af) {
127 case AF_INET:
128 route_cb.ip_count += cnt;
129 return;
130 #ifdef INET6
131 case AF_INET6:
132 route_cb.ip6_count += cnt;
133 return;
134 #endif
135 case AF_IPX:
136 route_cb.ipx_count += cnt;
137 return;
138 case AF_NS:
139 route_cb.ns_count += cnt;
140 return;
141 case AF_ISO:
142 route_cb.iso_count += cnt;
143 return;
144 }
145 }
146
147 /*ARGSUSED*/
148 int
149 route_usrreq(so, req, m, nam, control, p)
150 struct socket *so;
151 int req;
152 struct mbuf *m, *nam, *control;
153 struct proc *p;
154 {
155 int error = 0;
156 struct rawcb *rp = sotorawcb(so);
157 int s;
158
159 if (req == PRU_ATTACH) {
160 MALLOC(rp, struct rawcb *, sizeof(*rp), M_PCB, M_WAITOK);
161 if ((so->so_pcb = rp) != NULL)
162 bzero(so->so_pcb, sizeof(*rp));
163
164 }
165 if (req == PRU_DETACH && rp)
166 rt_adjustcount(rp->rcb_proto.sp_protocol, -1);
167 s = splsoftnet();
168
169 /*
170 * Don't call raw_usrreq() in the attach case, because
171 * we want to allow non-privileged processes to listen on
172 * and send "safe" commands to the routing socket.
173 */
174 if (req == PRU_ATTACH) {
175 if (p == 0)
176 error = EACCES;
177 else
178 error = raw_attach(so, (int)(long)nam);
179 } else
180 error = raw_usrreq(so, req, m, nam, control, p);
181
182 rp = sotorawcb(so);
183 if (req == PRU_ATTACH && rp) {
184 if (error) {
185 free((caddr_t)rp, M_PCB);
186 splx(s);
187 return (error);
188 }
189 rt_adjustcount(rp->rcb_proto.sp_protocol, 1);
190 rp->rcb_laddr = &route_src;
191 rp->rcb_faddr = &route_dst;
192 soisconnected(so);
193 so->so_options |= SO_USELOOPBACK;
194 }
195 splx(s);
196 return (error);
197 }
198
199 /*ARGSUSED*/
200 int
201 #if __STDC__
202 route_output(struct mbuf *m, ...)
203 #else
204 route_output(m, va_alist)
205 struct mbuf *m;
206 va_dcl
207 #endif
208 {
209 struct rt_msghdr *rtm = 0;
210 struct radix_node *rn = 0;
211 struct rtentry *rt = 0;
212 struct rtentry *saved_nrt = 0;
213 struct radix_node_head *rnh;
214 struct rt_addrinfo info;
215 int len, error = 0;
216 struct ifnet *ifp = 0;
217 struct ifaddr *ifa = 0;
218 struct socket *so;
219 va_list ap;
220
221 va_start(ap, m);
222 so = va_arg(ap, struct socket *);
223 va_end(ap);
224
225 #define senderr(e) do { error = e; goto flush;} while (0)
226 if (m == 0 || ((m->m_len < sizeof(int32_t)) &&
227 (m = m_pullup(m, sizeof(int32_t))) == 0))
228 return (ENOBUFS);
229 if ((m->m_flags & M_PKTHDR) == 0)
230 panic("route_output");
231 len = m->m_pkthdr.len;
232 if (len < sizeof(*rtm) ||
233 len != mtod(m, struct rt_msghdr *)->rtm_msglen) {
234 dst = 0;
235 senderr(EINVAL);
236 }
237 R_Malloc(rtm, struct rt_msghdr *, len);
238 if (rtm == 0) {
239 dst = 0;
240 senderr(ENOBUFS);
241 }
242 m_copydata(m, 0, len, (caddr_t)rtm);
243 if (rtm->rtm_version != RTM_VERSION) {
244 dst = 0;
245 senderr(EPROTONOSUPPORT);
246 }
247 rtm->rtm_pid = curproc->p_pid;
248 bzero(&info, sizeof(info));
249 info.rti_addrs = rtm->rtm_addrs;
250 if (rt_xaddrs((caddr_t)(rtm + 1), len + (caddr_t)rtm, &info))
251 senderr(EINVAL);
252 info.rti_flags = rtm->rtm_flags;
253 if (dst == 0 || (dst->sa_family >= AF_MAX))
254 senderr(EINVAL);
255 if (gate != 0 && (gate->sa_family >= AF_MAX))
256 senderr(EINVAL);
257 if (genmask) {
258 struct radix_node *t;
259 t = rn_addmask((caddr_t)genmask, 0, 1);
260 if (t && Bcmp(genmask, t->rn_key, *(u_char *)genmask) == 0)
261 genmask = (struct sockaddr *)(t->rn_key);
262 else
263 senderr(ENOBUFS);
264 }
265
266 /*
267 * Verify that the caller has the appropriate privilege; RTM_GET
268 * is the only operation the non-superuser is allowed.
