route_output: Correctly set family according to the rtinfo.rti_dst
[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;
1730f0e1 433 }
9d1f1679
SZ
434 } else if (rt->rt_ifp != NULL) {
435 rtm->rtm_index = rt->rt_ifp->if_index;
590b8cd4
JH
436 }
437
438 msglen = rt_msgsize(rtm->rtm_type, rtinfo);
439 if (rtm->rtm_msglen < msglen) {
440 rtm = reallocbuf(rtm, msglen, rtm->rtm_msglen);
441 if (rtm == NULL)
442 return (ENOBUFS);
443 *prtm = rtm;
444 }
445 rt_msg_buffer(rtm->rtm_type, rtinfo, rtm, msglen);
446
447 rtm->rtm_flags = rt->rt_flags;
448 rtm->rtm_rmx = rt->rt_rmx;
449 rtm->rtm_addrs = rtinfo->rti_addrs;
450
451 return (0);
452}
453
ecdefdda
MD
454static void route_output_add_callback(int, int, struct rt_addrinfo *,
455 struct rtentry *, void *);
456static void route_output_delete_callback(int, int, struct rt_addrinfo *,
457 struct rtentry *, void *);
00104f14 458static int route_output_get_callback(int, struct rt_addrinfo *,
b5e01dd3 459 struct rtentry *, void *, int);
2c62ae3b 460static int route_output_change_callback(int, struct rt_addrinfo *,
b5e01dd3 461 struct rtentry *, void *, int);
ae2a5ccd 462static int route_output_lock_callback(int, struct rt_addrinfo *,
b5e01dd3 463 struct rtentry *, void *, int);
ecdefdda 464
984263bc
MD
465/*ARGSUSED*/
466static int
deffea2e 467route_output(struct mbuf *m, struct socket *so, ...)
984263bc 468{
2e9572df 469 struct rt_msghdr *rtm = NULL;
f23061d4 470 struct rawcb *rp = NULL;
deffea2e 471 struct pr_output_info *oi;
590b8cd4 472 struct rt_addrinfo rtinfo;
a8c218b2 473 sa_family_t family;
ef87f48d 474 int len, error = 0;
deffea2e
JS
475 __va_list ap;
476
1cfcc9b6
SZ
477 M_ASSERTPKTHDR(m);
478
deffea2e
JS
479 __va_start(ap, so);
480 oi = __va_arg(ap, struct pr_output_info *);
481 __va_end(ap);
984263bc 482
a8c218b2
SZ
483 family = familyof(NULL);
484
ef87f48d 485#define gotoerr(e) { error = e; goto flush;}
590b8cd4
JH
486
487 if (m == NULL ||
488 (m->m_len < sizeof(long) &&
489 (m = m_pullup(m, sizeof(long))) == NULL))
984263bc 490 return (ENOBUFS);
984263bc 491 len = m->m_pkthdr.len;
590b8cd4 492 if (len < sizeof(struct rt_msghdr) ||
a8c218b2 493 len != mtod(m, struct rt_msghdr *)->rtm_msglen)
ef87f48d 494 gotoerr(EINVAL);
a8c218b2 495
efda3bd0 496 rtm = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
a8c218b2 497 if (rtm == NULL)
ef87f48d 498 gotoerr(ENOBUFS);
a8c218b2 499
984263bc 500 m_copydata(m, 0, len, (caddr_t)rtm);
a8c218b2 501 if (rtm->rtm_version != RTM_VERSION)
ef87f48d 502 gotoerr(EPROTONOSUPPORT);
a8c218b2 503
47e78ae8 504 rtm->rtm_pid = oi->p_pid;
590b8cd4
JH
505 bzero(&rtinfo, sizeof(struct rt_addrinfo));
506 rtinfo.rti_addrs = rtm->rtm_addrs;
a8c218b2 507 if (rt_xaddrs((char *)(rtm + 1), (char *)rtm + len, &rtinfo) != 0)
ef87f48d 508 gotoerr(EINVAL);
a8c218b2 509
590b8cd4
JH
510 rtinfo.rti_flags = rtm->rtm_flags;
511 if (rtinfo.rti_dst == NULL || rtinfo.rti_dst->sa_family >= AF_MAX ||
512 (rtinfo.rti_gateway && rtinfo.rti_gateway->sa_family >= AF_MAX))
ef87f48d
JH
513 gotoerr(EINVAL);
514
a8c218b2
SZ
515 family = familyof(rtinfo.rti_dst);
516
590b8cd4 517 if (rtinfo.rti_genmask != NULL) {
ac0f13c5
SZ
518 error = rtmask_add_global(rtinfo.rti_genmask);
519 if (error)
520 goto flush;
984263bc
MD
521 }
522
523 /*
524 * Verify that the caller has the appropriate privilege; RTM_GET
525 * is the only operation the non-superuser is allowed.
526 */
02650706
SZ
527 if (rtm->rtm_type != RTM_GET &&
528 priv_check_cred(so->so_cred, PRIV_ROOT, 0) != 0)
ef87f48d 529 gotoerr(EPERM);
984263bc
MD
530
531 switch (rtm->rtm_type) {
984263bc 532 case RTM_ADD:
ecdefdda
MD
533 if (rtinfo.rti_gateway == NULL) {
534 error = EINVAL;
535 } else {
536 error = rtrequest1_global(RTM_ADD, &rtinfo,
537 route_output_add_callback, rtm);
984263bc
MD
538 }
539 break;
984263bc 540 case RTM_DELETE:
ecdefdda
MD
541 /*
542 * note: &rtm passed as argument so 'rtm' can be replaced.
