rtsock.c revision 1.55 1 /* $NetBSD: rtsock.c,v 1.55 2002/02/22 17:26:31 christos 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 <sys/cdefs.h>
68 __KERNEL_RCSID(0, "$NetBSD: rtsock.c,v 1.55 2002/02/22 17:26:31 christos Exp $");
69
70 #include "opt_inet.h"
71
72 #include <sys/param.h>
73 #include <sys/systm.h>
74 #include <sys/proc.h>
75 #include <sys/mbuf.h>
76 #include <sys/socket.h>
77 #include <sys/socketvar.h>
78 #include <sys/domain.h>
79 #include <sys/protosw.h>
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 memset(so->so_pcb, 0, 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 sa_family_t family;
221
222 va_start(ap, m);
223 so = va_arg(ap, struct socket *);
224 va_end(ap);
225
226 #define senderr(e) do { error = e; goto flush;} while (0)
227 if (m == 0 || ((m->m_len < sizeof(int32_t)) &&
228 (m = m_pullup(m, sizeof(int32_t))) == 0))
229 return (ENOBUFS);
230 if ((m->m_flags & M_PKTHDR) == 0)
231 panic("route_output");
232 len = m->m_pkthdr.len;
233 if (len < sizeof(*rtm) ||
234 len != mtod(m, struct rt_msghdr *)->rtm_msglen) {
235 dst = 0;
236 senderr(EINVAL);
237 }
238 R_Malloc(rtm, struct rt_msghdr *, len);
239 if (rtm == 0) {
240 dst = 0;
241 senderr(ENOBUFS);
242 }
243 m_copydata(m, 0, len, (caddr_t)rtm);
244 if (rtm->rtm_version != RTM_VERSION) {
245 dst = 0;
246 senderr(EPROTONOSUPPORT);
247 }
248 rtm->rtm_pid = curproc->p_pid;
249 memset(&info, 0, sizeof(info));
250 info.rti_addrs = rtm->rtm_addrs;
251 if (rt_xaddrs((caddr_t)(rtm + 1), len + (caddr_t)rtm, &info))
252 senderr(EINVAL);
253 info.rti_flags = rtm->rtm_flags;
254 if (dst == 0 || (dst->sa_family >= AF_MAX))
255 senderr(EINVAL);
256 if (gate != 0 && (gate->sa_family >= AF_MAX))
257 senderr(EINVAL);
258 if (genmask) {
259 struct radix_node *t;
260 t = rn_addmask((caddr_t)genmask, 0, 1);
261 if (t && genmask->sa_len >= ((struct sockaddr *)t->rn_key)->sa_len &&
262 Bcmp((caddr_t *)genmask + 1, (caddr_t *)t->rn_key + 1,
263 ((struct sockaddr *)t->rn_key)->sa_len) - 1)
264 genmask = (struct sockaddr *)(t->rn_key);
265 else
266 senderr(ENOBUFS);
267 }
268
269 /*
270 * Verify that the caller has the appropriate privilege; RTM_GET
271 * is the only operation the non-superuser is allowed.
272 */
273 if (rtm->rtm_type != RTM_GET &&
274 suser(curproc->p_ucred, &curproc->p_acflag) != 0)
275 senderr(EACCES);
276
277 switch (rtm->rtm_type) {
278
279 case RTM_ADD:
280 if (gate == 0)
281 senderr(EINVAL);
282 error = rtrequest1(rtm->rtm_type, &info, &saved_nrt);
283 if (error == 0 && saved_nrt) {
284 rt_setmetrics(rtm->rtm_inits,
285 &rtm->rtm_rmx, &saved_nrt->rt_rmx);
286 saved_nrt->rt_refcnt--;
287 saved_nrt->rt_genmask = genmask;
288 }
289 break;
290
291 case RTM_DELETE:
292 error = rtrequest1(rtm->rtm_type, &info, &saved_nrt);
293 if (error == 0) {
294 (rt = saved_nrt)->rt_refcnt++;
295 goto report;
296 }
297 break;
298
299 case RTM_GET:
300 case RTM_CHANGE:
301 case RTM_LOCK:
302 if ((rnh = rt_tables[dst->sa_family]) == 0) {
303 senderr(EAFNOSUPPORT);
304 }
305 rn = rnh->rnh_lookup(dst, netmask, rnh);
306 if (rn == NULL || (rn->rn_flags & RNF_ROOT) != 0) {