269 */
270 if (rtm->rtm_type != RTM_GET &&
271 suser(curproc->p_ucred, &curproc->p_acflag) != 0)
272 senderr(EACCES);
273
274 switch (rtm->rtm_type) {
275
276 case RTM_ADD:
277 if (gate == 0)
278 senderr(EINVAL);
279 error = rtrequest1(rtm->rtm_type, &info, &saved_nrt);
280 if (error == 0 && saved_nrt) {
281 rt_setmetrics(rtm->rtm_inits,
282 &rtm->rtm_rmx, &saved_nrt->rt_rmx);
283 saved_nrt->rt_refcnt--;
284 saved_nrt->rt_genmask = genmask;
285 }
286 break;
287
288 case RTM_DELETE:
289 error = rtrequest1(rtm->rtm_type, &info, &saved_nrt);
290 if (error == 0) {
291 (rt = saved_nrt)->rt_refcnt++;
292 goto report;
293 }
294 break;
295
296 case RTM_GET:
297 case RTM_CHANGE:
298 case RTM_LOCK:
299 if ((rnh = rt_tables[dst->sa_family]) == 0) {
300 senderr(EAFNOSUPPORT);
301 }
302 rn = rnh->rnh_lookup(dst, netmask, rnh);
303 if (rn == NULL || (rn->rn_flags & RNF_ROOT) != 0) {
304 senderr(ESRCH);
305 }
306 rt = (struct rtentry *)rn;
307 rt->rt_refcnt++;
308
309 switch(rtm->rtm_type) {
310
311 case RTM_GET:
312 report:
313 dst = rt_key(rt);
314 gate = rt->rt_gateway;
315 netmask = rt_mask(rt);
316 genmask = rt->rt_genmask;
317 if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) {
318 if ((ifp = rt->rt_ifp) != NULL) {
319 ifpaddr = ifp->if_addrlist.tqh_first->ifa_addr;
320 ifaaddr = rt->rt_ifa->ifa_addr;
321 if (ifp->if_flags & IFF_POINTOPOINT)
322 brdaddr = rt->rt_ifa->ifa_dstaddr;
323 else
324 brdaddr = 0;
325 rtm->rtm_index = ifp->if_index;
326 } else {
327 ifpaddr = 0;
328 ifaaddr = 0;
329 }
330 }
331 (void)rt_msg2(rtm->rtm_type, &info, (caddr_t)0,
332 (struct walkarg *)0, &len);
333 if (len > rtm->rtm_msglen) {
334 struct rt_msghdr *new_rtm;
335 R_Malloc(new_rtm, struct rt_msghdr *, len);
336 if (new_rtm == 0)
337 senderr(ENOBUFS);
338 Bcopy(rtm, new_rtm, rtm->rtm_msglen);
339 Free(rtm); rtm = new_rtm;
340 }
341 (void)rt_msg2(rtm->rtm_type, &info, (caddr_t)rtm,
342 (struct walkarg *)0, 0);
343 rtm->rtm_flags = rt->rt_flags;
344 rtm->rtm_rmx = rt->rt_rmx;
345 rtm->rtm_addrs = info.rti_addrs;
346 break;
347
348 case RTM_CHANGE:
349 /*
350 * new gateway could require new ifaddr, ifp;
351 * flags may also be different; ifp may be specified
352 * by ll sockaddr when protocol address is ambiguous
353 */
354 if ((error = rt_getifa(&info)) != 0)
355 senderr(error);
356 if (gate && rt_setgate(rt, rt_key(rt), gate))
357 senderr(EDQUOT);
358 /* new gateway could require new ifaddr, ifp;
359 flags may also be different; ifp may be specified
360 by ll sockaddr when protocol address is ambiguous */
361 if (ifpaddr && (ifa = ifa_ifwithnet(ifpaddr)) &&
362 (ifp = ifa->ifa_ifp) && (ifaaddr || gate))
363 ifa = ifaof_ifpforaddr(ifaaddr ? ifaaddr : gate,
364 ifp);
365 else if ((ifaaddr && (ifa = ifa_ifwithaddr(ifaaddr))) ||
366 (gate && (ifa = ifa_ifwithroute(rt->rt_flags,
367 rt_key(rt), gate))))
368 ifp = ifa->ifa_ifp;
369 if (ifa) {
370 struct ifaddr *oifa = rt->rt_ifa;
371 if (oifa != ifa) {
372 if (oifa && oifa->ifa_rtrequest)
373 oifa->ifa_rtrequest(RTM_DELETE, rt,
374 &info);
375 IFAFREE(rt->rt_ifa);
376 rt->rt_ifa = ifa;
377 IFAREF(rt->rt_ifa);
378 rt->rt_ifp = ifp;
379 }
380 }
381 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx,
382 &rt->rt_rmx);
383 if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest)
384 rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, &info);
385 if (genmask)
386 rt->rt_genmask = genmask;
387 /*
388 * Fall into
389 */
390 case RTM_LOCK:
391 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits);
392 rt->rt_rmx.rmx_locks |=
393 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks);
394 break;
395 }
396 break;
397
398 default:
399 senderr(EOPNOTSUPP);
400 }
401
402 flush:
403 if (rtm) {
404 if (error)
405 rtm->rtm_errno = error;
406 else
407 rtm->rtm_flags |= RTF_DONE;
408 }
409 if (rt)
410 rtfree(rt);
411 {
412 struct rawcb *rp = 0;
413 /*
414 * Check to see if we don't want our own messages.
