route: Add rtsearch_global(), which is used to fix RTM_CHANGE support.
[dragonfly.git] / sys / net / rtsock.c
CommitLineData
f3ed2586
JH
1/*
2 * Copyright (c) 2004, 2005 The DragonFly Project. All rights reserved.
3 *
4 * This code is derived from software contributed to The DragonFly Project
5 * by Jeffrey M. Hsu.
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 DragonFly Project nor the names of its
16 * contributors may be used to endorse or promote products derived
17 * from this software without specific, prior written permission.
18 *
19 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
20 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
21 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
22 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
23 * COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
24 * INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING,
25 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
26 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
27 * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
28 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
29 * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
30 * SUCH DAMAGE.
31 */
32
984263bc
MD
33/*
34 * Copyright (c) 1988, 1991, 1993
35 * The Regents of the University of California. All rights reserved.
36 *
37 * Redistribution and use in source and binary forms, with or without
38 * modification, are permitted provided that the following conditions
39 * are met:
40 * 1. Redistributions of source code must retain the above copyright
41 * notice, this list of conditions and the following disclaimer.
42 * 2. Redistributions in binary form must reproduce the above copyright
43 * notice, this list of conditions and the following disclaimer in the
44 * documentation and/or other materials provided with the distribution.
45 * 3. All advertising materials mentioning features or use of this software
46 * must display the following acknowledgement:
47 * This product includes software developed by the University of
48 * California, Berkeley and its contributors.
49 * 4. Neither the name of the University nor the names of its contributors
50 * may be used to endorse or promote products derived from this software
51 * without specific prior written permission.
52 *
53 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
54 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
55 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
56 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
57 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
58 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
59 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
60 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
61 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
62 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
63 * SUCH DAMAGE.
64 *
65 * @(#)rtsock.c 8.7 (Berkeley) 10/12/95
66 * $FreeBSD: src/sys/net/rtsock.c,v 1.44.2.11 2002/12/04 14:05:41 ru Exp $
e3873585 67 * $DragonFly: src/sys/net/rtsock.c,v 1.45 2008/10/27 02:56:30 sephe Exp $
984263bc
MD
68 */
69
78812139
EN
70#include "opt_sctp.h"
71
984263bc
MD
72#include <sys/param.h>
73#include <sys/systm.h>
74#include <sys/kernel.h>
75#include <sys/sysctl.h>
76#include <sys/proc.h>
895c1f85 77#include <sys/priv.h>
984263bc
MD
78#include <sys/malloc.h>
79#include <sys/mbuf.h>
47e78ae8 80#include <sys/protosw.h>
984263bc
MD
81#include <sys/socket.h>
82#include <sys/socketvar.h>
83#include <sys/domain.h>
a49aa710 84#include <sys/thread2.h>
deffea2e 85
984263bc
MD
86#include <net/if.h>
87#include <net/route.h>
88#include <net/raw_cb.h>
47db7f9b 89#include <net/netmsg2.h>
984263bc 90
78812139
EN
91#ifdef SCTP
92extern void sctp_add_ip_address(struct ifaddr *ifa);
93extern void sctp_delete_ip_address(struct ifaddr *ifa);
94#endif /* SCTP */
95
984263bc
MD
96MALLOC_DEFINE(M_RTABLE, "routetbl", "routing tables");
97
f23061d4
JH
98static struct route_cb {
99 int ip_count;
100 int ip6_count;
101 int ipx_count;
102 int ns_count;
103 int any_count;
104} route_cb;
105
590b8cd4 106static const struct sockaddr route_src = { 2, PF_ROUTE, };
984263bc
MD
107
108struct walkarg {
109 int w_tmemsize;
110 int w_op, w_arg;
8e63efec 111 void *w_tmem;
984263bc
MD
112 struct sysctl_req *w_req;
113};
114
115static struct mbuf *
590b8cd4
JH
116 rt_msg_mbuf (int, struct rt_addrinfo *);
117static void rt_msg_buffer (int, struct rt_addrinfo *, void *buf, int len);
118static int rt_msgsize (int type, struct rt_addrinfo *rtinfo);
ef87f48d 119static int rt_xaddrs (char *, char *, struct rt_addrinfo *);
158abb01
RG
120static int sysctl_dumpentry (struct radix_node *rn, void *vw);
121static int sysctl_iflist (int af, struct walkarg *w);
deffea2e 122static int route_output(struct mbuf *, struct socket *, ...);
47e78ae8 123static void rt_setmetrics (u_long, struct rt_metrics *,
f23061d4 124 struct rt_metrics *);
984263bc 125
984263bc
MD
126/*
127 * It really doesn't make any sense at all for this code to share much
128 * with raw_usrreq.c, since its functionality is so restricted. XXX
129 */
130static int
131rts_abort(struct socket *so)
132{
a49aa710 133 int error;
f23061d4 134
a49aa710 135 crit_enter();
984263bc 136 error = raw_usrreqs.pru_abort(so);
a49aa710 137 crit_exit();
984263bc
MD
138 return error;
139}
140
141/* pru_accept is EOPNOTSUPP */
142
143static int
e4700d00 144rts_attach(struct socket *so, int proto, struct pru_attach_info *ai)
984263bc
MD
145{
146 struct rawcb *rp;
a49aa710 147 int error;
984263bc 148
ef87f48d 149 if (sotorawcb(so) != NULL)
984263bc 150 return EISCONN; /* XXX panic? */
f23061d4 151
efda3bd0 152 rp = kmalloc(sizeof *rp, M_PCB, M_WAITOK | M_ZERO);
984263bc
MD
153
154 /*
a49aa710 155 * The critical section is necessary to block protocols from sending
984263bc
MD
156 * error notifications (like RTM_REDIRECT or RTM_LOSING) while
157 * this PCB is extant but incompletely initialized.
158 * Probably we should try to do more of this work beforehand and
a49aa710 159 * eliminate the critical section.