543 */
544 error = rtrequest1_global(RTM_DELETE, &rtinfo,
545 route_output_delete_callback, &rtm);
984263bc 546 break;
984263bc 547 case RTM_GET:
00104f14
SZ
548 /*
549 * note: &rtm passed as argument so 'rtm' can be replaced.
550 */
551 error = rtsearch_global(RTM_GET, &rtinfo,
552 route_output_get_callback, &rtm,
553 RTS_NOEXACTMATCH);
ecdefdda
MD
554 break;
555 case RTM_CHANGE:
2c62ae3b
SZ
556 error = rtsearch_global(RTM_CHANGE, &rtinfo,
557 route_output_change_callback, rtm,
558 RTS_EXACTMATCH);
ecdefdda
MD
559 break;
560 case RTM_LOCK:
ae2a5ccd
SZ
561 error = rtsearch_global(RTM_LOCK, &rtinfo,
562 route_output_lock_callback, rtm,
563 RTS_EXACTMATCH);
590b8cd4 564 break;
984263bc 565 default:
ecdefdda
MD
566 error = EOPNOTSUPP;
567 break;
984263bc
MD
568 }
569
570flush:
590b8cd4
JH
571 if (rtm != NULL) {
572 if (error != 0)
984263bc
MD
573 rtm->rtm_errno = error;
574 else
575 rtm->rtm_flags |= RTF_DONE;
576 }
ecdefdda 577
984263bc
MD
578 /*
579 * Check to see if we don't want our own messages.
580 */
ef87f48d 581 if (!(so->so_options & SO_USELOOPBACK)) {
984263bc 582 if (route_cb.any_count <= 1) {
590b8cd4 583 if (rtm != NULL)
efda3bd0 584 kfree(rtm, M_RTABLE);
984263bc
MD
585 m_freem(m);
586 return (error);
587 }
588 /* There is another listener, so construct message */
589 rp = sotorawcb(so);
590 }
590b8cd4 591 if (rtm != NULL) {
984263bc
MD
592 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm);
593 if (m->m_pkthdr.len < rtm->rtm_msglen) {
594 m_freem(m);
595 m = NULL;
596 } else if (m->m_pkthdr.len > rtm->rtm_msglen)
597 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len);
efda3bd0 598 kfree(rtm, M_RTABLE);
984263bc 599 }
ef87f48d 600 if (m != NULL)
a8c218b2 601 rts_input_skip(m, family, rp);
984263bc
MD
602 return (error);
603}
604
ecdefdda
MD
605static void
606route_output_add_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
607 struct rtentry *rt, void *arg)
608{
609 struct rt_msghdr *rtm = arg;
610
611 if (error == 0 && rt != NULL) {
612 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx,
613 &rt->rt_rmx);
614 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits);
615 rt->rt_rmx.rmx_locks |=
616 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks);
ac0f13c5
SZ
617 if (rtinfo->rti_genmask != NULL) {
618 rt->rt_genmask = rtmask_purelookup(rtinfo->rti_genmask);
619 if (rt->rt_genmask == NULL) {
620 /*
621 * This should not happen, since we
622 * have already installed genmask
623 * on each CPU before we reach here.
624 */
625 panic("genmask is gone!?");
626 }
627 } else {
628 rt->rt_genmask = NULL;
629 }
31f89d97 630 rtm->rtm_index = rt->rt_ifp->if_index;
ecdefdda
MD
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
00104f14
SZ
650static int
651route_output_get_callback(int cmd, struct rt_addrinfo *rtinfo,
b5e01dd3 652 struct rtentry *rt, void *arg, int found_cnt)
00104f14
SZ
653{
654 struct rt_msghdr **rtm = arg;
655 int error, found = 0;
656
657 if (((rtinfo->rti_flags ^ rt->rt_flags) & RTF_HOST) == 0)
658 found = 1;
659
660 error = fillrtmsg(rtm, rt, rtinfo);
661 if (!error && found) {
662 /* Got the exact match, we could return now! */
663 error = EJUSTRETURN;
664 }
665 return error;
666}
667
2c62ae3b
SZ
668static int
669route_output_change_callback(int cmd, struct rt_addrinfo *rtinfo,
b5e01dd3 670 struct rtentry *rt, void *arg, int found_cnt)
ecdefdda
MD
671{
672 struct rt_msghdr *rtm = arg;
673 struct ifaddr *ifa;
2c62ae3b 674 int error = 0;
ecdefdda
MD
675
676 /*
677 * new gateway could require new ifaddr, ifp;
678 * flags may also be different; ifp may be specified
679 * by ll sockaddr when protocol address is ambiguous
680 */
681 if (((rt->rt_flags & RTF_GATEWAY) && rtinfo->rti_gateway != NULL) ||
fb41af81
SZ
682 rtinfo->rti_ifpaddr != NULL ||
683 (rtinfo->rti_ifaaddr != NULL &&
684 !sa_equal(rtinfo->rti_ifaaddr, rt->rt_ifa->ifa_addr))) {
ecdefdda
MD
685 error = rt_getifa(rtinfo);
686 if (error != 0)
687 goto done;
688 }
689 if (rtinfo->rti_gateway != NULL) {
b5e01dd3
SZ
690 /*
691 * We only need to generate rtmsg upon the
692 * first route to be changed.