415 */
416 if ((so->so_options & SO_USELOOPBACK) == 0) {
417 if (route_cb.any_count <= 1) {
418 if (rtm)
419 Free(rtm);
420 m_freem(m);
421 return (error);
422 }
423 /* There is another listener, so construct message */
424 rp = sotorawcb(so);
425 }
426 /*
427 * copy rtm into m.
428 * XXX is it okay if we omit responses when we are unable to reply?
429 */
430 if (rtm) {
431 if (rtm->rtm_msglen > MCLBYTES) {
432 m_freem(m);
433 m = NULL;
434 } else if (m->m_pkthdr.len > rtm->rtm_msglen) {
435 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm);
436 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len);
437 } else {
438 m_freem(m);
439 MGETHDR(m, M_DONTWAIT, MT_DATA);
440 if (m && rtm->rtm_msglen > MHLEN) {
441 MCLGET(m, M_DONTWAIT);
442 if ((m->m_flags & M_EXT) == 0) {
443 m_free(m);
444 m = NULL;
445 }
446 }
447 if (m) {
448 m->m_pkthdr.len = m->m_len = rtm->rtm_msglen;
449 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm);
450 }
451 }
452 Free(rtm);
453 } else {
454 m_freem(m);
455 m = NULL;
456 }
457 if (rp)
458 rp->rcb_proto.sp_family = 0; /* Avoid us */
459 if (dst)
460 route_proto.sp_protocol = dst->sa_family;
461 if (m)
462 raw_input(m, &route_proto, &route_src, &route_dst);
463 if (rp)
464 rp->rcb_proto.sp_family = PF_ROUTE;
465 }
466 return (error);
467 }
468
469 void
470 rt_setmetrics(which, in, out)
471 u_long which;
472 struct rt_metrics *in, *out;
473 {
474 #define metric(f, e) if (which & (f)) out->e = in->e;
475 metric(RTV_RPIPE, rmx_recvpipe);
476 metric(RTV_SPIPE, rmx_sendpipe);
477 metric(RTV_SSTHRESH, rmx_ssthresh);
478 metric(RTV_RTT, rmx_rtt);
479 metric(RTV_RTTVAR, rmx_rttvar);
480 metric(RTV_HOPCOUNT, rmx_hopcount);
481 metric(RTV_MTU, rmx_mtu);
482 metric(RTV_EXPIRE, rmx_expire);
483 #undef metric
484 }
485
486 #define ROUNDUP(a) \
487 ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long))
488 #define ADVANCE(x, n) (x += ROUNDUP((n)->sa_len))
489
490 static int
491 rt_xaddrs(cp, cplim, rtinfo)
492 caddr_t cp, cplim;
493 struct rt_addrinfo *rtinfo;
494 {
495 struct sockaddr *sa;
496 int i;
497
498 bzero(rtinfo->rti_info, sizeof(rtinfo->rti_info));
499 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) {
500 if ((rtinfo->rti_addrs & (1 << i)) == 0)
501 continue;
502 rtinfo->rti_info[i] = sa = (struct sockaddr *)cp;
503 ADVANCE(cp, sa);
504 }
505
506 /* Check for extra addresses specified. */
507 if ((rtinfo->rti_addrs & (~0 << i)) != 0)
508 return (1);
509 /* Check for bad data length. */
510 if (cp != cplim) {
511 if (i == RTAX_NETMASK + 1 &&
512 cp - ROUNDUP(sa->sa_len) + sa->sa_len == cplim)
513 /*
514 * The last sockaddr was netmask.
515 * We accept this for now for the sake of old
516 * binaries or third party softwares.