984263bc 160 */
a49aa710 161 crit_enter();
ef87f48d 162 so->so_pcb = rp;
e4700d00 163 error = raw_attach(so, proto, ai->sb_rlimit);
984263bc
MD
164 rp = sotorawcb(so);
165 if (error) {
a49aa710 166 crit_exit();
efda3bd0 167 kfree(rp, M_PCB);
984263bc
MD
168 return error;
169 }
170 switch(rp->rcb_proto.sp_protocol) {
171 case AF_INET:
172 route_cb.ip_count++;
173 break;
174 case AF_INET6:
175 route_cb.ip6_count++;
176 break;
177 case AF_IPX:
178 route_cb.ipx_count++;
179 break;
180 case AF_NS:
181 route_cb.ns_count++;
182 break;
183 }
184 rp->rcb_faddr = &route_src;
185 route_cb.any_count++;
186 soisconnected(so);
187 so->so_options |= SO_USELOOPBACK;
a49aa710 188 crit_exit();
984263bc
MD
189 return 0;
190}
191
192static int
dadab5e9 193rts_bind(struct socket *so, struct sockaddr *nam, struct thread *td)
984263bc 194{
a49aa710 195 int error;
f23061d4 196
a49aa710 197 crit_enter();
dadab5e9 198 error = raw_usrreqs.pru_bind(so, nam, td); /* xxx just EINVAL */
a49aa710 199 crit_exit();
984263bc
MD
200 return error;
201}
202
203static int
dadab5e9 204rts_connect(struct socket *so, struct sockaddr *nam, struct thread *td)
984263bc 205{
a49aa710 206 int error;
f23061d4 207
a49aa710 208 crit_enter();
dadab5e9 209 error = raw_usrreqs.pru_connect(so, nam, td); /* XXX just EINVAL */
a49aa710 210 crit_exit();
984263bc
MD
211 return error;
212}
213
214/* pru_connect2 is EOPNOTSUPP */
215/* pru_control is EOPNOTSUPP */
216
217static int
218rts_detach(struct socket *so)
219{
220 struct rawcb *rp = sotorawcb(so);
a49aa710 221 int error;
984263bc 222
a49aa710 223 crit_enter();
ef87f48d 224 if (rp != NULL) {
984263bc
MD
225 switch(rp->rcb_proto.sp_protocol) {
226 case AF_INET:
227 route_cb.ip_count--;
228 break;
229 case AF_INET6:
230 route_cb.ip6_count--;
231 break;
232 case AF_IPX:
233 route_cb.ipx_count--;
234 break;
235 case AF_NS:
236 route_cb.ns_count--;
237 break;
238 }
239 route_cb.any_count--;
240 }
241 error = raw_usrreqs.pru_detach(so);
a49aa710 242 crit_exit();
984263bc
MD
243 return error;
244}
245
246static int
247rts_disconnect(struct socket *so)
248{
a49aa710 249 int error;
f23061d4 250
a49aa710 251 crit_enter();
984263bc 252 error = raw_usrreqs.pru_disconnect(so);
a49aa710 253 crit_exit();
984263bc
MD
254 return error;
255}
256
257/* pru_listen is EOPNOTSUPP */
258
259static int
260rts_peeraddr(struct socket *so, struct sockaddr **nam)
261{
a49aa710 262 int error;
f23061d4 263
a49aa710 264 crit_enter();
984263bc 265 error = raw_usrreqs.pru_peeraddr(so, nam);
a49aa710 266 crit_exit();
984263bc
MD
267 return error;
268}
269
270/* pru_rcvd is EOPNOTSUPP */
271/* pru_rcvoob is EOPNOTSUPP */
272
273static int
274rts_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam,
dadab5e9 275 struct mbuf *control, struct thread *td)
984263bc 276{
a49aa710 277 int error;
f23061d4 278
a49aa710 279 crit_enter();
dadab5e9 280 error = raw_usrreqs.pru_send(so, flags, m, nam, control, td);
a49aa710 281 crit_exit();
984263bc
MD
282 return error;
283}
284
285/* pru_sense is null */
286
287static int
288rts_shutdown(struct socket *so)
289{
a49aa710 290 int error;
f23061d4 291
a49aa710 292 crit_enter();
984263bc 293 error = raw_usrreqs.pru_shutdown(so);
a49aa710 294 crit_exit();
984263bc
MD
295 return error;
296}
297
298static int
299rts_sockaddr(struct socket *so, struct sockaddr **nam)
300{
a49aa710 301 int error;
f23061d4 302
a49aa710 303 crit_enter();
984263bc 304 error = raw_usrreqs.pru_sockaddr(so, nam);
a49aa710 305 crit_exit();
984263bc
MD
306 return error;
307}
308
309static struct pr_usrreqs route_usrreqs = {
fa5e758c
MD
310 .pru_abort = rts_abort,
311 .pru_accept = pru_accept_notsupp,
312 .pru_attach = rts_attach,
313 .pru_bind = rts_bind,
314 .pru_connect = rts_connect,
315 .pru_connect2 = pru_connect2_notsupp,
316 .pru_control = pru_control_notsupp,
317 .pru_detach = rts_detach,
318 .pru_disconnect = rts_disconnect,
319 .pru_listen = pru_listen_notsupp,
320 .pru_peeraddr = rts_peeraddr,
321 .pru_rcvd = pru_rcvd_notsupp,
322 .pru_rcvoob = pru_rcvoob_notsupp,
323 .pru_send = rts_send,
324 .pru_sense = pru_sense_null,
325 .pru_shutdown = rts_shutdown,
326 .pru_sockaddr = rts_sockaddr,
327 .pru_sosend = sosend,
328 .pru_soreceive = soreceive,
329 .pru_sopoll = sopoll
984263bc
MD
330};
331
590b8cd4
JH
332static __inline sa_family_t
333familyof(struct sockaddr *sa)
334{
335 return (sa != NULL ? sa->sa_family : 0);
336}
337
47db7f9b
MD
338/*
339 * Routing socket input function. The packet must be serialized onto cpu 0.
340 * We use the cpu0_soport() netisr processing loop to handle it.
341 *
342 * This looks messy but it means that anyone, including interrupt code,
343 * can send a message to the routing socket.
344 */
590b8cd4 345static void
47db7f9b 346rts_input_handler(struct netmsg *msg)
590b8cd4
JH
347{
348 static const struct sockaddr route_dst = { 2, PF_ROUTE, };
47db7f9b
MD
349 struct sockproto route_proto;
350 struct netmsg_packet *pmsg;
351 struct mbuf *m;
352 sa_family_t family;
d16fa201 353 struct rawcb *skip;
47db7f9b
MD
354
355 pmsg = (void *)msg;
47db7f9b
MD
356 family = pmsg->nm_netmsg.nm_lmsg.u.ms_result;
357 route_proto.sp_family = PF_ROUTE;
358 route_proto.sp_protocol = family;
590b8cd4 359
d16fa201
SZ
360 m = pmsg->nm_packet;
361 M_ASSERTPKTHDR(m);
362
363 skip = m->m_pkthdr.header;
364 m->m_pkthdr.header = NULL;
365
366 raw_input(m, &route_proto, &route_src, &route_dst, skip);
590b8cd4
JH
367}
368
47db7f9b 369static void
d16fa201 370rts_input_skip(struct mbuf *m, sa_family_t family, struct rawcb *skip)
47db7f9b
MD
371{
372 struct netmsg_packet *pmsg;
373 lwkt_port_t port;
374
d16fa201
SZ
375 M_ASSERTPKTHDR(m);
376
934c6849 377 port = cpu0_soport(NULL, NULL, NULL, 0);
47db7f9b
MD
378 pmsg = &m->m_hdr.mh_netmsg;
379 netmsg_init(&pmsg->nm_netmsg, &netisr_apanic_rport,
380 0, rts_input_handler);
381 pmsg->nm_packet = m;
382 pmsg->nm_netmsg.nm_lmsg.u.ms_result = family;
d16fa201 383 m->m_pkthdr.header = skip; /* XXX steal field in pkthdr */
47db7f9b
MD
384 lwkt_sendmsg(port, &pmsg->nm_netmsg.nm_lmsg);
385}
386
d16fa201
SZ
387static __inline void
388rts_input(struct mbuf *m, sa_family_t family)
389{
390 rts_input_skip(m, family, NULL);
391}
392
590b8cd4
JH
393static void *
394reallocbuf(void *ptr, size_t len, size_t olen)
395{
396 void *newptr;
397
efda3bd0 398 newptr = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
590b8cd4
JH
399 if (newptr == NULL)
400 return NULL;
401 bcopy(ptr, newptr, olen);
efda3bd0 402 kfree(ptr, M_RTABLE);
590b8cd4
JH
403 return (newptr);
404}
405
8e63efec
JH
406/*
407 * Internal helper routine for route_output().