693 */
afe44b68 694 error = rt_setgate(rt, rt_key(rt), rtinfo->rti_gateway,
b5e01dd3 695 found_cnt == 1 ? RTL_REPORTMSG : RTL_DONTREPORT);
ecdefdda
MD
696 if (error != 0)
697 goto done;
698 }
699 if ((ifa = rtinfo->rti_ifa) != NULL) {
700 struct ifaddr *oifa = rt->rt_ifa;
701
702 if (oifa != ifa) {
703 if (oifa && oifa->ifa_rtrequest)
704 oifa->ifa_rtrequest(RTM_DELETE, rt, rtinfo);
705 IFAFREE(rt->rt_ifa);
706 IFAREF(ifa);
707 rt->rt_ifa = ifa;
708 rt->rt_ifp = rtinfo->rti_ifp;
709 }
710 }
711 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, &rt->rt_rmx);
712 if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest)
2c62ae3b 713 rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, rtinfo);
ac0f13c5
SZ
714 if (rtinfo->rti_genmask != NULL) {
715 rt->rt_genmask = rtmask_purelookup(rtinfo->rti_genmask);
716 if (rt->rt_genmask == NULL) {
717 /*
718 * This should not happen, since we
719 * have already installed genmask
720 * on each CPU before we reach here.
721 */
722 panic("genmask is gone!?\n");
723 }
724 }
31f89d97 725 rtm->rtm_index = rt->rt_ifp->if_index;
ecdefdda 726done:
2c62ae3b 727 return error;
ecdefdda
MD
728}
729
ae2a5ccd
SZ
730static int
731route_output_lock_callback(int cmd, struct rt_addrinfo *rtinfo,
b5e01dd3
SZ
732 struct rtentry *rt, void *arg,
733 int found_cnt __unused)
ecdefdda
MD
734{
735 struct rt_msghdr *rtm = arg;
736
737 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits);
738 rt->rt_rmx.rmx_locks |=
739 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks);
ae2a5ccd 740 return 0;
ecdefdda
MD
741}
742
984263bc 743static void
ef87f48d 744rt_setmetrics(u_long which, struct rt_metrics *in, struct rt_metrics *out)
984263bc 745{
2e9572df
JH
746#define setmetric(flag, elt) if (which & (flag)) out->elt = in->elt;
747 setmetric(RTV_RPIPE, rmx_recvpipe);
748 setmetric(RTV_SPIPE, rmx_sendpipe);
749 setmetric(RTV_SSTHRESH, rmx_ssthresh);
750 setmetric(RTV_RTT, rmx_rtt);
751 setmetric(RTV_RTTVAR, rmx_rttvar);
752 setmetric(RTV_HOPCOUNT, rmx_hopcount);
753 setmetric(RTV_MTU, rmx_mtu);
754 setmetric(RTV_EXPIRE, rmx_expire);
755#undef setmetric
984263bc
MD
756}
757
758#define ROUNDUP(a) \
759 ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long))
984263bc 760
984263bc
MD
761/*
762 * Extract the addresses of the passed sockaddrs.
763 * Do a little sanity checking so as to avoid bad memory references.
764 * This data is derived straight from userland.
765 */
766static int
ef87f48d 767rt_xaddrs(char *cp, char *cplim, struct rt_addrinfo *rtinfo)
984263bc 768{
82ed7fc2
RG
769 struct sockaddr *sa;
770 int i;
984263bc
MD
771
772 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) {
773 if ((rtinfo->rti_addrs & (1 << i)) == 0)
774 continue;
775 sa = (struct sockaddr *)cp;
776 /*
777 * It won't fit.
778 */
0c3c561c 779 if ((cp + sa->sa_len) > cplim) {
984263bc
MD
780 return (EINVAL);
781 }
782
783 /*
ef87f48d 784 * There are no more... Quit now.
984263bc 785 * If there are more bits, they are in error.
ef87f48d 786 * I've seen this. route(1) can evidently generate these.
984263bc 787 * This causes kernel to core dump.
ef87f48d 788 * For compatibility, if we see this, point to a safe address.
984263bc
MD
789 */
790 if (sa->sa_len == 0) {
590b8cd4
JH
791 static struct sockaddr sa_zero = {
792 sizeof sa_zero, AF_INET,
793 };
794
984263bc 795 rtinfo->rti_info[i] = &sa_zero;
9b42cabe 796 kprintf("rtsock: received more addr bits than sockaddrs.\n");
984263bc
MD
797 return (0); /* should be EINVAL but for compat */
798 }
799
ef87f48d 800 /* Accept the sockaddr. */
984263bc 801 rtinfo->rti_info[i] = sa;
590b8cd4 802 cp += ROUNDUP(sa->sa_len);
984263bc
MD
803 }
804 return (0);
805}
806
590b8cd4
JH
807static int
808rt_msghdrsize(int type)
984263bc 809{
984263bc 810 switch (type) {
984263bc
MD
811 case RTM_DELADDR:
812 case RTM_NEWADDR:
590b8cd4 813 return sizeof(struct ifa_msghdr);
984263bc
MD
814 case RTM_DELMADDR:
815 case RTM_NEWMADDR:
590b8cd4 816 return sizeof(struct ifma_msghdr);
984263bc 817 case RTM_IFINFO:
590b8cd4 818 return sizeof(struct if_msghdr);
984263bc 819 case RTM_IFANNOUNCE:
841ab66c 820 case RTM_IEEE80211:
590b8cd4 821 return sizeof(struct if_announcemsghdr);
984263bc 822 default:
590b8cd4 823 return sizeof(struct rt_msghdr);
984263bc 824 }
590b8cd4
JH
825}
826
827static int
828rt_msgsize(int type, struct rt_addrinfo *rtinfo)
829{
830 int len, i;
831
832 len = rt_msghdrsize(type);
833 for (i = 0; i < RTAX_MAX; i++) {
834 if (rtinfo->rti_info[i] != NULL)
835 len += ROUNDUP(rtinfo->rti_info[i]->sa_len);
984263bc 836 }
590b8cd4
JH
837 len = ALIGN(len);
838 return len;
839}
840
841/*
842 * Build a routing message in a buffer.