517 */
518 ;
519 else
520 return (1);
521 }
522 return (0);
523 }
524
525 static struct mbuf *
526 rt_msg1(type, rtinfo, data, datalen)
527 int type;
528 struct rt_addrinfo *rtinfo;
529 caddr_t data;
530 int datalen;
531 {
532 struct rt_msghdr *rtm;
533 struct mbuf *m;
534 int i;
535 struct sockaddr *sa;
536 int hlen, alen, len, dlen;
537
538 /* pre-compute length */
539 switch (type) {
540 case RTM_DELADDR:
541 case RTM_NEWADDR:
542 hlen = sizeof(struct ifa_msghdr);
543 break;
544
545 #ifdef COMPAT_14
546 case RTM_OIFINFO:
547 hlen = sizeof(struct if_msghdr14);
548 break;
549 #endif
550
551 case RTM_IFINFO:
552 hlen = sizeof(struct if_msghdr);
553 break;
554
555 case RTM_IFANNOUNCE:
556 hlen = sizeof(struct if_announcemsghdr);
557 break;
558
559 default:
560 hlen = sizeof(struct rt_msghdr);
561 break;
562 }
563 alen = 0;
564 for (i = 0; i < RTAX_MAX; i++) {
565 if ((sa = rtinfo->rti_info[i]) == NULL)
566 continue;
567 dlen = ROUNDUP(sa->sa_len);
568 alen += dlen;
569 }
570
571 if (hlen > MHLEN || alen > MCLBYTES)
572 panic("rt_msg1: message too long");
573 m = m_gethdr(M_DONTWAIT, MT_DATA);
574 if (!m)
575 return (NULL);
576 m->m_pkthdr.len = hlen;
577 m->m_len = hlen;
578 if (hlen + alen > MHLEN) {
579 m->m_next = m_get(M_DONTWAIT, MT_DATA);
580 if (m->m_next && alen > MLEN) {
581 MCLGET(m->m_next, M_DONTWAIT);
582 if ((m->m_next->m_flags & M_EXT) == 0) {
583 m_freem(m->m_next);
584 m->m_next = NULL;
585 }
586 }
587 if (!m->m_next) {
588 m_freem(m);
589 return (NULL);
590 }
591 m->m_pkthdr.len += alen;
592 m->m_next->m_len = alen;
593 } else {
594 m->m_pkthdr.len += alen;
595 m->m_len += alen;
596 }
597 m->m_pkthdr.rcvif = NULL;
598 m_copyback(m, 0, datalen, data);
599 rtm = mtod(m, struct rt_msghdr *);
600 len = hlen;
601 for (i = 0; i < RTAX_MAX; i++) {
602 if ((sa = rtinfo->rti_info[i]) == NULL)
603 continue;
604 rtinfo->rti_addrs |= (1 << i);
605 dlen = ROUNDUP(sa->sa_len);
606 m_copyback(m, len, dlen, (caddr_t)sa);
607 len += dlen;
608 }
609 rtm->rtm_msglen = len;
610 rtm->rtm_version = RTM_VERSION;
611 rtm->rtm_type = type;
612 return (m);
613 }
614
615 /*
616 * rt_msg2
617 *
618 * fills 'cp' or 'w'.w_tmem with the routing socket message and
619 * returns the length of the message in 'lenp'.
620 *
621 * if walkarg is 0, cp is expected to be 0 or a buffer large enough to hold
622 * the message
623 * otherwise walkarg's w_needed is updated and if the user buffer is
624 * specified and w_needed indicates space exists the information is copied
625 * into the temp space (w_tmem). w_tmem is [re]allocated if necessary,
626 * if the allocation fails ENOBUFS is returned.
627 */
628 static int
629 rt_msg2(type, rtinfo, cp, w, lenp)
630 int type;
631 struct rt_addrinfo *rtinfo;
632 caddr_t cp;
633 struct walkarg *w;
634 int *lenp;
635 {
636 int i;
637 int len, dlen, second_time = 0;
638 caddr_t cp0;
639
640 rtinfo->rti_addrs = 0;
641 again:
642 switch (type) {
643
644 case RTM_DELADDR:
645 case RTM_NEWADDR:
646 len = sizeof(struct ifa_msghdr);
647 break;
648 #ifdef COMPAT_14
649 case RTM_OIFINFO:
650 len = sizeof(struct if_msghdr14);
651 break;
652 #endif
653
654 case RTM_IFINFO:
655 len = sizeof(struct if_msghdr);
656 break;
657
658 default:
659 len = sizeof(struct rt_msghdr);
660 }
661 if ((cp0 = cp) != NULL)
662 cp += len;
663 for (i = 0; i < RTAX_MAX; i++) {
664 struct sockaddr *sa;
665
666 if ((sa = rtinfo->rti_info[i]) == 0)
667 continue;
668 rtinfo->rti_addrs |= (1 << i);
669 dlen = ROUNDUP(sa->sa_len);
670 if (cp) {
671 bcopy(sa, cp, (unsigned)dlen);
672 cp += dlen;
673 }
674 len += dlen;
675 }
676 if (cp == 0 && w != NULL && !second_time) {
677 struct walkarg *rw = w;
678
679 rw->w_needed += len;
680 if (rw->w_needed <= 0 && rw->w_where) {
681 if (rw->w_tmemsize < len) {
682 if (rw->w_tmem)
683 free(rw->w_tmem, M_RTABLE);
684 rw->w_tmem = (caddr_t) malloc(len, M_RTABLE,
685 M_NOWAIT);
686 if (rw->w_tmem)
687 rw->w_tmemsize = len;
688 }
689 if (rw->w_tmem) {
690 cp = rw->w_tmem;
691 second_time = 1;
692 goto again;
693 } else {
694 rw->w_tmemneeded = len;
695 return (ENOBUFS);
696 }
697 }
698 }
699 if (cp) {
700 struct rt_msghdr *rtm = (struct rt_msghdr *)cp0;
701
702 rtm->rtm_version = RTM_VERSION;
703 rtm->rtm_type = type;
704 rtm->rtm_msglen = len;
705 }
706 if (lenp)
707 *lenp = len;
708 return (0);
709 }
710
711 /*
712 * This routine is called to generate a message from the routing
713 * socket indicating that a redirect has occured, a routing lookup
714 * has failed, or that a protocol has detected timeouts to a particular
715 * destination.