408 */
590b8cd4
JH
409static int
410fillrtmsg(struct rt_msghdr **prtm, struct rtentry *rt,
411 struct rt_addrinfo *rtinfo)
412{
413 int msglen;
414 struct rt_msghdr *rtm = *prtm;
415
416 /* Fill in rt_addrinfo for call to rt_msg_buffer(). */
417 rtinfo->rti_dst = rt_key(rt);
418 rtinfo->rti_gateway = rt->rt_gateway;
419 rtinfo->rti_netmask = rt_mask(rt); /* might be NULL */
420 rtinfo->rti_genmask = rt->rt_genmask; /* might be NULL */
421 if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) {
422 if (rt->rt_ifp != NULL) {
423 rtinfo->rti_ifpaddr =
b2632176
SZ
424 TAILQ_FIRST(&rt->rt_ifp->if_addrheads[mycpuid])
425 ->ifa->ifa_addr;
590b8cd4
JH
426 rtinfo->rti_ifaaddr = rt->rt_ifa->ifa_addr;
427 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT)
428 rtinfo->rti_bcastaddr = rt->rt_ifa->ifa_dstaddr;
429 rtm->rtm_index = rt->rt_ifp->if_index;
430 } else {
431 rtinfo->rti_ifpaddr = NULL;
432 rtinfo->rti_ifaaddr = NULL;
433 }
434 }
435
436 msglen = rt_msgsize(rtm->rtm_type, rtinfo);
437 if (rtm->rtm_msglen < msglen) {
438 rtm = reallocbuf(rtm, msglen, rtm->rtm_msglen);
439 if (rtm == NULL)
440 return (ENOBUFS);
441 *prtm = rtm;
442 }
443 rt_msg_buffer(rtm->rtm_type, rtinfo, rtm, msglen);
444
445 rtm->rtm_flags = rt->rt_flags;
446 rtm->rtm_rmx = rt->rt_rmx;
447 rtm->rtm_addrs = rtinfo->rti_addrs;
448
449 return (0);
450}
451
ecdefdda
MD
452static void route_output_add_callback(int, int, struct rt_addrinfo *,
453 struct rtentry *, void *);
454static void route_output_delete_callback(int, int, struct rt_addrinfo *,
455 struct rtentry *, void *);
2c62ae3b 456static int route_output_change_callback(int, struct rt_addrinfo *,
ecdefdda
MD
457 struct rtentry *, void *);
458static void route_output_lock_callback(int, int, struct rt_addrinfo *,
459 struct rtentry *, void *);
460
984263bc
MD
461/*ARGSUSED*/
462static int
deffea2e 463route_output(struct mbuf *m, struct socket *so, ...)
984263bc 464{
2e9572df 465 struct rt_msghdr *rtm = NULL;
ecdefdda 466 struct rtentry *rt;
984263bc 467 struct radix_node_head *rnh;
f23061d4 468 struct rawcb *rp = NULL;
deffea2e 469 struct pr_output_info *oi;
590b8cd4 470 struct rt_addrinfo rtinfo;
ef87f48d 471 int len, error = 0;
deffea2e
JS
472 __va_list ap;
473
474 __va_start(ap, so);
475 oi = __va_arg(ap, struct pr_output_info *);
476 __va_end(ap);
984263bc 477
ef87f48d 478#define gotoerr(e) { error = e; goto flush;}
590b8cd4
JH
479
480 if (m == NULL ||
481 (m->m_len < sizeof(long) &&
482 (m = m_pullup(m, sizeof(long))) == NULL))
984263bc 483 return (ENOBUFS);
ef87f48d 484 if (!(m->m_flags & M_PKTHDR))
984263bc
MD
485 panic("route_output");
486 len = m->m_pkthdr.len;
590b8cd4 487 if (len < sizeof(struct rt_msghdr) ||
984263bc 488 len != mtod(m, struct rt_msghdr *)->rtm_msglen) {
590b8cd4 489 rtinfo.rti_dst = NULL;
ef87f48d 490 gotoerr(EINVAL);
984263bc 491 }
efda3bd0 492 rtm = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
ef87f48d 493 if (rtm == NULL) {
590b8cd4 494 rtinfo.rti_dst = NULL;
ef87f48d 495 gotoerr(ENOBUFS);
984263bc
MD
496 }
497 m_copydata(m, 0, len, (caddr_t)rtm);
498 if (rtm->rtm_version != RTM_VERSION) {
590b8cd4 499 rtinfo.rti_dst = NULL;
ef87f48d 500 gotoerr(EPROTONOSUPPORT);
984263bc 501 }
47e78ae8 502 rtm->rtm_pid = oi->p_pid;
590b8cd4
JH
503 bzero(&rtinfo, sizeof(struct rt_addrinfo));
504 rtinfo.rti_addrs = rtm->rtm_addrs;
505 if (rt_xaddrs((char *)(rtm + 1), (char *)rtm + len, &rtinfo) != 0) {
506 rtinfo.rti_dst = NULL;
ef87f48d 507 gotoerr(EINVAL);
984263bc 508 }
590b8cd4
JH
509 rtinfo.rti_flags = rtm->rtm_flags;
510 if (rtinfo.rti_dst == NULL || rtinfo.rti_dst->sa_family >= AF_MAX ||
511 (rtinfo.rti_gateway && rtinfo.rti_gateway->sa_family >= AF_MAX))
ef87f48d
JH
512 gotoerr(EINVAL);
513
590b8cd4
JH
514 if (rtinfo.rti_genmask != NULL) {
515 struct radix_node *n;
ef87f48d 516
590b8cd4
JH
517#define clen(s) (*(u_char *)(s))
518 n = rn_addmask((char *)rtinfo.rti_genmask, TRUE, 1);
519 if (n != NULL &&
520 rtinfo.rti_genmask->sa_len >= clen(n->rn_key) &&
521 bcmp((char *)rtinfo.rti_genmask + 1,
522 (char *)n->rn_key + 1, clen(n->rn_key) - 1) == 0)
523 rtinfo.rti_genmask = (struct sockaddr *)n->rn_key;
984263bc 524 else
ef87f48d 525 gotoerr(ENOBUFS);
984263bc
MD
526 }
527
528 /*
529 * Verify that the caller has the appropriate privilege; RTM_GET
530 * is the only operation the non-superuser is allowed.
531 */
895c1f85 532 if (rtm->rtm_type != RTM_GET && priv_check_cred(so->so_cred, PRIV_ROOT, 0) != 0)
ef87f48d 533 gotoerr(EPERM);
984263bc
MD
534
535 switch (rtm->rtm_type) {
984263bc 536 case RTM_ADD:
ecdefdda
MD
537 if (rtinfo.rti_gateway == NULL) {
538 error = EINVAL;
539 } else {
540 error = rtrequest1_global(RTM_ADD, &rtinfo,
541 route_output_add_callback, rtm);
984263bc
MD
542 }
543 break;
984263bc 544 case RTM_DELETE:
ecdefdda
MD
545 /*
546 * note: &rtm passed as argument so 'rtm' can be replaced.