307 senderr(ESRCH);
308 }
309 rt = (struct rtentry *)rn;
310 rt->rt_refcnt++;
311
312 switch(rtm->rtm_type) {
313
314 case RTM_GET:
315 report:
316 dst = rt_key(rt);
317 gate = rt->rt_gateway;
318 netmask = rt_mask(rt);
319 genmask = rt->rt_genmask;
320 if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) {
321 if ((ifp = rt->rt_ifp) != NULL) {
322 ifpaddr = TAILQ_FIRST(&ifp->if_addrlist)->ifa_addr;
323 ifaaddr = rt->rt_ifa->ifa_addr;
324 if (ifp->if_flags & IFF_POINTOPOINT)
325 brdaddr = rt->rt_ifa->ifa_dstaddr;
326 else
327 brdaddr = 0;
328 rtm->rtm_index = ifp->if_index;
329 } else {
330 ifpaddr = 0;
331 ifaaddr = 0;
332 }
333 }
334 (void)rt_msg2(rtm->rtm_type, &info, (caddr_t)0,
335 (struct walkarg *)0, &len);
336 if (len > rtm->rtm_msglen) {
337 struct rt_msghdr *new_rtm;
338 R_Malloc(new_rtm, struct rt_msghdr *, len);
339 if (new_rtm == 0)
340 senderr(ENOBUFS);
341 Bcopy(rtm, new_rtm, rtm->rtm_msglen);
342 Free(rtm); rtm = new_rtm;
343 }
344 (void)rt_msg2(rtm->rtm_type, &info, (caddr_t)rtm,
345 (struct walkarg *)0, 0);
346 rtm->rtm_flags = rt->rt_flags;
347 rtm->rtm_rmx = rt->rt_rmx;
348 rtm->rtm_addrs = info.rti_addrs;
349 break;
350
351 case RTM_CHANGE:
352 /*
353 * new gateway could require new ifaddr, ifp;
354 * flags may also be different; ifp may be specified
355 * by ll sockaddr when protocol address is ambiguous
356 */
357 if ((error = rt_getifa(&info)) != 0)
358 senderr(error);
359 if (gate && rt_setgate(rt, rt_key(rt), gate))
360 senderr(EDQUOT);
361 /* new gateway could require new ifaddr, ifp;
362 flags may also be different; ifp may be specified
363 by ll sockaddr when protocol address is ambiguous */
364 if (ifpaddr && (ifa = ifa_ifwithnet(ifpaddr)) &&
365 (ifp = ifa->ifa_ifp) && (ifaaddr || gate))
366 ifa = ifaof_ifpforaddr(ifaaddr ? ifaaddr : gate,
367 ifp);
368 else if ((ifaaddr && (ifa = ifa_ifwithaddr(ifaaddr))) ||
369 (gate && (ifa = ifa_ifwithroute(rt->rt_flags,
370 rt_key(rt), gate))))
371 ifp = ifa->ifa_ifp;
372 if (ifa) {
373 struct ifaddr *oifa = rt->rt_ifa;
374 if (oifa != ifa) {
375 if (oifa && oifa->ifa_rtrequest)
376 oifa->ifa_rtrequest(RTM_DELETE, rt,
377 &info);
378 IFAFREE(rt->rt_ifa);
379 rt->rt_ifa = ifa;
380 IFAREF(rt->rt_ifa);
381 rt->rt_ifp = ifp;
382 }
383 }
384 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx,
385 &rt->rt_rmx);
386 if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest)
387 rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, &info);
388 if (genmask)
389 rt->rt_genmask = genmask;
390 /*
391 * Fall into
392 */
393 case RTM_LOCK:
394 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits);
395 rt->rt_rmx.rmx_locks |=
396 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks);
397 break;
398 }
399 break;
400
401 default:
402 senderr(EOPNOTSUPP);
403 }
404
405 flush:
406 if (rtm) {
407 if (error)
408 rtm->rtm_errno = error;
409 else
410 rtm->rtm_flags |= RTF_DONE;
411 }
412 family = dst ? dst->sa_family : 0;
413 if (rt)
414 rtfree(rt);
415 {
416 struct rawcb *rp = 0;
417 /*
418 * Check to see if we don't want our own messages.