716 */
717 void
718 rt_missmsg(type, rtinfo, flags, error)
719 int type, flags, error;
720 struct rt_addrinfo *rtinfo;
721 {
722 struct rt_msghdr rtm;
723 struct mbuf *m;
724 struct sockaddr *sa = rtinfo->rti_info[RTAX_DST];
725
726 if (route_cb.any_count == 0)
727 return;
728 bzero(&rtm, sizeof(rtm));
729 rtm.rtm_flags = RTF_DONE | flags;
730 rtm.rtm_errno = error;
731 m = rt_msg1(type, rtinfo, (caddr_t)&rtm, sizeof(rtm));
732 if (m == 0)
733 return;
734 mtod(m, struct rt_msghdr *)->rtm_addrs = rtinfo->rti_addrs;
735 route_proto.sp_protocol = sa ? sa->sa_family : 0;
736 raw_input(m, &route_proto, &route_src, &route_dst);
737 }
738
739 /*
740 * This routine is called to generate a message from the routing
741 * socket indicating that the status of a network interface has changed.
742 */
743 void
744 rt_ifmsg(ifp)
745 struct ifnet *ifp;
746 {
747 struct if_msghdr ifm;
748 #ifdef COMPAT_14
749 struct if_msghdr14 oifm;
750 #endif
751 struct mbuf *m;
752 struct rt_addrinfo info;
753
754 if (route_cb.any_count == 0)
755 return;
756 bzero(&info, sizeof(info));
757 bzero(&ifm, sizeof(ifm));
758 ifm.ifm_index = ifp->if_index;
759 ifm.ifm_flags = ifp->if_flags;
760 ifm.ifm_data = ifp->if_data;
761 ifm.ifm_addrs = 0;
762 m = rt_msg1(RTM_IFINFO, &info, (caddr_t)&ifm, sizeof(ifm));
763 if (m == 0)
764 return;
765 route_proto.sp_protocol = 0;
766 raw_input(m, &route_proto, &route_src, &route_dst);
767 #ifdef COMPAT_14
768 bzero(&info, sizeof(info));
769 bzero(&oifm, sizeof(oifm));
770 oifm.ifm_index = ifp->if_index;
771 oifm.ifm_flags = ifp->if_flags;
772 oifm.ifm_data.ifi_type = ifp->if_data.ifi_type;
773 oifm.ifm_data.ifi_addrlen = ifp->if_data.ifi_addrlen;
774 oifm.ifm_data.ifi_hdrlen = ifp->if_data.ifi_hdrlen;
775 oifm.ifm_data.ifi_mtu = ifp->if_data.ifi_mtu;
776 oifm.ifm_data.ifi_metric = ifp->if_data.ifi_metric;
777 oifm.ifm_data.ifi_baudrate = ifp->if_data.ifi_baudrate;
778 oifm.ifm_data.ifi_ipackets = ifp->if_data.ifi_ipackets;
779 oifm.ifm_data.ifi_ierrors = ifp->if_data.ifi_ierrors;
780 oifm.ifm_data.ifi_opackets = ifp->if_data.ifi_opackets;
781 oifm.ifm_data.ifi_oerrors = ifp->if_data.ifi_oerrors;
782 oifm.ifm_data.ifi_collisions = ifp->if_data.ifi_collisions;
783 oifm.ifm_data.ifi_ibytes = ifp->if_data.ifi_ibytes;
784 oifm.ifm_data.ifi_obytes = ifp->if_data.ifi_obytes;
785 oifm.ifm_data.ifi_imcasts = ifp->if_data.ifi_imcasts;
786 oifm.ifm_data.ifi_omcasts = ifp->if_data.ifi_omcasts;
787 oifm.ifm_data.ifi_iqdrops = ifp->if_data.ifi_iqdrops;
788 oifm.ifm_data.ifi_noproto = ifp->if_data.ifi_noproto;
789 oifm.ifm_data.ifi_lastchange = ifp->if_data.ifi_lastchange;
790 oifm.ifm_addrs = 0;
791 m = rt_msg1(RTM_OIFINFO, &info, (caddr_t)&oifm, sizeof(oifm));
792 if (m == 0)
793 return;
794 route_proto.sp_protocol = 0;
795 raw_input(m, &route_proto, &route_src, &route_dst);
796 #endif
797 }
798
799 /*
800 * This is called to generate messages from the routing socket
801 * indicating a network interface has had addresses associated with it.