547 */
548 error = rtrequest1_global(RTM_DELETE, &rtinfo,
549 route_output_delete_callback, &rtm);
984263bc 550 break;
984263bc 551 case RTM_GET:
ecdefdda
MD
552 rnh = rt_tables[mycpuid][rtinfo.rti_dst->sa_family];
553 if (rnh == NULL) {
554 error = EAFNOSUPPORT;
555 break;
556 }
590b8cd4
JH
557 rt = (struct rtentry *)
558 rnh->rnh_lookup((char *)rtinfo.rti_dst,
559 (char *)rtinfo.rti_netmask, rnh);
ecdefdda
MD
560 if (rt == NULL) {
561 error = ESRCH;
984263bc
MD
562 break;
563 }
ecdefdda
MD
564 rt->rt_refcnt++;
565 if (fillrtmsg(&rtm, rt, &rtinfo) != 0)
566 gotoerr(ENOBUFS);
567 --rt->rt_refcnt;
568 break;
569 case RTM_CHANGE:
2c62ae3b
SZ
570 error = rtsearch_global(RTM_CHANGE, &rtinfo,
571 route_output_change_callback, rtm,
572 RTS_EXACTMATCH);
ecdefdda
MD
573 break;
574 case RTM_LOCK:
575 error = rtrequest1_global(RTM_GET, &rtinfo,
576 route_output_lock_callback, rtm);
590b8cd4 577 break;
984263bc 578 default:
ecdefdda
MD
579 error = EOPNOTSUPP;
580 break;
984263bc
MD
581 }
582
583flush:
590b8cd4
JH
584 if (rtm != NULL) {
585 if (error != 0)
984263bc
MD
586 rtm->rtm_errno = error;
587 else
588 rtm->rtm_flags |= RTF_DONE;
589 }
ecdefdda 590
984263bc
MD
591 /*
592 * Check to see if we don't want our own messages.
593 */
ef87f48d 594 if (!(so->so_options & SO_USELOOPBACK)) {
984263bc 595 if (route_cb.any_count <= 1) {
590b8cd4 596 if (rtm != NULL)
efda3bd0 597 kfree(rtm, M_RTABLE);
984263bc
MD
598 m_freem(m);
599 return (error);
600 }
601 /* There is another listener, so construct message */
602 rp = sotorawcb(so);
603 }
590b8cd4 604 if (rtm != NULL) {
984263bc
MD
605 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm);
606 if (m->m_pkthdr.len < rtm->rtm_msglen) {
607 m_freem(m);
608 m = NULL;
609 } else if (m->m_pkthdr.len > rtm->rtm_msglen)
610 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len);
efda3bd0 611 kfree(rtm, M_RTABLE);
984263bc 612 }
ef87f48d 613 if (m != NULL)
d16fa201 614 rts_input_skip(m, familyof(rtinfo.rti_dst), rp);
984263bc
MD
615 return (error);
616}
617
ecdefdda
MD
618static void
619route_output_add_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
620 struct rtentry *rt, void *arg)
621{
622 struct rt_msghdr *rtm = arg;
623
624 if (error == 0 && rt != NULL) {
625 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx,
626 &rt->rt_rmx);
627 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits);
628 rt->rt_rmx.rmx_locks |=
629 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks);
630 rt->rt_genmask = rtinfo->rti_genmask;
631 }
632}
633
634static void
635route_output_delete_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
636 struct rtentry *rt, void *arg)
637{
638 struct rt_msghdr **rtm = arg;
639
640 if (error == 0 && rt) {
641 ++rt->rt_refcnt;
642 if (fillrtmsg(rtm, rt, rtinfo) != 0) {
643 error = ENOBUFS;
644 /* XXX no way to return the error */
645 }
646 --rt->rt_refcnt;
647 }
648}
649
2c62ae3b
SZ
650static int
651route_output_change_callback(int cmd, struct rt_addrinfo *rtinfo,
652 struct rtentry *rt, void *arg)
ecdefdda
MD
653{
654 struct rt_msghdr *rtm = arg;
655 struct ifaddr *ifa;
2c62ae3b 656 int error = 0;
ecdefdda
MD
657
658 /*
659 * new gateway could require new ifaddr, ifp;
660 * flags may also be different; ifp may be specified
661 * by ll sockaddr when protocol address is ambiguous
662 */
663 if (((rt->rt_flags & RTF_GATEWAY) && rtinfo->rti_gateway != NULL) ||
fb41af81
SZ
664 rtinfo->rti_ifpaddr != NULL ||
665 (rtinfo->rti_ifaaddr != NULL &&
666 !sa_equal(rtinfo->rti_ifaaddr, rt->rt_ifa->ifa_addr))) {
ecdefdda
MD
667 error = rt_getifa(rtinfo);
668 if (error != 0)
669 goto done;
670 }
671 if (rtinfo->rti_gateway != NULL) {
672 error = rt_setgate(rt, rt_key(rt), rtinfo->rti_gateway);
673 if (error != 0)
674 goto done;
675 }
676 if ((ifa = rtinfo->rti_ifa) != NULL) {
677 struct ifaddr *oifa = rt->rt_ifa;
678
679 if (oifa != ifa) {
680 if (oifa && oifa->ifa_rtrequest)
681 oifa->ifa_rtrequest(RTM_DELETE, rt, rtinfo);
682 IFAFREE(rt->rt_ifa);
683 IFAREF(ifa);
684 rt->rt_ifa = ifa;
685 rt->rt_ifp = rtinfo->rti_ifp;
686 }
687 }
688 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, &rt->rt_rmx);
689 if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest)
2c62ae3b 690 rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, rtinfo);
ecdefdda
MD
691 if (rtinfo->rti_genmask != NULL)
692 rt->rt_genmask = rtinfo->rti_genmask;
693done:
2c62ae3b 694 return error;
ecdefdda
MD
695}
696
697static void
698route_output_lock_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
699 struct rtentry *rt, void *arg)
700{
701 struct rt_msghdr *rtm = arg;
702
703 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits);
704 rt->rt_rmx.rmx_locks |=
705 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks);
706}
707
984263bc 708static void
ef87f48d 709rt_setmetrics(u_long which, struct rt_metrics *in, struct rt_metrics *out)
984263bc 710{
2e9572df
JH
711#define setmetric(flag, elt) if (which & (flag)) out->elt = in->elt;
712 setmetric(RTV_RPIPE, rmx_recvpipe);
713 setmetric(RTV_SPIPE, rmx_sendpipe);
714 setmetric(RTV_SSTHRESH, rmx_ssthresh);
715 setmetric(RTV_RTT, rmx_rtt);
716 setmetric(RTV_RTTVAR, rmx_rttvar);
717 setmetric(RTV_HOPCOUNT, rmx_hopcount);
718 setmetric(RTV_MTU, rmx_mtu);
719 setmetric(RTV_EXPIRE, rmx_expire);
720#undef setmetric
984263bc
MD
721}
722
723#define ROUNDUP(a) \
724 ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long))
984263bc 725
984263bc
MD
726/*
727 * Extract the addresses of the passed sockaddrs.
728 * Do a little sanity checking so as to avoid bad memory references.
729 * This data is derived straight from userland.
730 */
731static int
ef87f48d 732rt_xaddrs(char *cp, char *cplim, struct rt_addrinfo *rtinfo)
984263bc 733{
82ed7fc2
RG
734 struct sockaddr *sa;
735 int i;
984263bc
MD
736
737 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) {
738 if ((rtinfo->rti_addrs & (1 << i)) == 0)
739 continue;
740 sa = (struct sockaddr *)cp;
741 /*
742 * It won't fit.
743 */
0c3c561c 744 if ((cp + sa->sa_len) > cplim) {
984263bc
MD
745 return (EINVAL);
746 }
747
748 /*
ef87f48d 749 * There are no more... Quit now.