843 * Copy the addresses in the rtinfo->rti_info[] sockaddr array
844 * to the end of the buffer after the message header.
845 *
846 * Set the rtinfo->rti_addrs bitmask of addresses present in rtinfo->rti_info[].
847 * This side-effect can be avoided if we reorder the addrs bitmask field in all
848 * the route messages to line up so we can set it here instead of back in the
849 * calling routine.
850 */
851static void
852rt_msg_buffer(int type, struct rt_addrinfo *rtinfo, void *buf, int msglen)
853{
854 struct rt_msghdr *rtm;
855 char *cp;
856 int dlen, i;
857
858 rtm = (struct rt_msghdr *) buf;
859 rtm->rtm_version = RTM_VERSION;
860 rtm->rtm_type = type;
861 rtm->rtm_msglen = msglen;
862
863 cp = (char *)buf + rt_msghdrsize(type);
864 rtinfo->rti_addrs = 0;
984263bc 865 for (i = 0; i < RTAX_MAX; i++) {
590b8cd4
JH
866 struct sockaddr *sa;
867
984263bc
MD
868 if ((sa = rtinfo->rti_info[i]) == NULL)
869 continue;
870 rtinfo->rti_addrs |= (1 << i);
871 dlen = ROUNDUP(sa->sa_len);
590b8cd4
JH
872 bcopy(sa, cp, dlen);
873 cp += dlen;
984263bc 874 }
984263bc
MD
875}
876
590b8cd4
JH
877/*
878 * Build a routing message in a mbuf chain.
879 * Copy the addresses in the rtinfo->rti_info[] sockaddr array
880 * to the end of the mbuf after the message header.
881 *
882 * Set the rtinfo->rti_addrs bitmask of addresses present in rtinfo->rti_info[].
883 * This side-effect can be avoided if we reorder the addrs bitmask field in all
884 * the route messages to line up so we can set it here instead of back in the
885 * calling routine.
886 */
887static struct mbuf *
888rt_msg_mbuf(int type, struct rt_addrinfo *rtinfo)
984263bc 889{
590b8cd4
JH
890 struct mbuf *m;
891 struct rt_msghdr *rtm;
892 int hlen, len;
82ed7fc2 893 int i;
984263bc 894
590b8cd4
JH
895 hlen = rt_msghdrsize(type);
896 KASSERT(hlen <= MCLBYTES, ("rt_msg_mbuf: hlen %d doesn't fit", hlen));
984263bc 897
50503f0f 898 m = m_getl(hlen, MB_DONTWAIT, MT_DATA, M_PKTHDR, NULL);
590b8cd4
JH
899 if (m == NULL)
900 return (NULL);
e9fa4b60 901 mbuftrackid(m, 32);
590b8cd4
JH
902 m->m_pkthdr.len = m->m_len = hlen;
903 m->m_pkthdr.rcvif = NULL;
904 rtinfo->rti_addrs = 0;
905 len = hlen;
984263bc 906 for (i = 0; i < RTAX_MAX; i++) {
82ed7fc2 907 struct sockaddr *sa;
590b8cd4 908 int dlen;
984263bc 909
ef87f48d 910 if ((sa = rtinfo->rti_info[i]) == NULL)
984263bc
MD
911 continue;
912 rtinfo->rti_addrs |= (1 << i);
913 dlen = ROUNDUP(sa->sa_len);
590b8cd4 914 m_copyback(m, len, dlen, (caddr_t)sa); /* can grow mbuf chain */
984263bc
MD
915 len += dlen;
916 }
590b8cd4
JH
917 if (m->m_pkthdr.len != len) { /* one of the m_copyback() calls failed */
918 m_freem(m);
919 return (NULL);
984263bc 920 }
590b8cd4
JH
921 rtm = mtod(m, struct rt_msghdr *);
922 bzero(rtm, hlen);
923 rtm->rtm_msglen = len;
924 rtm->rtm_version = RTM_VERSION;
925 rtm->rtm_type = type;
926 return (m);
984263bc
MD
927}
928
929/*
930 * This routine is called to generate a message from the routing
f23061d4 931 * socket indicating that a redirect has occurred, a routing lookup
984263bc
MD
932 * has failed, or that a protocol has detected timeouts to a particular
933 * destination.