802 * if we ever reverse the logic and replace messages TO the routing
803 * socket indicate a request to configure interfaces, then it will
804 * be unnecessary as the routing socket will automatically generate
805 * copies of it.
806 */
807 void
808 rt_newaddrmsg(cmd, ifa, error, rt)
809 int cmd, error;
810 struct ifaddr *ifa;
811 struct rtentry *rt;
812 {
813 struct rt_addrinfo info;
814 struct sockaddr *sa = NULL;
815 int pass;
816 struct mbuf *m = NULL;
817 struct ifnet *ifp = ifa->ifa_ifp;
818
819 if (route_cb.any_count == 0)
820 return;
821 for (pass = 1; pass < 3; pass++) {
822 bzero(&info, sizeof(info));
823 if ((cmd == RTM_ADD && pass == 1) ||
824 (cmd == RTM_DELETE && pass == 2)) {
825 struct ifa_msghdr ifam;
826 int ncmd = cmd == RTM_ADD ? RTM_NEWADDR : RTM_DELADDR;
827
828 ifaaddr = sa = ifa->ifa_addr;
829 ifpaddr = ifp->if_addrlist.tqh_first->ifa_addr;
830 netmask = ifa->ifa_netmask;
831 brdaddr = ifa->ifa_dstaddr;
832 bzero(&ifam, sizeof(ifam));
833 ifam.ifam_index = ifp->if_index;
834 ifam.ifam_metric = ifa->ifa_metric;
835 ifam.ifam_flags = ifa->ifa_flags;
836 m = rt_msg1(ncmd, &info, (caddr_t)&ifam, sizeof(ifam));
837 if (m == NULL)
838 continue;
839 mtod(m, struct ifa_msghdr *)->ifam_addrs =
840 info.rti_addrs;
841 }
842 if ((cmd == RTM_ADD && pass == 2) ||
843 (cmd == RTM_DELETE && pass == 1)) {
844 struct rt_msghdr rtm;
845
846 if (rt == 0)
847 continue;
848 netmask = rt_mask(rt);
849 dst = sa = rt_key(rt);
850 gate = rt->rt_gateway;
851 bzero(&rtm, sizeof(rtm));
852 rtm.rtm_index = ifp->if_index;
853 rtm.rtm_flags |= rt->rt_flags;
854 rtm.rtm_errno = error;
855 m = rt_msg1(cmd, &info, (caddr_t)&rtm, sizeof(rtm));
856 if (m == NULL)
857 continue;
858 mtod(m, struct rt_msghdr *)->rtm_addrs = info.rti_addrs;
859 }
860 route_proto.sp_protocol = sa ? sa->sa_family : 0;
861 raw_input(m, &route_proto, &route_src, &route_dst);
862 }
863 }
864
865 /*
866 * This is called to generate routing socket messages indicating
867 * network interface arrival and departure.
868 */
869 void
870 rt_ifannouncemsg(ifp, what)
871 struct ifnet *ifp;
872 int what;
873 {
874 struct if_announcemsghdr ifan;
875 struct mbuf *m;
876 struct rt_addrinfo info;
877
878 if (route_cb.any_count == 0)
879 return;
880 bzero(&info, sizeof(info));
881 bzero(&ifan, sizeof(ifan));
882 ifan.ifan_index = ifp->if_index;
883 strcpy(ifan.ifan_name, ifp->if_xname);
884 ifan.ifan_what = what;
885 m = rt_msg1(RTM_IFANNOUNCE, &info, (caddr_t)&ifan, sizeof(ifan));
886 if (m == 0)
887 return;
888 route_proto.sp_protocol = 0;
889 raw_input(m, &route_proto, &route_src, &route_dst);
890 }
891
892 /*
893 * This is used in dumping the kernel table via sysctl().