984263bc 750 * If there are more bits, they are in error.
ef87f48d 751 * I've seen this. route(1) can evidently generate these.
984263bc 752 * This causes kernel to core dump.
ef87f48d 753 * For compatibility, if we see this, point to a safe address.
984263bc
MD
754 */
755 if (sa->sa_len == 0) {
590b8cd4
JH
756 static struct sockaddr sa_zero = {
757 sizeof sa_zero, AF_INET,
758 };
759
984263bc 760 rtinfo->rti_info[i] = &sa_zero;
9b42cabe 761 kprintf("rtsock: received more addr bits than sockaddrs.\n");
984263bc
MD
762 return (0); /* should be EINVAL but for compat */
763 }
764
ef87f48d 765 /* Accept the sockaddr. */
984263bc 766 rtinfo->rti_info[i] = sa;
590b8cd4 767 cp += ROUNDUP(sa->sa_len);
984263bc
MD
768 }
769 return (0);
770}
771
590b8cd4
JH
772static int
773rt_msghdrsize(int type)
984263bc 774{
984263bc 775 switch (type) {
984263bc
MD
776 case RTM_DELADDR:
777 case RTM_NEWADDR:
590b8cd4 778 return sizeof(struct ifa_msghdr);
984263bc
MD
779 case RTM_DELMADDR:
780 case RTM_NEWMADDR:
590b8cd4 781 return sizeof(struct ifma_msghdr);
984263bc 782 case RTM_IFINFO:
590b8cd4 783 return sizeof(struct if_msghdr);
984263bc 784 case RTM_IFANNOUNCE:
841ab66c 785 case RTM_IEEE80211:
590b8cd4 786 return sizeof(struct if_announcemsghdr);
984263bc 787 default:
590b8cd4 788 return sizeof(struct rt_msghdr);
984263bc 789 }
590b8cd4
JH
790}
791
792static int
793rt_msgsize(int type, struct rt_addrinfo *rtinfo)
794{
795 int len, i;
796
797 len = rt_msghdrsize(type);
798 for (i = 0; i < RTAX_MAX; i++) {
799 if (rtinfo->rti_info[i] != NULL)
800 len += ROUNDUP(rtinfo->rti_info[i]->sa_len);
984263bc 801 }
590b8cd4
JH
802 len = ALIGN(len);
803 return len;
804}
805
806/*
807 * Build a routing message in a buffer.
808 * Copy the addresses in the rtinfo->rti_info[] sockaddr array
809 * to the end of the buffer after the message header.
810 *
811 * Set the rtinfo->rti_addrs bitmask of addresses present in rtinfo->rti_info[].
812 * This side-effect can be avoided if we reorder the addrs bitmask field in all
813 * the route messages to line up so we can set it here instead of back in the
814 * calling routine.
815 */
816static void
817rt_msg_buffer(int type, struct rt_addrinfo *rtinfo, void *buf, int msglen)
818{
819 struct rt_msghdr *rtm;
820 char *cp;
821 int dlen, i;
822
823 rtm = (struct rt_msghdr *) buf;
824 rtm->rtm_version = RTM_VERSION;
825 rtm->rtm_type = type;
826 rtm->rtm_msglen = msglen;
827
828 cp = (char *)buf + rt_msghdrsize(type);
829 rtinfo->rti_addrs = 0;
984263bc 830 for (i = 0; i < RTAX_MAX; i++) {
590b8cd4
JH
831 struct sockaddr *sa;
832
984263bc
MD
833 if ((sa = rtinfo->rti_info[i]) == NULL)
834 continue;
835 rtinfo->rti_addrs |= (1 << i);
836 dlen = ROUNDUP(sa->sa_len);
590b8cd4
JH
837 bcopy(sa, cp, dlen);
838 cp += dlen;
984263bc 839 }
984263bc
MD
840}
841
590b8cd4
JH
842/*
843 * Build a routing message in a mbuf chain.
844 * Copy the addresses in the rtinfo->rti_info[] sockaddr array
845 * to the end of the mbuf after the message header.
846 *
847 * Set the rtinfo->rti_addrs bitmask of addresses present in rtinfo->rti_info[].
848 * This side-effect can be avoided if we reorder the addrs bitmask field in all
849 * the route messages to line up so we can set it here instead of back in the
850 * calling routine.
851 */
852static struct mbuf *
853rt_msg_mbuf(int type, struct rt_addrinfo *rtinfo)
984263bc 854{
590b8cd4
JH
855 struct mbuf *m;
856 struct rt_msghdr *rtm;
857 int hlen, len;
82ed7fc2 858 int i;
984263bc 859
590b8cd4
JH
860 hlen = rt_msghdrsize(type);
861 KASSERT(hlen <= MCLBYTES, ("rt_msg_mbuf: hlen %d doesn't fit", hlen));
984263bc 862
50503f0f 863 m = m_getl(hlen, MB_DONTWAIT, MT_DATA, M_PKTHDR, NULL);
590b8cd4
JH
864 if (m == NULL)
865 return (NULL);
e9fa4b60 866 mbuftrackid(m, 32);
590b8cd4
JH
867 m->m_pkthdr.len = m->m_len = hlen;
868 m->m_pkthdr.rcvif = NULL;
869 rtinfo->rti_addrs = 0;
870 len = hlen;
984263bc 871 for (i = 0; i < RTAX_MAX; i++) {
82ed7fc2 872 struct sockaddr *sa;
590b8cd4 873 int dlen;
984263bc 874
ef87f48d 875 if ((sa = rtinfo->rti_info[i]) == NULL)
984263bc
MD
876 continue;
877 rtinfo->rti_addrs |= (1 << i);
878 dlen = ROUNDUP(sa->sa_len);
590b8cd4 879 m_copyback(m, len, dlen, (caddr_t)sa); /* can grow mbuf chain */
984263bc
MD
880 len += dlen;
881 }
590b8cd4
JH
882 if (m->m_pkthdr.len != len) { /* one of the m_copyback() calls failed */
883 m_freem(m);
884 return (NULL);
984263bc 885 }
590b8cd4
JH
886 rtm = mtod(m, struct rt_msghdr *);
887 bzero(rtm, hlen);
888 rtm->rtm_msglen = len;
889 rtm->rtm_version = RTM_VERSION;
890 rtm->rtm_type = type;
891 return (m);
984263bc
MD
892}
893
894/*
895 * This routine is called to generate a message from the routing
f23061d4 896 * socket indicating that a redirect has occurred, a routing lookup
984263bc
MD
897 * has failed, or that a protocol has detected timeouts to a particular
898 * destination.