419 */
420 if ((so->so_options & SO_USELOOPBACK) == 0) {
421 if (route_cb.any_count <= 1) {
422 if (rtm)
423 Free(rtm);
424 m_freem(m);
425 return (error);
426 }
427 /* There is another listener, so construct message */
428 rp = sotorawcb(so);
429 }
430 if (rtm) {
431 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm);
432 if (m->m_pkthdr.len < rtm->rtm_msglen) {
433 m_freem(m);
434 m = NULL;
435 } else if (m->m_pkthdr.len > rtm->rtm_msglen)
436 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len);
437 Free(rtm);
438 }
439 if (rp)
440 rp->rcb_proto.sp_family = 0; /* Avoid us */
441 if (family)
442 route_proto.sp_protocol = family;
443 if (m)
444 raw_input(m, &route_proto, &route_src, &route_dst);
445 if (rp)
446 rp->rcb_proto.sp_family = PF_ROUTE;
447 }
448 return (error);
449 }
450
451 void
452 rt_setmetrics(which, in, out)
453 u_long which;
454 struct rt_metrics *in, *out;
455 {
456 #define metric(f, e) if (which & (f)) out->e = in->e;
457 metric(RTV_RPIPE, rmx_recvpipe);
458 metric(RTV_SPIPE, rmx_sendpipe);
459 metric(RTV_SSTHRESH, rmx_ssthresh);
460 metric(RTV_RTT, rmx_rtt);
461 metric(RTV_RTTVAR, rmx_rttvar);
462 metric(RTV_HOPCOUNT, rmx_hopcount);
463 metric(RTV_MTU, rmx_mtu);
464 metric(RTV_EXPIRE, rmx_expire);
465 #undef metric
466 }
467
468 #define ROUNDUP(a) \
469 ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long))
470 #define ADVANCE(x, n) (x += ROUNDUP((n)->sa_len))
471
472 static int
473 rt_xaddrs(cp, cplim, rtinfo)
474 caddr_t cp, cplim;
475 struct rt_addrinfo *rtinfo;
476 {
477 struct sockaddr *sa;
478 int i;
479
480 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) {
481 if ((rtinfo->rti_addrs & (1 << i)) == 0)
482 continue;
483 rtinfo->rti_info[i] = sa = (struct sockaddr *)cp;
484 ADVANCE(cp, sa);
485 }
486
487 /* Check for extra addresses specified. */
488 if ((rtinfo->rti_addrs & (~0 << i)) != 0)
489 return (1);
490 /* Check for bad data length. */
491 if (cp != cplim) {
492 if (i == RTAX_NETMASK + 1 &&
493 cp - ROUNDUP(sa->sa_len) + sa->sa_len == cplim)
494 /*
495 * The last sockaddr was netmask.
496 * We accept this for now for the sake of old
497 * binaries or third party softwares.
498 */
499 ;
500 else
501 return (1);
502 }
503 return (0);
504 }
505
506 static struct mbuf *
507 rt_msg1(type, rtinfo, data, datalen)
508 int type;
509 struct rt_addrinfo *rtinfo;
510 caddr_t data;
511 int datalen;
512 {
513 struct rt_msghdr *rtm;
514 struct mbuf *m;
515 int i;
516 struct sockaddr *sa;
517 int len, dlen;
518
519 m = m_gethdr(M_DONTWAIT, MT_DATA);
520 if (m == 0)
521 return (m);
522 switch (type) {
523
524 case RTM_DELADDR:
525 case RTM_NEWADDR:
526 len = sizeof(struct ifa_msghdr);
527 break;
528
529 #ifdef COMPAT_14
530 case RTM_OIFINFO:
531 len = sizeof(struct if_msghdr14);
532 break;
533 #endif
534
535 case RTM_IFINFO:
536 len = sizeof(struct if_msghdr);
537 break;
538
539 case RTM_IFANNOUNCE:
540 len = sizeof(struct if_announcemsghdr);
541 break;
542
543 default:
544 len = sizeof(struct rt_msghdr);
545 }
546 if (len > MHLEN + MLEN)
547 panic("rt_msg1: message too long");
548 else if (len > MHLEN) {
549 m->m_next = m_get(M_DONTWAIT, MT_DATA);
550 if (m->m_next == NULL) {
551 m_freem(m);
552 return (NULL);
553 }
554 m->m_pkthdr.len = len;
555 m->m_len = MHLEN;
556 m->m_next->m_len = len - MHLEN;
557 } else {
558 m->m_pkthdr.len = m->m_len = len;
559 }
560 m->m_pkthdr.rcvif = 0;
561 m_copyback(m, 0, datalen, data);
562 rtm = mtod(m, struct rt_msghdr *);
563 for (i = 0; i < RTAX_MAX; i++) {
564 if ((sa = rtinfo->rti_info[i]) == NULL)
565 continue;
566 rtinfo->rti_addrs |= (1 << i);
567 dlen = ROUNDUP(sa->sa_len);
568 m_copyback(m, len, dlen, (caddr_t)sa);
569 len += dlen;
570 }
571 if (m->m_pkthdr.len != len) {
572 m_freem(m);
573 return (NULL);
574 }
575 rtm->rtm_msglen = len;
576 rtm->rtm_version = RTM_VERSION;
577 rtm->rtm_type = type;
578 return (m);
579 }
580
581 /*
582 * rt_msg2
583 *
584 * fills 'cp' or 'w'.w_tmem with the routing socket message and
585 * returns the length of the message in 'lenp'.