934 */
935void
f23061d4 936rt_missmsg(int type, struct rt_addrinfo *rtinfo, int flags, int error)
984263bc 937{
0c3c561c 938 struct sockaddr *dst = rtinfo->rti_info[RTAX_DST];
82ed7fc2
RG
939 struct rt_msghdr *rtm;
940 struct mbuf *m;
984263bc
MD
941
942 if (route_cb.any_count == 0)
943 return;
590b8cd4 944 m = rt_msg_mbuf(type, rtinfo);
ef87f48d 945 if (m == NULL)
984263bc
MD
946 return;
947 rtm = mtod(m, struct rt_msghdr *);
948 rtm->rtm_flags = RTF_DONE | flags;
949 rtm->rtm_errno = error;
950 rtm->rtm_addrs = rtinfo->rti_addrs;
590b8cd4 951 rts_input(m, familyof(dst));
984263bc
MD
952}
953
f3ed2586
JH
954void
955rt_dstmsg(int type, struct sockaddr *dst, int error)
956{
957 struct rt_msghdr *rtm;
958 struct rt_addrinfo addrs;
959 struct mbuf *m;
960
961 if (route_cb.any_count == 0)
962 return;
963 bzero(&addrs, sizeof(struct rt_addrinfo));
964 addrs.rti_info[RTAX_DST] = dst;
590b8cd4 965 m = rt_msg_mbuf(type, &addrs);
f3ed2586
JH
966 if (m == NULL)
967 return;
968 rtm = mtod(m, struct rt_msghdr *);
969 rtm->rtm_flags = RTF_DONE;
970 rtm->rtm_errno = error;
971 rtm->rtm_addrs = addrs.rti_addrs;
590b8cd4 972 rts_input(m, familyof(dst));
f3ed2586
JH
973}
974
984263bc
MD
975/*
976 * This routine is called to generate a message from the routing
977 * socket indicating that the status of a network interface has changed.
978 */
979void
0c3c561c 980rt_ifmsg(struct ifnet *ifp)
984263bc 981{
82ed7fc2 982 struct if_msghdr *ifm;
984263bc 983 struct mbuf *m;
590b8cd4 984 struct rt_addrinfo rtinfo;
984263bc
MD
985
986 if (route_cb.any_count == 0)
987 return;
590b8cd4
JH
988 bzero(&rtinfo, sizeof(struct rt_addrinfo));
989 m = rt_msg_mbuf(RTM_IFINFO, &rtinfo);
ef87f48d 990 if (m == NULL)
984263bc
MD
991 return;
992 ifm = mtod(m, struct if_msghdr *);
993 ifm->ifm_index = ifp->if_index;
9c095379 994 ifm->ifm_flags = ifp->if_flags;
984263bc 995 ifm->ifm_data = ifp->if_data;
590b8cd4
JH
996 ifm->ifm_addrs = 0;
997 rts_input(m, 0);
984263bc
MD
998}
999
372316d9
JH
1000static void
1001rt_ifamsg(int cmd, struct ifaddr *ifa)
1002{
1003 struct ifa_msghdr *ifam;
590b8cd4 1004 struct rt_addrinfo rtinfo;
372316d9 1005 struct mbuf *m;
372316d9
JH
1006 struct ifnet *ifp = ifa->ifa_ifp;
1007
590b8cd4
JH
1008 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1009 rtinfo.rti_ifaaddr = ifa->ifa_addr;
b2632176
SZ
1010 rtinfo.rti_ifpaddr =
1011 TAILQ_FIRST(&ifp->if_addrheads[mycpuid])->ifa->ifa_addr;
590b8cd4
JH
1012 rtinfo.rti_netmask = ifa->ifa_netmask;
1013 rtinfo.rti_bcastaddr = ifa->ifa_dstaddr;
f23061d4 1014
590b8cd4 1015 m = rt_msg_mbuf(cmd, &rtinfo);
f23061d4 1016 if (m == NULL)
372316d9 1017 return;
f23061d4 1018
372316d9
JH
1019 ifam = mtod(m, struct ifa_msghdr *);
1020 ifam->ifam_index = ifp->if_index;
1021 ifam->ifam_metric = ifa->ifa_metric;
1022 ifam->ifam_flags = ifa->ifa_flags;
590b8cd4 1023 ifam->ifam_addrs = rtinfo.rti_addrs;
f23061d4 1024
590b8cd4 1025 rts_input(m, familyof(ifa->ifa_addr));
372316d9
JH
1026}
1027
f3ed2586
JH
1028void
1029rt_rtmsg(int cmd, struct rtentry *rt, struct ifnet *ifp, int error)
372316d9
JH
1030{
1031 struct rt_msghdr *rtm;
590b8cd4 1032 struct rt_addrinfo rtinfo;
372316d9 1033 struct mbuf *m;
f3ed2586 1034 struct sockaddr *dst;
372316d9
JH
1035
1036 if (rt == NULL)
1037 return;
f23061d4 1038
590b8cd4
JH
1039 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1040 rtinfo.rti_dst = dst = rt_key(rt);
1041 rtinfo.rti_gateway = rt->rt_gateway;
1042 rtinfo.rti_netmask = rt_mask(rt);
b2632176
SZ
1043 if (ifp != NULL) {
1044 rtinfo.rti_ifpaddr =
1045 TAILQ_FIRST(&ifp->if_addrheads[mycpuid])->ifa->ifa_addr;
1046 }
590b8cd4 1047 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr;
f23061d4 1048
590b8cd4 1049 m = rt_msg_mbuf(cmd, &rtinfo);
f23061d4 1050 if (m == NULL)
372316d9 1051 return;
f23061d4 1052
372316d9 1053 rtm = mtod(m, struct rt_msghdr *);
f3ed2586
JH
1054 if (ifp != NULL)
1055 rtm->rtm_index = ifp->if_index;
372316d9
JH
1056 rtm->rtm_flags |= rt->rt_flags;
1057 rtm->rtm_errno = error;
590b8cd4 1058 rtm->rtm_addrs = rtinfo.rti_addrs;
f23061d4 1059
590b8cd4 1060 rts_input(m, familyof(dst));
372316d9
JH
1061}
1062
984263bc
MD
1063/*
1064 * This is called to generate messages from the routing socket
1065 * indicating a network interface has had addresses associated with it.
1066 * if we ever reverse the logic and replace messages TO the routing
1067 * socket indicate a request to configure interfaces, then it will
1068 * be unnecessary as the routing socket will automatically generate
1069 * copies of it.