899 */
900void
f23061d4 901rt_missmsg(int type, struct rt_addrinfo *rtinfo, int flags, int error)
984263bc 902{
0c3c561c 903 struct sockaddr *dst = rtinfo->rti_info[RTAX_DST];
82ed7fc2
RG
904 struct rt_msghdr *rtm;
905 struct mbuf *m;
984263bc
MD
906
907 if (route_cb.any_count == 0)
908 return;
590b8cd4 909 m = rt_msg_mbuf(type, rtinfo);
ef87f48d 910 if (m == NULL)
984263bc
MD
911 return;
912 rtm = mtod(m, struct rt_msghdr *);
913 rtm->rtm_flags = RTF_DONE | flags;
914 rtm->rtm_errno = error;
915 rtm->rtm_addrs = rtinfo->rti_addrs;
590b8cd4 916 rts_input(m, familyof(dst));
984263bc
MD
917}
918
f3ed2586
JH
919void
920rt_dstmsg(int type, struct sockaddr *dst, int error)
921{
922 struct rt_msghdr *rtm;
923 struct rt_addrinfo addrs;
924 struct mbuf *m;
925
926 if (route_cb.any_count == 0)
927 return;
928 bzero(&addrs, sizeof(struct rt_addrinfo));
929 addrs.rti_info[RTAX_DST] = dst;
590b8cd4 930 m = rt_msg_mbuf(type, &addrs);
f3ed2586
JH
931 if (m == NULL)
932 return;
933 rtm = mtod(m, struct rt_msghdr *);
934 rtm->rtm_flags = RTF_DONE;
935 rtm->rtm_errno = error;
936 rtm->rtm_addrs = addrs.rti_addrs;
590b8cd4 937 rts_input(m, familyof(dst));
f3ed2586
JH
938}
939
984263bc
MD
940/*
941 * This routine is called to generate a message from the routing
942 * socket indicating that the status of a network interface has changed.
943 */
944void
0c3c561c 945rt_ifmsg(struct ifnet *ifp)
984263bc 946{
82ed7fc2 947 struct if_msghdr *ifm;
984263bc 948 struct mbuf *m;
590b8cd4 949 struct rt_addrinfo rtinfo;
984263bc
MD
950
951 if (route_cb.any_count == 0)
952 return;
590b8cd4
JH
953 bzero(&rtinfo, sizeof(struct rt_addrinfo));
954 m = rt_msg_mbuf(RTM_IFINFO, &rtinfo);
ef87f48d 955 if (m == NULL)
984263bc
MD
956 return;
957 ifm = mtod(m, struct if_msghdr *);
958 ifm->ifm_index = ifp->if_index;
9c095379 959 ifm->ifm_flags = ifp->if_flags;
984263bc 960 ifm->ifm_data = ifp->if_data;
590b8cd4
JH
961 ifm->ifm_addrs = 0;
962 rts_input(m, 0);
984263bc
MD
963}
964
372316d9
JH
965static void
966rt_ifamsg(int cmd, struct ifaddr *ifa)
967{
968 struct ifa_msghdr *ifam;
590b8cd4 969 struct rt_addrinfo rtinfo;
372316d9 970 struct mbuf *m;
372316d9
JH
971 struct ifnet *ifp = ifa->ifa_ifp;
972
590b8cd4
JH
973 bzero(&rtinfo, sizeof(struct rt_addrinfo));
974 rtinfo.rti_ifaaddr = ifa->ifa_addr;
b2632176
SZ
975 rtinfo.rti_ifpaddr =
976 TAILQ_FIRST(&ifp->if_addrheads[mycpuid])->ifa->ifa_addr;
590b8cd4
JH
977 rtinfo.rti_netmask = ifa->ifa_netmask;
978 rtinfo.rti_bcastaddr = ifa->ifa_dstaddr;
f23061d4 979
590b8cd4 980 m = rt_msg_mbuf(cmd, &rtinfo);
f23061d4 981 if (m == NULL)
372316d9 982 return;
f23061d4 983
372316d9
JH
984 ifam = mtod(m, struct ifa_msghdr *);
985 ifam->ifam_index = ifp->if_index;
986 ifam->ifam_metric = ifa->ifa_metric;
987 ifam->ifam_flags = ifa->ifa_flags;
590b8cd4 988 ifam->ifam_addrs = rtinfo.rti_addrs;
f23061d4 989
590b8cd4 990 rts_input(m, familyof(ifa->ifa_addr));
372316d9
JH
991}
992
f3ed2586
JH
993void
994rt_rtmsg(int cmd, struct rtentry *rt, struct ifnet *ifp, int error)
372316d9
JH
995{
996 struct rt_msghdr *rtm;
590b8cd4 997 struct rt_addrinfo rtinfo;
372316d9 998 struct mbuf *m;
f3ed2586 999 struct sockaddr *dst;
372316d9
JH
1000
1001 if (rt == NULL)
1002 return;
f23061d4 1003
590b8cd4
JH
1004 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1005 rtinfo.rti_dst = dst = rt_key(rt);
1006 rtinfo.rti_gateway = rt->rt_gateway;
1007 rtinfo.rti_netmask = rt_mask(rt);
b2632176
SZ
1008 if (ifp != NULL) {
1009 rtinfo.rti_ifpaddr =
1010 TAILQ_FIRST(&ifp->if_addrheads[mycpuid])->ifa->ifa_addr;
1011 }
590b8cd4 1012 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr;
f23061d4 1013
590b8cd4 1014 m = rt_msg_mbuf(cmd, &rtinfo);
f23061d4 1015 if (m == NULL)
372316d9 1016 return;
f23061d4 1017
372316d9 1018 rtm = mtod(m, struct rt_msghdr *);
f3ed2586
JH
1019 if (ifp != NULL)
1020 rtm->rtm_index = ifp->if_index;
372316d9
JH
1021 rtm->rtm_flags |= rt->rt_flags;
1022 rtm->rtm_errno = error;
590b8cd4 1023 rtm->rtm_addrs = rtinfo.rti_addrs;
f23061d4 1024
590b8cd4 1025 rts_input(m, familyof(dst));
372316d9
JH
1026}
1027
984263bc
MD
1028/*
1029 * This is called to generate messages from the routing socket
1030 * indicating a network interface has had addresses associated with it.
1031 * if we ever reverse the logic and replace messages TO the routing
1032 * socket indicate a request to configure interfaces, then it will
1033 * be unnecessary as the routing socket will automatically generate
1034 * copies of it.
1035 */
1036void
0c3c561c 1037rt_newaddrmsg(int cmd, struct ifaddr *ifa, int error, struct rtentry *rt)
984263bc 1038{
78812139
EN
1039#ifdef SCTP
1040 /*
1041 * notify the SCTP stack
1042 * this will only get called when an address is added/deleted
1043 * XXX pass the ifaddr struct instead if ifa->ifa_addr...
1044 */
1045 if (cmd == RTM_ADD)
1046 sctp_add_ip_address(ifa);
1047 else if (cmd == RTM_DELETE)
1048 sctp_delete_ip_address(ifa);
1049#endif /* SCTP */
1050
984263bc
MD
1051 if (route_cb.any_count == 0)
1052 return;
984263bc 1053
372316d9
JH
1054 if (cmd == RTM_ADD) {
1055 rt_ifamsg(RTM_NEWADDR, ifa);
f3ed2586 1056 rt_rtmsg(RTM_ADD, rt, ifa->ifa_ifp, error);
372316d9
JH
1057 } else {
1058 KASSERT((cmd == RTM_DELETE), ("unknown cmd %d", cmd));
f3ed2586 1059 rt_rtmsg(RTM_DELETE, rt, ifa->ifa_ifp, error);
372316d9 1060 rt_ifamsg(RTM_DELADDR, ifa);
984263bc
MD
1061 }
1062}
1063
1064/*
1065 * This is the analogue to the rt_newaddrmsg which performs the same
1066 * function but for multicast group memberhips. This is easier since
1067 * there is no route state to worry about.
