rtsock: Fill in the correct rtm_index for RTM_ADD and RTM_CHANGE messages.
[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
SZ
458static int route_output_get_callback(int, struct rt_addrinfo *,
459 struct rtentry *, void *);
2c62ae3b 460static int route_output_change_callback(int, struct rt_addrinfo *,
ecdefdda 461 struct rtentry *, void *);
ae2a5ccd
SZ
462static int route_output_lock_callback(int, struct rt_addrinfo *,
463 struct rtentry *, void *);
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;
ef87f48d 473 int len, error = 0;
deffea2e
JS
474 __va_list ap;
475
1cfcc9b6
SZ
476 M_ASSERTPKTHDR(m);
477
deffea2e
JS
478 __va_start(ap, so);
479 oi = __va_arg(ap, struct pr_output_info *);
480 __va_end(ap);
984263bc 481
ef87f48d 482#define gotoerr(e) { error = e; goto flush;}
590b8cd4
JH
483
484 if (m == NULL ||
485 (m->m_len < sizeof(long) &&
486 (m = m_pullup(m, sizeof(long))) == NULL))
984263bc 487 return (ENOBUFS);
984263bc 488 len = m->m_pkthdr.len;
590b8cd4 489 if (len < sizeof(struct rt_msghdr) ||
984263bc 490 len != mtod(m, struct rt_msghdr *)->rtm_msglen) {
590b8cd4 491 rtinfo.rti_dst = NULL;
ef87f48d 492 gotoerr(EINVAL);
984263bc 493 }
efda3bd0 494 rtm = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
ef87f48d 495 if (rtm == NULL) {
590b8cd4 496 rtinfo.rti_dst = NULL;
ef87f48d 497 gotoerr(ENOBUFS);
984263bc
MD
498 }
499 m_copydata(m, 0, len, (caddr_t)rtm);
500 if (rtm->rtm_version != RTM_VERSION) {
590b8cd4 501 rtinfo.rti_dst = NULL;
ef87f48d 502 gotoerr(EPROTONOSUPPORT);
984263bc 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;
507 if (rt_xaddrs((char *)(rtm + 1), (char *)rtm + len, &rtinfo) != 0) {
508 rtinfo.rti_dst = NULL;
ef87f48d 509 gotoerr(EINVAL);
984263bc 510 }
590b8cd4
JH
511 rtinfo.rti_flags = rtm->rtm_flags;
512 if (rtinfo.rti_dst == NULL || rtinfo.rti_dst->sa_family >= AF_MAX ||
513 (rtinfo.rti_gateway && rtinfo.rti_gateway->sa_family >= AF_MAX))
ef87f48d
JH
514 gotoerr(EINVAL);
515
590b8cd4
JH
516 if (rtinfo.rti_genmask != NULL) {
517 struct radix_node *n;
ef87f48d 518
590b8cd4
JH
519#define clen(s) (*(u_char *)(s))
520 n = rn_addmask((char *)rtinfo.rti_genmask, TRUE, 1);
521 if (n != NULL &&
522 rtinfo.rti_genmask->sa_len >= clen(n->rn_key) &&
523 bcmp((char *)rtinfo.rti_genmask + 1,
524 (char *)n->rn_key + 1, clen(n->rn_key) - 1) == 0)
525 rtinfo.rti_genmask = (struct sockaddr *)n->rn_key;
984263bc 526 else
ef87f48d 527 gotoerr(ENOBUFS);
984263bc
MD
528 }
529
530 /*
531 * Verify that the caller has the appropriate privilege; RTM_GET
532 * is the only operation the non-superuser is allowed.