1070 */
1071void
0c3c561c 1072rt_newaddrmsg(int cmd, struct ifaddr *ifa, int error, struct rtentry *rt)
984263bc 1073{
78812139
EN
1074#ifdef SCTP
1075 /*
1076 * notify the SCTP stack
1077 * this will only get called when an address is added/deleted
1078 * XXX pass the ifaddr struct instead if ifa->ifa_addr...
1079 */
1080 if (cmd == RTM_ADD)
1081 sctp_add_ip_address(ifa);
1082 else if (cmd == RTM_DELETE)
1083 sctp_delete_ip_address(ifa);
1084#endif /* SCTP */
1085
984263bc
MD
1086 if (route_cb.any_count == 0)
1087 return;
984263bc 1088
372316d9
JH
1089 if (cmd == RTM_ADD) {
1090 rt_ifamsg(RTM_NEWADDR, ifa);
f3ed2586 1091 rt_rtmsg(RTM_ADD, rt, ifa->ifa_ifp, error);
372316d9
JH
1092 } else {
1093 KASSERT((cmd == RTM_DELETE), ("unknown cmd %d", cmd));
f3ed2586 1094 rt_rtmsg(RTM_DELETE, rt, ifa->ifa_ifp, error);
372316d9 1095 rt_ifamsg(RTM_DELADDR, ifa);
984263bc
MD
1096 }
1097}
1098
1099/*
1100 * This is the analogue to the rt_newaddrmsg which performs the same
1101 * function but for multicast group memberhips. This is easier since
1102 * there is no route state to worry about.
1103 */
1104void
0c3c561c 1105rt_newmaddrmsg(int cmd, struct ifmultiaddr *ifma)
984263bc 1106{
590b8cd4 1107 struct rt_addrinfo rtinfo;
ef87f48d 1108 struct mbuf *m = NULL;
984263bc
MD
1109 struct ifnet *ifp = ifma->ifma_ifp;
1110 struct ifma_msghdr *ifmam;
1111
1112 if (route_cb.any_count == 0)
1113 return;
1114
590b8cd4
JH
1115 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1116 rtinfo.rti_ifaaddr = ifma->ifma_addr;
b2632176
SZ
1117 if (ifp != NULL && !TAILQ_EMPTY(&ifp->if_addrheads[mycpuid])) {
1118 rtinfo.rti_ifpaddr =
1119 TAILQ_FIRST(&ifp->if_addrheads[mycpuid])->ifa->ifa_addr;
1120 }
984263bc
MD
1121 /*
1122 * If a link-layer address is present, present it as a ``gateway''
1123 * (similarly to how ARP entries, e.g., are presented).
1124 */
590b8cd4 1125 rtinfo.rti_gateway = ifma->ifma_lladdr;
f23061d4 1126
590b8cd4 1127 m = rt_msg_mbuf(cmd, &rtinfo);
f23061d4 1128 if (m == NULL)
984263bc 1129 return;
f23061d4 1130
984263bc
MD
1131 ifmam = mtod(m, struct ifma_msghdr *);
1132 ifmam->ifmam_index = ifp->if_index;
590b8cd4 1133 ifmam->ifmam_addrs = rtinfo.rti_addrs;
f23061d4 1134
590b8cd4 1135 rts_input(m, familyof(ifma->ifma_addr));
984263bc
MD
1136}
1137
841ab66c
SZ
1138static struct mbuf *
1139rt_makeifannouncemsg(struct ifnet *ifp, int type, int what,
1140 struct rt_addrinfo *info)
984263bc 1141{
590b8cd4 1142 struct if_announcemsghdr *ifan;
841ab66c 1143 struct mbuf *m;
984263bc
MD
1144
1145 if (route_cb.any_count == 0)
841ab66c 1146 return NULL;
f23061d4 1147
841ab66c
SZ
1148 bzero(info, sizeof(*info));
1149 m = rt_msg_mbuf(type, info);
984263bc 1150 if (m == NULL)
841ab66c 1151 return NULL;
f23061d4 1152
984263bc
MD
1153 ifan = mtod(m, struct if_announcemsghdr *);
1154 ifan->ifan_index = ifp->if_index;
5fe66e68 1155 strlcpy(ifan->ifan_name, ifp->if_xname, sizeof ifan->ifan_name);
984263bc 1156 ifan->ifan_what = what;
841ab66c
SZ
1157 return m;
1158}
1159
1160/*
1161 * This is called to generate routing socket messages indicating
1162 * IEEE80211 wireless events.
1163 * XXX we piggyback on the RTM_IFANNOUNCE msg format in a clumsy way.
1164 */
1165void
1166rt_ieee80211msg(struct ifnet *ifp, int what, void *data, size_t data_len)
1167{
1168 struct rt_addrinfo info;
1169 struct mbuf *m;
f23061d4 1170
841ab66c
SZ
1171 m = rt_makeifannouncemsg(ifp, RTM_IEEE80211, what, &info);
1172 if (m == NULL)
1173 return;
1174
1175 /*
1176 * Append the ieee80211 data. Try to stick it in the
1177 * mbuf containing the ifannounce msg; otherwise allocate
1178 * a new mbuf and append.
1179 *
1180 * NB: we assume m is a single mbuf.