894 */
895 static int
896 sysctl_dumpentry(rn, v)
897 struct radix_node *rn;
898 void *v;
899 {
900 struct walkarg *w = v;
901 struct rtentry *rt = (struct rtentry *)rn;
902 int error = 0, size;
903 struct rt_addrinfo info;
904
905 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg))
906 return 0;
907 bzero(&info, sizeof(info));
908 dst = rt_key(rt);
909 gate = rt->rt_gateway;
910 netmask = rt_mask(rt);
911 genmask = rt->rt_genmask;
912 if (rt->rt_ifp) {
913 ifpaddr = rt->rt_ifp->if_addrlist.tqh_first->ifa_addr;
914 ifaaddr = rt->rt_ifa->ifa_addr;
915 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT)
916 brdaddr = rt->rt_ifa->ifa_dstaddr;
917 }
918 if ((error = rt_msg2(RTM_GET, &info, 0, w, &size)))
919 return (error);
920 if (w->w_where && w->w_tmem && w->w_needed <= 0) {
921 struct rt_msghdr *rtm = (struct rt_msghdr *)w->w_tmem;
922
923 rtm->rtm_flags = rt->rt_flags;
924 rtm->rtm_use = rt->rt_use;
925 rtm->rtm_rmx = rt->rt_rmx;
926 rtm->rtm_index = rt->rt_ifp->if_index;
927 rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0;
928 rtm->rtm_addrs = info.rti_addrs;
929 if ((error = copyout(rtm, w->w_where, size)) != 0)
930 w->w_where = NULL;
931 else
932 w->w_where += size;
933 }
934 return (error);
935 }
936
937 static int
938 sysctl_iflist(af, w, type)
939 int af;
940 struct walkarg *w;
941 int type;
942 {
943 struct ifnet *ifp;
944 struct ifaddr *ifa;
945 struct rt_addrinfo info;
946 int len, error = 0;
947
948 bzero(&info, sizeof(info));
949 for (ifp = ifnet.tqh_first; ifp != 0; ifp = ifp->if_list.tqe_next) {
950 if (w->w_arg && w->w_arg != ifp->if_index)
951 continue;
952 ifa = ifp->if_addrlist.tqh_first;
953 ifpaddr = ifa->ifa_addr;
954 switch(type) {
955 case NET_RT_IFLIST:
956 error =
957 rt_msg2(RTM_IFINFO, &info, (caddr_t)0, w, &len);
958 break;
959 #ifdef COMPAT_14
960 case NET_RT_OIFLIST:
961 error =
962 rt_msg2(RTM_OIFINFO, &info, (caddr_t)0, w, &len);
963 break;
964 #endif
965 default:
966 panic("sysctl_iflist(1)");
967 }
968 if (error)
969 return (error);
970 ifpaddr = 0;
971 if (w->w_where && w->w_tmem && w->w_needed <= 0) {
972 switch(type) {
973 case NET_RT_IFLIST: {
974 struct if_msghdr *ifm;
975
976 ifm = (struct if_msghdr *)w->w_tmem;
977 ifm->ifm_index = ifp->if_index;
978 ifm->ifm_flags = ifp->if_flags;
979 ifm->ifm_data = ifp->if_data;
980 ifm->ifm_addrs = info.rti_addrs;
981 error = copyout(ifm, w->w_where, len);
982 if (error)
983 return (error);
984 w->w_where += len;
985 break;
986 }
987
988 #ifdef COMPAT_14
989 case NET_RT_OIFLIST: {
990 struct if_msghdr14 *ifm;
991
992 ifm = (struct if_msghdr14 *)w->w_tmem;
993 ifm->ifm_index = ifp->if_index;
994 ifm->ifm_flags = ifp->if_flags;
995 ifm->ifm_data.ifi_type = ifp->if_data.ifi_type;
996 ifm->ifm_data.ifi_addrlen =
997 ifp->if_data.ifi_addrlen;
998 ifm->ifm_data.ifi_hdrlen =
999 ifp->if_data.ifi_hdrlen;
1000 ifm->ifm_data.ifi_mtu = ifp->if_data.ifi_mtu;
1001 ifm->ifm_data.ifi_metric =
1002 ifp->if_data.ifi_metric;
1003 ifm->ifm_data.ifi_baudrate =
1004 ifp->if_data.ifi_baudrate;
1005 ifm->ifm_data.ifi_ipackets =
1006 ifp->if_data.ifi_ipackets;
1007 ifm->ifm_data.ifi_ierrors =
1008 ifp->if_data.ifi_ierrors;
1009 ifm->ifm_data.ifi_opackets =
1010 ifp->if_data.ifi_opackets;
1011 ifm->ifm_data.ifi_oerrors =
1012 ifp->if_data.ifi_oerrors;
1013 ifm->ifm_data.ifi_collisions =
1014 ifp->if_data.ifi_collisions;
1015 ifm->ifm_data.ifi_ibytes =
1016 ifp->if_data.ifi_ibytes;
1017 ifm->ifm_data.ifi_obytes =
1018 ifp->if_data.ifi_obytes;
1019 ifm->ifm_data.ifi_imcasts =
1020 ifp->if_data.ifi_imcasts;
1021 ifm->ifm_data.ifi_omcasts =