586 *
587 * if walkarg is 0, cp is expected to be 0 or a buffer large enough to hold
588 * the message
589 * otherwise walkarg's w_needed is updated and if the user buffer is
590 * specified and w_needed indicates space exists the information is copied
591 * into the temp space (w_tmem). w_tmem is [re]allocated if necessary,
592 * if the allocation fails ENOBUFS is returned.
593 */
594 static int
595 rt_msg2(type, rtinfo, cp, w, lenp)
596 int type;
597 struct rt_addrinfo *rtinfo;
598 caddr_t cp;
599 struct walkarg *w;
600 int *lenp;
601 {
602 int i;
603 int len, dlen, second_time = 0;
604 caddr_t cp0;
605
606 rtinfo->rti_addrs = 0;
607 again:
608 switch (type) {
609
610 case RTM_DELADDR:
611 case RTM_NEWADDR:
612 len = sizeof(struct ifa_msghdr);
613 break;
614 #ifdef COMPAT_14
615 case RTM_OIFINFO:
616 len = sizeof(struct if_msghdr14);
617 break;
618 #endif
619
620 case RTM_IFINFO:
621 len = sizeof(struct if_msghdr);
622 break;
623
624 default:
625 len = sizeof(struct rt_msghdr);
626 }
627 if ((cp0 = cp) != NULL)
628 cp += len;
629 for (i = 0; i < RTAX_MAX; i++) {
630 struct sockaddr *sa;
631
632 if ((sa = rtinfo->rti_info[i]) == 0)
633 continue;
634 rtinfo->rti_addrs |= (1 << i);
635 dlen = ROUNDUP(sa->sa_len);
636 if (cp) {
637 bcopy(sa, cp, (unsigned)dlen);
638 cp += dlen;
639 }
640 len += dlen;
641 }
642 if (cp == 0 && w != NULL && !second_time) {
643 struct walkarg *rw = w;
644
645 rw->w_needed += len;
646 if (rw->w_needed <= 0 && rw->w_where) {
647 if (rw->w_tmemsize < len) {
648 if (rw->w_tmem)
649 free(rw->w_tmem, M_RTABLE);
650 rw->w_tmem = (caddr_t) malloc(len, M_RTABLE,
651 M_NOWAIT);
652 if (rw->w_tmem)
653 rw->w_tmemsize = len;
654 }
655 if (rw->w_tmem) {
656 cp = rw->w_tmem;
657 second_time = 1;
658 goto again;
659 } else {
660 rw->w_tmemneeded = len;
661 return (ENOBUFS);
662 }
663 }
664 }
665 if (cp) {
666 struct rt_msghdr *rtm = (struct rt_msghdr *)cp0;
667
668 rtm->rtm_version = RTM_VERSION;
669 rtm->rtm_type = type;
670 rtm->rtm_msglen = len;
671 }
672 if (lenp)
673 *lenp = len;
674 return (0);
675 }
676
677 /*
678 * This routine is called to generate a message from the routing
679 * socket indicating that a redirect has occurred, a routing lookup
680 * has failed, or that a protocol has detected timeouts to a particular
681 * destination.
682 */
683 void
684 rt_missmsg(type, rtinfo, flags, error)
685 int type, flags, error;
686 struct rt_addrinfo *rtinfo;
687 {
688 struct rt_msghdr rtm;
689 struct mbuf *m;
690 struct sockaddr *sa = rtinfo->rti_info[RTAX_DST];
691
692 if (route_cb.any_count == 0)
693 return;
694 memset(&rtm, 0, sizeof(rtm));
695 rtm.rtm_flags = RTF_DONE | flags;
696 rtm.rtm_errno = error;
697 m = rt_msg1(type, rtinfo, (caddr_t)&rtm, sizeof(rtm));
698 if (m == 0)
699 return;
700 mtod(m, struct rt_msghdr *)->rtm_addrs = rtinfo->rti_addrs;
701 route_proto.sp_protocol = sa ? sa->sa_family : 0;
702 raw_input(m, &route_proto, &route_src, &route_dst);
703 }
704
705 /*
706 * This routine is called to generate a message from the routing
707 * socket indicating that the status of a network interface has changed.