1181 */
1182 if (data_len > M_TRAILINGSPACE(m)) {
b11e1686 1183 struct mbuf *n = m_get(MB_DONTWAIT, MT_DATA);
841ab66c
SZ
1184 if (n == NULL) {
1185 m_freem(m);
1186 return;
1187 }
1188 bcopy(data, mtod(n, void *), data_len);
1189 n->m_len = data_len;
1190 m->m_next = n;
1191 } else if (data_len > 0) {
1192 bcopy(data, mtod(m, u_int8_t *) + m->m_len, data_len);
1193 m->m_len += data_len;
1194 }
e9fa4b60 1195 mbuftrackid(m, 33);
841ab66c
SZ
1196 if (m->m_flags & M_PKTHDR)
1197 m->m_pkthdr.len += data_len;
1198 mtod(m, struct if_announcemsghdr *)->ifan_msglen += data_len;
590b8cd4 1199 rts_input(m, 0);
841ab66c
SZ
1200}
1201
1202/*
1203 * This is called to generate routing socket messages indicating
1204 * network interface arrival and departure.
1205 */
1206void
1207rt_ifannouncemsg(struct ifnet *ifp, int what)
1208{
1209 struct rt_addrinfo addrinfo;
1210 struct mbuf *m;
1211
1212 m = rt_makeifannouncemsg(ifp, RTM_IFANNOUNCE, what, &addrinfo);
1213 if (m != NULL)
1214 rts_input(m, 0);
590b8cd4 1215}
f23061d4 1216
590b8cd4
JH
1217static int
1218resizewalkarg(struct walkarg *w, int len)
1219{
1220 void *newptr;
1221
efda3bd0 1222 newptr = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
590b8cd4
JH
1223 if (newptr == NULL)
1224 return (ENOMEM);
1225 if (w->w_tmem != NULL)
efda3bd0 1226 kfree(w->w_tmem, M_RTABLE);
590b8cd4
JH
1227 w->w_tmem = newptr;
1228 w->w_tmemsize = len;
1229 return (0);
1230}
984263bc
MD
1231
1232/*
1233 * This is used in dumping the kernel table via sysctl().
1234 */
1235int
590b8cd4 1236sysctl_dumpentry(struct radix_node *rn, void *vw)
984263bc 1237{
82ed7fc2
RG
1238 struct walkarg *w = vw;
1239 struct rtentry *rt = (struct rtentry *)rn;
590b8cd4
JH
1240 struct rt_addrinfo rtinfo;
1241 int error, msglen;
984263bc
MD
1242
1243 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg))
1244 return 0;
f23061d4 1245
590b8cd4
JH
1246 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1247 rtinfo.rti_dst = rt_key(rt);
1248 rtinfo.rti_gateway = rt->rt_gateway;
1249 rtinfo.rti_netmask = rt_mask(rt);
1250 rtinfo.rti_genmask = rt->rt_genmask;
ef87f48d 1251 if (rt->rt_ifp != NULL) {
590b8cd4 1252 rtinfo.rti_ifpaddr =
b2632176 1253 TAILQ_FIRST(&rt->rt_ifp->if_addrheads[mycpuid])->ifa->ifa_addr;
590b8cd4 1254 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr;
984263bc 1255 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT)
590b8cd4 1256 rtinfo.rti_bcastaddr = rt->rt_ifa->ifa_dstaddr;
984263bc 1257 }
590b8cd4
JH
1258 msglen = rt_msgsize(RTM_GET, &rtinfo);
1259 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0)
1260 return (ENOMEM);
1261 rt_msg_buffer(RTM_GET, &rtinfo, w->w_tmem, msglen);
1262 if (w->w_req != NULL) {
8e63efec 1263 struct rt_msghdr *rtm = w->w_tmem;
984263bc
MD
1264
1265 rtm->rtm_flags = rt->rt_flags;
1266 rtm->rtm_use = rt->rt_use;
1267 rtm->rtm_rmx = rt->rt_rmx;
1268 rtm->rtm_index = rt->rt_ifp->if_index;
1269 rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0;
590b8cd4
JH
1270 rtm->rtm_addrs = rtinfo.rti_addrs;
1271 error = SYSCTL_OUT(w->w_req, rtm, msglen);
984263bc
MD
1272 return (error);
1273 }
590b8cd4 1274 return (0);
984263bc
MD
1275}
1276
590b8cd4
JH
1277static int
1278sysctl_iflist(int af, struct walkarg *w)
984263bc 1279{
82ed7fc2 1280 struct ifnet *ifp;
590b8cd4
JH
1281 struct rt_addrinfo rtinfo;
1282 int msglen, error;
984263bc 1283
590b8cd4 1284 bzero(&rtinfo, sizeof(struct rt_addrinfo));
984263bc 1285 TAILQ_FOREACH(ifp, &ifnet, if_link) {
b2632176
SZ
1286 struct ifaddr_container *ifac;
1287 struct ifaddr *ifa;
1288
984263bc
MD
1289 if (w->w_arg && w->w_arg != ifp->if_index)
1290 continue;
b2632176
SZ
1291 ifac = TAILQ_FIRST(&ifp->if_addrheads[mycpuid]);
1292 ifa = ifac->ifa;
590b8cd4
JH
1293 rtinfo.rti_ifpaddr = ifa->ifa_addr;
1294 msglen = rt_msgsize(RTM_IFINFO, &rtinfo);
1295 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0)
1296 return (ENOMEM);
1297 rt_msg_buffer(RTM_IFINFO, &rtinfo, w->w_tmem, msglen);