1022 ifp->if_data.ifi_omcasts;
1023 ifm->ifm_data.ifi_iqdrops =
1024 ifp->if_data.ifi_iqdrops;
1025 ifm->ifm_data.ifi_noproto =
1026 ifp->if_data.ifi_noproto;
1027 ifm->ifm_data.ifi_lastchange =
1028 ifp->if_data.ifi_lastchange;
1029 ifm->ifm_addrs = info.rti_addrs;
1030 error = copyout(ifm, w->w_where, len);
1031 if (error)
1032 return (error);
1033 w->w_where += len;
1034 break;
1035 }
1036 #endif
1037 default:
1038 panic("sysctl_iflist(2)");
1039 }
1040 }
1041 while ((ifa = ifa->ifa_list.tqe_next) != NULL) {
1042 if (af && af != ifa->ifa_addr->sa_family)
1043 continue;
1044 ifaaddr = ifa->ifa_addr;
1045 netmask = ifa->ifa_netmask;
1046 brdaddr = ifa->ifa_dstaddr;
1047 if ((error = rt_msg2(RTM_NEWADDR, &info, 0, w, &len)))
1048 return (error);
1049 if (w->w_where && w->w_tmem && w->w_needed <= 0) {
1050 struct ifa_msghdr *ifam;
1051
1052 ifam = (struct ifa_msghdr *)w->w_tmem;
1053 ifam->ifam_index = ifa->ifa_ifp->if_index;
1054 ifam->ifam_flags = ifa->ifa_flags;
1055 ifam->ifam_metric = ifa->ifa_metric;
1056 ifam->ifam_addrs = info.rti_addrs;
1057 error = copyout(w->w_tmem, w->w_where, len);
1058 if (error)
1059 return (error);
1060 w->w_where += len;
1061 }
1062 }
1063 ifaaddr = netmask = brdaddr = 0;
1064 }
1065 return (0);
1066 }
1067
1068 static int
1069 sysctl_rtable(name, namelen, where, given, new, newlen)
1070 int *name;
1071 u_int namelen;
1072 void *where;
1073 size_t *given;
1074 void *new;
1075 size_t newlen;
1076 {
1077 struct radix_node_head *rnh;
1078 int i, s, error = EINVAL;
1079 u_char af;
1080 struct walkarg w;
1081
1082 if (new)
1083 return (EPERM);
1084 if (namelen != 3)
1085 return (EINVAL);
1086 af = name[0];
1087 w.w_tmemneeded = 0;
1088 w.w_tmemsize = 0;
1089 w.w_tmem = NULL;
1090 again:
1091 /* we may return here if a later [re]alloc of the t_mem buffer fails */
1092 if (w.w_tmemneeded) {
1093 w.w_tmem = (caddr_t) malloc(w.w_tmemneeded, M_RTABLE, M_WAITOK);
1094 w.w_tmemsize = w.w_tmemneeded;
1095 w.w_tmemneeded = 0;
1096 }
1097 w.w_op = name[1];
1098 w.w_arg = name[2];
1099 w.w_given = *given;
1100 w.w_needed = 0 - w.w_given;
1101 w.w_where = where;
1102
1103 s = splsoftnet();
1104 switch (w.w_op) {
1105
1106 case NET_RT_DUMP:
1107 case NET_RT_FLAGS:
1108 for (i = 1; i <= AF_MAX; i++)
1109 if ((rnh = rt_tables[i]) && (af == 0 || af == i) &&
1110 (error = (*rnh->rnh_walktree)(rnh,
1111 sysctl_dumpentry, &w)))
1112 break;
1113 break;
1114
1115 #ifdef COMPAT_14
1116 case NET_RT_OIFLIST:
1117 error = sysctl_iflist(af, &w, w.w_op);
1118 break;
1119 #endif
1120
1121 case NET_RT_IFLIST:
1122 error = sysctl_iflist(af, &w, w.w_op);
1123 }
1124 splx(s);
1125
1126 /* check to see if we couldn't allocate memory with NOWAIT */
1127 if (error == ENOBUFS && w.w_tmem == 0 && w.w_tmemneeded)
1128 goto again;
1129
1130 if (w.w_tmem)
1131 free(w.w_tmem, M_RTABLE);
1132 w.w_needed += w.w_given;
1133 if (where) {
1134 *given = w.w_where - (caddr_t) where;
1135 if (*given < w.w_needed)
1136 return (ENOMEM);
1137 } else {
1138 *given = (11 * w.w_needed) / 10;
1139 }
1140 return (error);
1141 }
1142
1143 /*
1144 * Definitions of protocols supported in the ROUTE domain.
1145 */
1146
1147 extern struct domain routedomain; /* or at least forward */
1148
1149 struct protosw routesw[] = {
1150 { SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR,
1151 raw_input, route_output, raw_ctlinput, 0,
1152 route_usrreq,
1153 raw_init, 0, 0, 0,
1154 sysctl_rtable,
1155 }
1156 };
1157
1158 struct domain routedomain =
1159 { PF_ROUTE, "route", route_init, 0, 0,
1160 routesw, &routesw[sizeof(routesw)/sizeof(routesw[0])] };
1161