533 */
02650706
SZ
534 if (rtm->rtm_type != RTM_GET &&
535 priv_check_cred(so->so_cred, PRIV_ROOT, 0) != 0)
ef87f48d 536 gotoerr(EPERM);
984263bc
MD
537
538 switch (rtm->rtm_type) {
984263bc 539 case RTM_ADD:
ecdefdda
MD
540 if (rtinfo.rti_gateway == NULL) {
541 error = EINVAL;
542 } else {
543 error = rtrequest1_global(RTM_ADD, &rtinfo,
544 route_output_add_callback, rtm);
984263bc
MD
545 }
546 break;
984263bc 547 case RTM_DELETE:
ecdefdda
MD
548 /*
549 * note: &rtm passed as argument so 'rtm' can be replaced.
550 */
551 error = rtrequest1_global(RTM_DELETE, &rtinfo,
552 route_output_delete_callback, &rtm);
984263bc 553 break;
984263bc 554 case RTM_GET:
00104f14
SZ
555 /*
556 * note: &rtm passed as argument so 'rtm' can be replaced.
557 */
558 error = rtsearch_global(RTM_GET, &rtinfo,
559 route_output_get_callback, &rtm,
560 RTS_NOEXACTMATCH);
ecdefdda
MD
561 break;
562 case RTM_CHANGE:
2c62ae3b
SZ
563 error = rtsearch_global(RTM_CHANGE, &rtinfo,
564 route_output_change_callback, rtm,
565 RTS_EXACTMATCH);
ecdefdda
MD
566 break;
567 case RTM_LOCK:
ae2a5ccd
SZ
568 error = rtsearch_global(RTM_LOCK, &rtinfo,
569 route_output_lock_callback, rtm,
570 RTS_EXACTMATCH);
590b8cd4 571 break;
984263bc 572 default:
ecdefdda
MD
573 error = EOPNOTSUPP;
574 break;
984263bc
MD
575 }
576
577flush:
590b8cd4
JH
578 if (rtm != NULL) {
579 if (error != 0)
984263bc
MD
580 rtm->rtm_errno = error;
581 else
582 rtm->rtm_flags |= RTF_DONE;
583 }
ecdefdda 584
984263bc
MD
585 /*
586 * Check to see if we don't want our own messages.
587 */
ef87f48d 588 if (!(so->so_options & SO_USELOOPBACK)) {
984263bc 589 if (route_cb.any_count <= 1) {
590b8cd4 590 if (rtm != NULL)
efda3bd0 591 kfree(rtm, M_RTABLE);
984263bc
MD
592 m_freem(m);
593 return (error);
594 }
595 /* There is another listener, so construct message */
596 rp = sotorawcb(so);
597 }
590b8cd4 598 if (rtm != NULL) {
984263bc
MD
599 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm);
600 if (m->m_pkthdr.len < rtm->rtm_msglen) {
601 m_freem(m);
602 m = NULL;
603 } else if (m->m_pkthdr.len > rtm->rtm_msglen)
604 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len);
efda3bd0 605 kfree(rtm, M_RTABLE);
984263bc 606 }
ef87f48d 607 if (m != NULL)
d16fa201 608 rts_input_skip(m, familyof(rtinfo.rti_dst), rp);
984263bc
MD
609 return (error);
610}
611
ecdefdda
MD
612static void
613route_output_add_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
614 struct rtentry *rt, void *arg)
615{
616 struct rt_msghdr *rtm = arg;
617
618 if (error == 0 && rt != NULL) {
619 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx,
620 &rt->rt_rmx);
621 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits);
622 rt->rt_rmx.rmx_locks |=
623 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks);
624 rt->rt_genmask = rtinfo->rti_genmask;
31f89d97 625 rtm->rtm_index = rt->rt_ifp->if_index;
ecdefdda
MD
626 }
627}
628
629static void
630route_output_delete_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
631 struct rtentry *rt, void *arg)
632{