1298 rtinfo.rti_ifpaddr = NULL;
ef87f48d 1299 if (w->w_req != NULL && w->w_tmem != NULL) {
8e63efec 1300 struct if_msghdr *ifm = w->w_tmem;
984263bc 1301
984263bc 1302 ifm->ifm_index = ifp->if_index;
9c095379 1303 ifm->ifm_flags = ifp->if_flags;
984263bc 1304 ifm->ifm_data = ifp->if_data;
590b8cd4
JH
1305 ifm->ifm_addrs = rtinfo.rti_addrs;
1306 error = SYSCTL_OUT(w->w_req, ifm, msglen);
984263bc
MD
1307 if (error)
1308 return (error);
1309 }
b2632176
SZ
1310 while ((ifac = TAILQ_NEXT(ifac, ifa_link)) != NULL) {
1311 ifa = ifac->ifa;
1312
984263bc
MD
1313 if (af && af != ifa->ifa_addr->sa_family)
1314 continue;
590b8cd4 1315 if (curproc->p_ucred->cr_prison &&
87de5057 1316 prison_if(curproc->p_ucred, ifa->ifa_addr))
984263bc 1317 continue;
590b8cd4
JH
1318 rtinfo.rti_ifaaddr = ifa->ifa_addr;
1319 rtinfo.rti_netmask = ifa->ifa_netmask;
1320 rtinfo.rti_bcastaddr = ifa->ifa_dstaddr;
1321 msglen = rt_msgsize(RTM_NEWADDR, &rtinfo);
1322 if (w->w_tmemsize < msglen &&
1323 resizewalkarg(w, msglen) != 0)
1324 return (ENOMEM);
1325 rt_msg_buffer(RTM_NEWADDR, &rtinfo, w->w_tmem, msglen);
1326 if (w->w_req != NULL) {
8e63efec 1327 struct ifa_msghdr *ifam = w->w_tmem;
984263bc 1328
984263bc
MD
1329 ifam->ifam_index = ifa->ifa_ifp->if_index;
1330 ifam->ifam_flags = ifa->ifa_flags;
1331 ifam->ifam_metric = ifa->ifa_metric;
590b8cd4
JH
1332 ifam->ifam_addrs = rtinfo.rti_addrs;
1333 error = SYSCTL_OUT(w->w_req, w->w_tmem, msglen);
984263bc
MD
1334 if (error)
1335 return (error);
1336 }
1337 }
590b8cd4
JH
1338 rtinfo.rti_netmask = NULL;
1339 rtinfo.rti_ifaaddr = NULL;
1340 rtinfo.rti_bcastaddr = NULL;
984263bc
MD
1341 }
1342 return (0);
1343}
1344
1345static int
1346sysctl_rtsock(SYSCTL_HANDLER_ARGS)
1347{
1348 int *name = (int *)arg1;
1349 u_int namelen = arg2;
82ed7fc2 1350 struct radix_node_head *rnh;
a49aa710 1351 int i, error = EINVAL;
ecdefdda 1352 int origcpu;
984263bc
MD
1353 u_char af;
1354 struct walkarg w;
1355
1356 name ++;
1357 namelen--;
1358 if (req->newptr)
1359 return (EPERM);
ecdefdda 1360 if (namelen != 3 && namelen != 4)
984263bc
MD
1361 return (EINVAL);
1362 af = name[0];
5fe66e68 1363 bzero(&w, sizeof w);
984263bc
MD
1364 w.w_op = name[1];
1365 w.w_arg = name[2];
1366 w.w_req = req;
1367
ecdefdda
MD
1368 /*
1369 * Optional third argument specifies cpu, used primarily for
1370 * debugging the route table.
1371 */
1372 if (namelen == 4) {
1373 if (name[3] < 0 || name[3] >= ncpus)
1374 return (EINVAL);
1375 origcpu = mycpuid;
1376 lwkt_migratecpu(name[3]);
1377 } else {
1378 origcpu = -1;
1379 }
a49aa710 1380 crit_enter();
984263bc 1381 switch (w.w_op) {
984263bc
MD
1382 case NET_RT_DUMP:
1383 case NET_RT_FLAGS:
1384 for (i = 1; i <= AF_MAX; i++)
ecdefdda
MD
1385 if ((rnh = rt_tables[mycpuid][i]) &&
1386 (af == 0 || af == i) &&
984263bc 1387 (error = rnh->rnh_walktree(rnh,
f23061d4 1388 sysctl_dumpentry, &w)))
984263bc
MD
1389 break;
1390 break;
1391
1392 case NET_RT_IFLIST:
1393 error = sysctl_iflist(af, &w);
1394 }
a49aa710 1395 crit_exit();
590b8cd4 1396 if (w.w_tmem != NULL)
efda3bd0 1397 kfree(w.w_tmem, M_RTABLE);
ecdefdda
MD
1398 if (origcpu >= 0)
1399 lwkt_migratecpu(origcpu);
984263bc
MD
1400 return (error);
1401}
1402
1403SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, "");
1404
1405/*
1406 * Definitions of protocols supported in the ROUTE domain.
1407 */
1408
ba4257b8 1409static struct domain routedomain; /* or at least forward */
984263bc
MD
1410
1411static struct protosw routesw[] = {
1412{ SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR,
1413 0, route_output, raw_ctlinput, 0,
e3873585 1414 cpu0_soport, cpu0_ctlport,
984263bc
MD
1415 raw_init, 0, 0, 0,
1416 &route_usrreqs
1417}
1418};
1419
ba4257b8 1420static struct domain routedomain = {
9c70fe43
JH
1421 PF_ROUTE, "route", NULL, NULL, NULL,
1422 routesw, &routesw[(sizeof routesw)/(sizeof routesw[0])],
1423};
984263bc
MD
1424
1425DOMAIN_SET(route);
47db7f9b 1426