708 */
709 void
710 rt_ifmsg(ifp)
711 struct ifnet *ifp;
712 {
713 struct if_msghdr ifm;
714 #ifdef COMPAT_14
715 struct if_msghdr14 oifm;
716 #endif
717 struct mbuf *m;
718 struct rt_addrinfo info;
719
720 if (route_cb.any_count == 0)
721 return;
722 memset(&info, 0, sizeof(info));
723 memset(&ifm, 0, sizeof(ifm));
724 ifm.ifm_index = ifp->if_index;
725 ifm.ifm_flags = ifp->if_flags;
726 ifm.ifm_data = ifp->if_data;
727 ifm.ifm_addrs = 0;
728 m = rt_msg1(RTM_IFINFO, &info, (caddr_t)&ifm, sizeof(ifm));
729 if (m == 0)
730 return;
731 route_proto.sp_protocol = 0;
732 raw_input(m, &route_proto, &route_src, &route_dst);
733 #ifdef COMPAT_14
734 memset(&info, 0, sizeof(info));
735 memset(&oifm, 0, sizeof(oifm));
736 oifm.ifm_index = ifp->if_index;
737 oifm.ifm_flags = ifp->if_flags;
738 oifm.ifm_data.ifi_type = ifp->if_data.ifi_type;
739 oifm.ifm_data.ifi_addrlen = ifp->if_data.ifi_addrlen;
740 oifm.ifm_data.ifi_hdrlen = ifp->if_data.ifi_hdrlen;
741 oifm.ifm_data.ifi_mtu = ifp->if_data.ifi_mtu;
742 oifm.ifm_data.ifi_metric = ifp->if_data.ifi_metric;
743 oifm.ifm_data.ifi_baudrate = ifp->if_data.ifi_baudrate;
744 oifm.ifm_data.ifi_ipackets = ifp->if_data.ifi_ipackets;
745 oifm.ifm_data.ifi_ierrors = ifp->if_data.ifi_ierrors;
746 oifm.ifm_data.ifi_opackets = ifp->if_data.ifi_opackets;
747 oifm.ifm_data.ifi_oerrors = ifp->if_data.ifi_oerrors;
748 oifm.ifm_data.ifi_collisions = ifp->if_data.ifi_collisions;
749 oifm.ifm_data.ifi_ibytes = ifp->if_data.ifi_ibytes;
750 oifm.ifm_data.ifi_obytes = ifp->if_data.ifi_obytes;
751 oifm.ifm_data.ifi_imcasts = ifp->if_data.ifi_imcasts;
752 oifm.ifm_data.ifi_omcasts = ifp->if_data.ifi_omcasts;
753 oifm.ifm_data.ifi_iqdrops = ifp->if_data.ifi_iqdrops;
754 oifm.ifm_data.ifi_noproto = ifp->if_data.ifi_noproto;
755 oifm.ifm_data.ifi_lastchange = ifp->if_data.ifi_lastchange;
756 oifm.ifm_addrs = 0;
757 m = rt_msg1(RTM_OIFINFO, &info, (caddr_t)&oifm, sizeof(oifm));
758 if (m == 0)
759 return;
760 route_proto.sp_protocol = 0;
761 raw_input(m, &route_proto, &route_src, &route_dst);
762 #endif
763 }
764
765 /*
766 * This is called to generate messages from the routing socket
767 * indicating a network interface has had addresses associated with it.
768 * if we ever reverse the logic and replace messages TO the routing
769 * socket indicate a request to configure interfaces, then it will
770 * be unnecessary as the routing socket will automatically generate
771 * copies of it.