1068 */
1069void
0c3c561c 1070rt_newmaddrmsg(int cmd, struct ifmultiaddr *ifma)
984263bc 1071{
590b8cd4 1072 struct rt_addrinfo rtinfo;
ef87f48d 1073 struct mbuf *m = NULL;
984263bc
MD
1074 struct ifnet *ifp = ifma->ifma_ifp;
1075 struct ifma_msghdr *ifmam;
1076
1077 if (route_cb.any_count == 0)
1078 return;
1079
590b8cd4
JH
1080 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1081 rtinfo.rti_ifaaddr = ifma->ifma_addr;
b2632176
SZ
1082 if (ifp != NULL && !TAILQ_EMPTY(&ifp->if_addrheads[mycpuid])) {
1083 rtinfo.rti_ifpaddr =
1084 TAILQ_FIRST(&ifp->if_addrheads[mycpuid])->ifa->ifa_addr;
1085 }
984263bc
MD
1086 /*
1087 * If a link-layer address is present, present it as a ``gateway''
1088 * (similarly to how ARP entries, e.g., are presented).
1089 */
590b8cd4 1090 rtinfo.rti_gateway = ifma->ifma_lladdr;
f23061d4 1091
590b8cd4 1092 m = rt_msg_mbuf(cmd, &rtinfo);
f23061d4 1093 if (m == NULL)
984263bc 1094 return;
f23061d4 1095
984263bc
MD
1096 ifmam = mtod(m, struct ifma_msghdr *);
1097 ifmam->ifmam_index = ifp->if_index;
590b8cd4 1098 ifmam->ifmam_addrs = rtinfo.rti_addrs;
f23061d4 1099
590b8cd4 1100 rts_input(m, familyof(ifma->ifma_addr));
984263bc
MD
1101}
1102
841ab66c
SZ
1103static struct mbuf *
1104rt_makeifannouncemsg(struct ifnet *ifp, int type, int what,
1105 struct rt_addrinfo *info)
984263bc 1106{
590b8cd4 1107 struct if_announcemsghdr *ifan;
841ab66c 1108 struct mbuf *m;
984263bc
MD
1109
1110 if (route_cb.any_count == 0)
841ab66c 1111 return NULL;
f23061d4 1112
841ab66c
SZ
1113 bzero(info, sizeof(*info));
1114 m = rt_msg_mbuf(type, info);
984263bc 1115 if (m == NULL)
841ab66c 1116 return NULL;
f23061d4 1117
984263bc
MD
1118 ifan = mtod(m, struct if_announcemsghdr *);
1119 ifan->ifan_index = ifp->if_index;
5fe66e68 1120 strlcpy(ifan->ifan_name, ifp->if_xname, sizeof ifan->ifan_name);
984263bc 1121 ifan->ifan_what = what;
841ab66c
SZ
1122 return m;
1123}
1124
1125/*
1126 * This is called to generate routing socket messages indicating
1127 * IEEE80211 wireless events.
1128 * XXX we piggyback on the RTM_IFANNOUNCE msg format in a clumsy way.
1129 */
1130void
1131rt_ieee80211msg(struct ifnet *ifp, int what, void *data, size_t data_len)
1132{
1133 struct rt_addrinfo info;
1134 struct mbuf *m;
f23061d4 1135
841ab66c
SZ
1136 m = rt_makeifannouncemsg(ifp, RTM_IEEE80211, what, &info);
1137 if (m == NULL)
1138 return;
1139
1140 /*
1141 * Append the ieee80211 data. Try to stick it in the
1142 * mbuf containing the ifannounce msg; otherwise allocate
1143 * a new mbuf and append.
1144 *
1145 * NB: we assume m is a single mbuf.
1146 */
1147 if (data_len > M_TRAILINGSPACE(m)) {
b11e1686 1148 struct mbuf *n = m_get(MB_DONTWAIT, MT_DATA);
841ab66c
SZ
1149 if (n == NULL) {
1150 m_freem(m);
1151 return;
1152 }
1153 bcopy(data, mtod(n, void *), data_len);
1154 n->m_len = data_len;
1155 m->m_next = n;
1156 } else if (data_len > 0) {
1157 bcopy(data, mtod(m, u_int8_t *) + m->m_len, data_len);
1158 m->m_len += data_len;
1159 }
e9fa4b60 1160 mbuftrackid(m, 33);
841ab66c
SZ
1161 if (m->m_flags & M_PKTHDR)
1162 m->m_pkthdr.len += data_len;
1163 mtod(m, struct if_announcemsghdr *)->ifan_msglen += data_len;
590b8cd4 1164 rts_input(m, 0);
841ab66c
SZ
1165}
1166
1167/*
1168 * This is called to generate routing socket messages indicating
1169 * network interface arrival and departure.
1170 */
1171void
1172rt_ifannouncemsg(struct ifnet *ifp, int what)
1173{
1174 struct rt_addrinfo addrinfo;
1175 struct mbuf *m;
1176
1177 m = rt_makeifannouncemsg(ifp, RTM_IFANNOUNCE, what, &addrinfo);
1178 if (m != NULL)
1179 rts_input(m, 0);
590b8cd4 1180}
f23061d4 1181
590b8cd4
JH
1182static int
1183resizewalkarg(struct walkarg *w, int len)
1184{
1185 void *newptr;
1186
efda3bd0 1187 newptr = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
590b8cd4
JH
1188 if (newptr == NULL)
1189 return (ENOMEM);
1190 if (w->w_tmem != NULL)
efda3bd0 1191 kfree(w->w_tmem, M_RTABLE);
590b8cd4
JH
1192 w->w_tmem = newptr;
1193 w->w_tmemsize = len;
1194 return (0);
1195}
984263bc
MD
1196
1197/*
1198 * This is used in dumping the kernel table via sysctl().