772 */
773 void
774 rt_newaddrmsg(cmd, ifa, error, rt)
775 int cmd, error;
776 struct ifaddr *ifa;
777 struct rtentry *rt;
778 {
779 struct rt_addrinfo info;
780 struct sockaddr *sa = NULL;
781 int pass;
782 struct mbuf *m = NULL;
783 struct ifnet *ifp = ifa->ifa_ifp;
784
785 if (route_cb.any_count == 0)
786 return;
787 for (pass = 1; pass < 3; pass++) {
788 memset(&info, 0, sizeof(info));
789 if ((cmd == RTM_ADD && pass == 1) ||
790 (cmd == RTM_DELETE && pass == 2)) {
791 struct ifa_msghdr ifam;
792 int ncmd = cmd == RTM_ADD ? RTM_NEWADDR : RTM_DELADDR;
793
794 ifaaddr = sa = ifa->ifa_addr;
795 ifpaddr = TAILQ_FIRST(&ifp->if_addrlist)->ifa_addr;
796 netmask = ifa->ifa_netmask;
797 brdaddr = ifa->ifa_dstaddr;
798 memset(&ifam, 0, sizeof(ifam));
799 ifam.ifam_index = ifp->if_index;
800 ifam.ifam_metric = ifa->ifa_metric;
801 ifam.ifam_flags = ifa->ifa_flags;
802 m = rt_msg1(ncmd, &info, (caddr_t)&ifam, sizeof(ifam));
803 if (m == NULL)
804 continue;
805 mtod(m, struct ifa_msghdr *)->ifam_addrs =
806 info.rti_addrs;
807 }
808 if ((cmd == RTM_ADD && pass == 2) ||
809 (cmd == RTM_DELETE && pass == 1)) {
810 struct rt_msghdr rtm;
811
812 if (rt == 0)
813 continue;
814 netmask = rt_mask(rt);
815 dst = sa = rt_key(rt);
816 gate = rt->rt_gateway;
817 memset(&rtm, 0, sizeof(rtm));
818 rtm.rtm_index = ifp->if_index;
819 rtm.rtm_flags |= rt->rt_flags;
820 rtm.rtm_errno = error;
821 m = rt_msg1(cmd, &info, (caddr_t)&rtm, sizeof(rtm));
822 if (m == NULL)
823 continue;
824 mtod(m, struct rt_msghdr *)->rtm_addrs = info.rti_addrs;
825 }
826 route_proto.sp_protocol = sa ? sa->sa_family : 0;
827 raw_input(m, &route_proto, &route_src, &route_dst);
828 }
829 }
830
831 /*
832 * This is called to generate routing socket messages indicating
833 * network interface arrival and departure.
834 */
835 void
836 rt_ifannouncemsg(ifp, what)
837 struct ifnet *ifp;
838 int what;
839 {
840 struct if_announcemsghdr ifan;
841 struct mbuf *m;
842 struct rt_addrinfo info;
843
844 if (route_cb.any_count == 0)
845 return;
846 memset(&info, 0, sizeof(info));
847 memset(&ifan, 0, sizeof(ifan));
848 ifan.ifan_index = ifp->if_index;
849 strcpy(ifan.ifan_name, ifp->if_xname);
850 ifan.ifan_what = what;
851 m = rt_msg1(RTM_IFANNOUNCE, &info, (caddr_t)&ifan, sizeof(ifan));
852 if (m == 0)
853 return;
854 route_proto.sp_protocol = 0;
855 raw_input(m, &route_proto, &route_src, &route_dst);
856 }
857
858 /*
859 * This is used in dumping the kernel table via sysctl().
860 */
861 static int
862 sysctl_dumpentry(rn, v)
863 struct radix_node *rn;
864 void *v;
865 {
866 struct walkarg *w = v;
867 struct rtentry *rt = (struct rtentry *)rn;
868 int error = 0, size;
869 struct rt_addrinfo info;
870
871 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg))
872 return 0;
873 memset(&info, 0, sizeof(info));
874 dst = rt_key(rt);
875 gate = rt->rt_gateway;
876 netmask = rt_mask(rt);
877 genmask = rt->rt_genmask;
878 if (rt->rt_ifp) {
879 ifpaddr = TAILQ_FIRST(&rt->rt_ifp->if_addrlist)->ifa_addr;
880 ifaaddr = rt->rt_ifa->ifa_addr;
881 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT)
882 brdaddr = rt->rt_ifa->ifa_dstaddr;
883 }
884 if ((error = rt_msg2(RTM_GET, &info, 0, w, &size)))
885 return (error);
886 if (w->w_where && w->w_tmem && w->w_needed <= 0) {
887 struct rt_msghdr *rtm = (struct rt_msghdr *)w->w_tmem;
888
889 rtm->rtm_flags = rt->rt_flags;
890 rtm->rtm_use = rt->rt_use;
891 rtm->rtm_rmx = rt->rt_rmx;
892 rtm->rtm_index = rt->rt_ifp->if_index;
893 rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0;
894 rtm->rtm_addrs = info.rti_addrs;
895 if ((error = copyout(rtm, w->w_where, size)) != 0)
896 w->w_where = NULL;
897 else
898 w->w_where += size;
899 }
900 return (error);
901 }
902
903 static int
904 sysctl_iflist(af, w, type)
905 int af;
906 struct walkarg *w;
907 int type;
908 {
909 struct ifnet *ifp;
910 struct ifaddr *ifa;
911 struct rt_addrinfo info;
912 int len, error = 0;
913
914 memset(&info, 0, sizeof(info));
915 TAILQ_FOREACH(ifp, &ifnet, if_list) {
916 if (w->w_arg && w->w_arg != ifp->if_index)
917 continue;
918 ifa = TAILQ_FIRST(&ifp->if_addrlist);
919 ifpaddr = ifa->ifa_addr;
920 switch(type) {
921 case NET_RT_IFLIST:
922 error =
923 rt_msg2(RTM_IFINFO, &info, (caddr_t)0, w, &len);
924 break;
925 #ifdef COMPAT_14
926 case NET_RT_OIFLIST:
927 error =
928 rt_msg2(RTM_OIFINFO, &info, (caddr_t)0, w, &len);
929 break;
930 #endif
931 default:
932 panic("sysctl_iflist(1)");
933 }
934 if (error)
935 return (error);
936 ifpaddr = 0;
937 if (w->w_where && w->w_tmem && w->w_needed <= 0) {
938 switch(type) {
939 case NET_RT_IFLIST: {
940 struct if_msghdr *ifm;
941
942 ifm = (struct if_msghdr *)w->w_tmem;
943 ifm->ifm_index = ifp->if_index;
944 ifm->ifm_flags = ifp->if_flags;
945 ifm->ifm_data = ifp->if_data;
946 ifm->ifm_addrs = info.rti_addrs;
947 error = copyout(ifm, w->w_where, len);
948 if (error)
949 return (error);
950 w->w_where += len;
951 break;
952 }
953
954 #ifdef COMPAT_14
955 case NET_RT_OIFLIST: {
956 struct if_msghdr14 *ifm;
957
958 ifm = (struct if_msghdr14 *)w->w_tmem;
959 ifm->ifm_index = ifp->if_index;
960 ifm->ifm_flags = ifp->if_flags;
961 ifm->ifm_data.ifi_type = ifp->if_data.ifi_type;
962 ifm->ifm_data.ifi_addrlen =
963 ifp->if_data.ifi_addrlen;
964 ifm->ifm_data.ifi_hdrlen =
965 ifp->if_data.ifi_hdrlen;
966 ifm->ifm_data.ifi_mtu = ifp->if_data.ifi_mtu;
967 ifm->ifm_data.ifi_metric =
968 ifp->if_data.ifi_metric;
969 ifm->ifm_data.ifi_baudrate =
970 ifp->if_data.ifi_baudrate;
971 ifm->ifm_data.ifi_ipackets =
972 ifp->if_data.ifi_ipackets;
973 ifm->ifm_data.ifi_ierrors =
974 ifp->if_data.ifi_ierrors;
975 ifm->ifm_data.ifi_opackets =
976 ifp->if_data.ifi_opackets;
977 ifm->ifm_data.ifi_oerrors =
978 ifp->if_data.ifi_oerrors;
979 ifm->ifm_data.ifi_collisions =
980 ifp->if_data.ifi_collisions;
981 ifm->ifm_data.ifi_ibytes =
982 ifp->if_data.ifi_ibytes;
983 ifm->ifm_data.ifi_obytes =
984 ifp->if_data.ifi_obytes;
985 ifm->ifm_data.ifi_imcasts =
986 ifp->if_data.ifi_imcasts;
987 ifm->ifm_data.ifi_omcasts =
988 ifp->if_data.ifi_omcasts;
989 ifm->ifm_data.ifi_iqdrops =
990 ifp->if_data.ifi_iqdrops;
991 ifm->ifm_data.ifi_noproto =
992 ifp->if_data.ifi_noproto;
993 ifm->ifm_data.ifi_lastchange =
994 ifp->if_data.ifi_lastchange;
995 ifm->ifm_addrs = info.rti_addrs;
996 error = copyout(ifm, w->w_where, len);
997 if (error)
998 return (error);
999 w->w_where += len;
1000 break;
1001 }