1199 */
1200int
590b8cd4 1201sysctl_dumpentry(struct radix_node *rn, void *vw)
984263bc 1202{
82ed7fc2
RG
1203 struct walkarg *w = vw;
1204 struct rtentry *rt = (struct rtentry *)rn;
590b8cd4
JH
1205 struct rt_addrinfo rtinfo;
1206 int error, msglen;
984263bc
MD
1207
1208 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg))
1209 return 0;
f23061d4 1210
590b8cd4
JH
1211 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1212 rtinfo.rti_dst = rt_key(rt);
1213 rtinfo.rti_gateway = rt->rt_gateway;
1214 rtinfo.rti_netmask = rt_mask(rt);
1215 rtinfo.rti_genmask = rt->rt_genmask;
ef87f48d 1216 if (rt->rt_ifp != NULL) {
590b8cd4 1217 rtinfo.rti_ifpaddr =
b2632176 1218 TAILQ_FIRST(&rt->rt_ifp->if_addrheads[mycpuid])->ifa->ifa_addr;
590b8cd4 1219 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr;
984263bc 1220 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT)
590b8cd4 1221 rtinfo.rti_bcastaddr = rt->rt_ifa->ifa_dstaddr;
984263bc 1222 }
590b8cd4
JH
1223 msglen = rt_msgsize(RTM_GET, &rtinfo);
1224 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0)
1225 return (ENOMEM);
1226 rt_msg_buffer(RTM_GET, &rtinfo, w->w_tmem, msglen);
1227 if (w->w_req != NULL) {
8e63efec 1228 struct rt_msghdr *rtm = w->w_tmem;
984263bc
MD
1229
1230 rtm->rtm_flags = rt->rt_flags;
1231 rtm->rtm_use = rt->rt_use;
1232 rtm->rtm_rmx = rt->rt_rmx;
1233 rtm->rtm_index = rt->rt_ifp->if_index;
1234 rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0;
590b8cd4
JH
1235 rtm->rtm_addrs = rtinfo.rti_addrs;
1236 error = SYSCTL_OUT(w->w_req, rtm, msglen);
984263bc
MD
1237 return (error);
1238 }
590b8cd4 1239 return (0);
984263bc
MD
1240}
1241
590b8cd4
JH
1242static int
1243sysctl_iflist(int af, struct walkarg *w)
984263bc 1244{
82ed7fc2 1245 struct ifnet *ifp;
590b8cd4
JH
1246 struct rt_addrinfo rtinfo;
1247 int msglen, error;
984263bc 1248
590b8cd4 1249 bzero(&rtinfo, sizeof(struct rt_addrinfo));
984263bc 1250 TAILQ_FOREACH(ifp, &ifnet, if_link) {
b2632176
SZ
1251 struct ifaddr_container *ifac;
1252 struct ifaddr *ifa;
1253
984263bc
MD
1254 if (w->w_arg && w->w_arg != ifp->if_index)
1255 continue;
b2632176
SZ
1256 ifac = TAILQ_FIRST(&ifp->if_addrheads[mycpuid]);
1257 ifa = ifac->ifa;
590b8cd4
JH
1258 rtinfo.rti_ifpaddr = ifa->ifa_addr;
1259 msglen = rt_msgsize(RTM_IFINFO, &rtinfo);
1260 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0)
1261 return (ENOMEM);
1262 rt_msg_buffer(RTM_IFINFO, &rtinfo, w->w_tmem, msglen);
1263 rtinfo.rti_ifpaddr = NULL;
ef87f48d 1264 if (w->w_req != NULL && w->w_tmem != NULL) {
8e63efec 1265 struct if_msghdr *ifm = w->w_tmem;
984263bc 1266
984263bc 1267 ifm->ifm_index = ifp->if_index;
9c095379 1268 ifm->ifm_flags = ifp->if_flags;
984263bc 1269 ifm->ifm_data = ifp->if_data;
590b8cd4
JH
1270 ifm->ifm_addrs = rtinfo.rti_addrs;
1271 error = SYSCTL_OUT(w->w_req, ifm, msglen);
984263bc
MD
1272 if (error)
1273 return (error);
1274 }
b2632176
SZ
1275 while ((ifac = TAILQ_NEXT(ifac, ifa_link)) != NULL) {
1276 ifa = ifac->ifa;
1277
984263bc
MD
1278 if (af && af != ifa->ifa_addr->sa_family)
1279 continue;
590b8cd4 1280 if (curproc->p_ucred->cr_prison &&
87de5057 1281 prison_if(curproc->p_ucred, ifa->ifa_addr))
984263bc 1282 continue;
590b8cd4
JH
1283 rtinfo.rti_ifaaddr = ifa->ifa_addr;
1284 rtinfo.rti_netmask = ifa->ifa_netmask;
1285 rtinfo.rti_bcastaddr = ifa->ifa_dstaddr;
1286 msglen = rt_msgsize(RTM_NEWADDR, &rtinfo);
1287 if (w->w_tmemsize < msglen &&
1288 resizewalkarg(w, msglen) != 0)
1289 return (ENOMEM);
1290 rt_msg_buffer(RTM_NEWADDR, &rtinfo, w->w_tmem, msglen);
1291 if (w->w_req != NULL) {
8e63efec 1292 struct ifa_msghdr *ifam = w->w_tmem;
984263bc 1293
984263bc
MD
1294 ifam->ifam_index = ifa->ifa_ifp->if_index;
1295 ifam->ifam_flags = ifa->ifa_flags;
1296 ifam->ifam_metric = ifa->ifa_metric;
590b8cd4
JH
1297 ifam->ifam_addrs = rtinfo.rti_addrs;
1298 error = SYSCTL_OUT(w->w_req, w->w_tmem, msglen);
984263bc
MD
1299 if (error)
1300 return (error);
1301 }
1302 }
590b8cd4
JH
1303 rtinfo.rti_netmask = NULL;
1304 rtinfo.rti_ifaaddr = NULL;
1305 rtinfo.rti_bcastaddr = NULL;
984263bc
MD
1306 }
1307 return (0);
1308}
1309
1310static int
1311sysctl_rtsock(SYSCTL_HANDLER_ARGS)
1312{
1313 int *name = (int *)arg1;
1314 u_int namelen = arg2;
82ed7fc2 1315 struct radix_node_head *rnh;
a49aa710 1316 int i, error = EINVAL;
ecdefdda 1317 int origcpu;
984263bc
MD
1318 u_char af;
1319 struct walkarg w;
1320
1321 name ++;
1322 namelen--;
1323 if (req->newptr)
1324 return (EPERM);
ecdefdda 1325 if (namelen != 3 && namelen != 4)
984263bc
MD
1326 return (EINVAL);
1327 af = name[0];
5fe66e68 1328 bzero(&w, sizeof w);
984263bc
MD
1329 w.w_op = name[1];
1330 w.w_arg = name[2];
1331 w.w_req = req;
1332
ecdefdda
MD
1333 /*
1334 * Optional third argument specifies cpu, used primarily for
1335 * debugging the route table.
1336 */
1337 if (namelen == 4) {
1338 if (name[3] < 0 || name[3] >= ncpus)
1339 return (EINVAL);
1340 origcpu = mycpuid;
1341 lwkt_migratecpu(name[3]);
1342 } else {
1343 origcpu = -1;
1344 }
a49aa710 1345 crit_enter();
984263bc 1346 switch (w.w_op) {
984263bc
MD
1347 case NET_RT_DUMP:
1348 case NET_RT_FLAGS:
1349 for (i = 1; i <= AF_MAX; i++)
ecdefdda
MD
1350 if ((rnh = rt_tables[mycpuid][i]) &&
1351 (af == 0 || af == i) &&
984263bc 1352 (error = rnh->rnh_walktree(rnh,
f23061d4 1353 sysctl_dumpentry, &w)))
984263bc
MD
1354 break;
1355 break;
1356
1357 case NET_RT_IFLIST:
1358 error = sysctl_iflist(af, &w);
1359 }
a49aa710 1360 crit_exit();
590b8cd4 1361 if (w.w_tmem != NULL)
efda3bd0 1362 kfree(w.w_tmem, M_RTABLE);
ecdefdda
MD
1363 if (origcpu >= 0)
1364 lwkt_migratecpu(origcpu);
984263bc
MD
1365 return (error);
1366}
1367
1368SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, "");
1369
1370/*
1371 * Definitions of protocols supported in the ROUTE domain.
1372 */
1373
ba4257b8 1374static struct domain routedomain; /* or at least forward */
984263bc
MD
1375
1376static struct protosw routesw[] = {
1377{ SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR,
1378 0, route_output, raw_ctlinput, 0,
e3873585 1379 cpu0_soport, cpu0_ctlport,
984263bc
MD
1380 raw_init, 0, 0, 0,
1381 &route_usrreqs
1382}
1383};
1384
ba4257b8 1385static struct domain routedomain = {
9c70fe43
JH
1386 PF_ROUTE, "route", NULL, NULL, NULL,
1387 routesw, &routesw[(sizeof routesw)/(sizeof routesw[0])],
1388};
984263bc
MD
1389
1390DOMAIN_SET(route);
47db7f9b 1391