1002 #endif
1003 default:
1004 panic("sysctl_iflist(2)");
1005 }
1006 }
1007 while ((ifa = TAILQ_NEXT(ifa, ifa_list)) != NULL) {
1008 if (af && af != ifa->ifa_addr->sa_family)
1009 continue;
1010 ifaaddr = ifa->ifa_addr;
1011 netmask = ifa->ifa_netmask;
1012 brdaddr = ifa->ifa_dstaddr;
1013 if ((error = rt_msg2(RTM_NEWADDR, &info, 0, w, &len)))
1014 return (error);
1015 if (w->w_where && w->w_tmem && w->w_needed <= 0) {
1016 struct ifa_msghdr *ifam;
1017
1018 ifam = (struct ifa_msghdr *)w->w_tmem;
1019 ifam->ifam_index = ifa->ifa_ifp->if_index;
1020 ifam->ifam_flags = ifa->ifa_flags;
1021 ifam->ifam_metric = ifa->ifa_metric;
1022 ifam->ifam_addrs = info.rti_addrs;
1023 error = copyout(w->w_tmem, w->w_where, len);
1024 if (error)
1025 return (error);
1026 w->w_where += len;
1027 }
1028 }
1029 ifaaddr = netmask = brdaddr = 0;
1030 }
1031 return (0);
1032 }
1033
1034 static int
1035 sysctl_rtable(name, namelen, where, given, new, newlen)
1036 int *name;
1037 u_int namelen;
1038 void *where;
1039 size_t *given;
1040 void *new;
1041 size_t newlen;
1042 {
1043 struct radix_node_head *rnh;
1044 int i, s, error = EINVAL;
1045 u_char af;
1046 struct walkarg w;
1047
1048 if (new)
1049 return (EPERM);
1050 if (namelen != 3)
1051 return (EINVAL);
1052 af = name[0];
1053 w.w_tmemneeded = 0;
1054 w.w_tmemsize = 0;
1055 w.w_tmem = NULL;
1056 again:
1057 /* we may return here if a later [re]alloc of the t_mem buffer fails */
1058 if (w.w_tmemneeded) {
1059 w.w_tmem = (caddr_t) malloc(w.w_tmemneeded, M_RTABLE, M_WAITOK);
1060 w.w_tmemsize = w.w_tmemneeded;
1061 w.w_tmemneeded = 0;
1062 }
1063 w.w_op = name[1];
1064 w.w_arg = name[2];
1065 w.w_given = *given;
1066 w.w_needed = 0 - w.w_given;
1067 w.w_where = where;
1068
1069 s = splsoftnet();
1070 switch (w.w_op) {
1071
1072 case NET_RT_DUMP:
1073 case NET_RT_FLAGS:
1074 for (i = 1; i <= AF_MAX; i++)
1075 if ((rnh = rt_tables[i]) && (af == 0 || af == i) &&
1076 (error = (*rnh->rnh_walktree)(rnh,
1077 sysctl_dumpentry, &w)))
1078 break;
1079 break;
1080
1081 #ifdef COMPAT_14
1082 case NET_RT_OIFLIST:
1083 error = sysctl_iflist(af, &w, w.w_op);
1084 break;
1085 #endif
1086
1087 case NET_RT_IFLIST:
1088 error = sysctl_iflist(af, &w, w.w_op);
1089 }
1090 splx(s);
1091
1092 /* check to see if we couldn't allocate memory with NOWAIT */
1093 if (error == ENOBUFS && w.w_tmem == 0 && w.w_tmemneeded)
1094 goto again;
1095
1096 if (w.w_tmem)
1097 free(w.w_tmem, M_RTABLE);
1098 w.w_needed += w.w_given;
1099 if (where) {
1100 *given = w.w_where - (caddr_t) where;
1101 if (*given < w.w_needed)
1102 return (ENOMEM);
1103 } else {
1104 *given = (11 * w.w_needed) / 10;
1105 }
1106 return (error);
1107 }
1108
1109 /*
1110 * Definitions of protocols supported in the ROUTE domain.
1111 */
1112
1113 extern struct domain routedomain; /* or at least forward */
1114
1115 struct protosw routesw[] = {
1116 { SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR,
1117 raw_input, route_output, raw_ctlinput, 0,
1118 route_usrreq,
1119 raw_init, 0, 0, 0,
1120 sysctl_rtable,
1121 }
1122 };
1123
1124 struct domain routedomain =
1125 { PF_ROUTE, "route", route_init, 0, 0,
1126 routesw, &routesw[sizeof(routesw)/sizeof(routesw[0])] };
1127