633 struct rt_msghdr **rtm = arg;
634
635 if (error == 0 && rt) {
636 ++rt->rt_refcnt;
637 if (fillrtmsg(rtm, rt, rtinfo) != 0) {
638 error = ENOBUFS;
639 /* XXX no way to return the error */
640 }
641 --rt->rt_refcnt;
642 }
643}
644
00104f14
SZ
645static int
646route_output_get_callback(int cmd, struct rt_addrinfo *rtinfo,
647 struct rtentry *rt, void *arg)
648{
649 struct rt_msghdr **rtm = arg;
650 int error, found = 0;
651
652 if (((rtinfo->rti_flags ^ rt->rt_flags) & RTF_HOST) == 0)
653 found = 1;
654
655 error = fillrtmsg(rtm, rt, rtinfo);
656 if (!error && found) {
657 /* Got the exact match, we could return now! */
658 error = EJUSTRETURN;
659 }
660 return error;
661}
662
2c62ae3b
SZ
663static int
664route_output_change_callback(int cmd, struct rt_addrinfo *rtinfo,
665 struct rtentry *rt, void *arg)
ecdefdda
MD
666{
667 struct rt_msghdr *rtm = arg;
668 struct ifaddr *ifa;
2c62ae3b 669 int error = 0;
ecdefdda
MD
670
671 /*
672 * new gateway could require new ifaddr, ifp;
673 * flags may also be different; ifp may be specified
674 * by ll sockaddr when protocol address is ambiguous
675 */
676 if (((rt->rt_flags & RTF_GATEWAY) && rtinfo->rti_gateway != NULL) ||
fb41af81
SZ
677 rtinfo->rti_ifpaddr != NULL ||
678 (rtinfo->rti_ifaaddr != NULL &&
679 !sa_equal(rtinfo->rti_ifaaddr, rt->rt_ifa->ifa_addr))) {
ecdefdda
MD
680 error = rt_getifa(rtinfo);
681 if (error != 0)
682 goto done;
683 }
684 if (rtinfo->rti_gateway != NULL) {
685 error = rt_setgate(rt, rt_key(rt), rtinfo->rti_gateway);
686 if (error != 0)
687 goto done;
688 }
689 if ((ifa = rtinfo->rti_ifa) != NULL) {
690 struct ifaddr *oifa = rt->rt_ifa;
691
692 if (oifa != ifa) {
693 if (oifa && oifa->ifa_rtrequest)
694 oifa->ifa_rtrequest(RTM_DELETE, rt, rtinfo);
695 IFAFREE(rt->rt_ifa);
696 IFAREF(ifa);
697 rt->rt_ifa = ifa;
698 rt->rt_ifp = rtinfo->rti_ifp;
699 }
700 }
701 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, &rt->rt_rmx);
702 if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest)
2c62ae3b 703 rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, rtinfo);
ecdefdda
MD
704 if (rtinfo->rti_genmask != NULL)
705 rt->rt_genmask = rtinfo->rti_genmask;
31f89d97 706 rtm->rtm_index = rt->rt_ifp->if_index;
ecdefdda 707done:
2c62ae3b 708 return error;
ecdefdda
MD
709}
710
ae2a5ccd
SZ
711static int
712route_output_lock_callback(int cmd, struct rt_addrinfo *rtinfo,
ecdefdda
MD
713 struct rtentry *rt, void *arg)
714{
715 struct rt_msghdr *rtm = arg;
716
717 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits);
718 rt->rt_rmx.rmx_locks |=
719 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks);
ae2a5ccd 720 return 0;
ecdefdda
MD
721}
722
984263bc 723static void
ef87f48d 724rt_setmetrics(u_long which, struct rt_metrics *in, struct rt_metrics *out)
984263bc 725{
2e9572df
JH
726#define setmetric(flag, elt) if (which & (flag)) out->elt = in->elt;
727 setmetric(RTV_RPIPE, rmx_recvpipe);
728 setmetric(RTV_SPIPE, rmx_sendpipe);
729 setmetric(RTV_SSTHRESH, rmx_ssthresh);
730 setmetric(RTV_RTT, rmx_rtt);
731 setmetric(RTV_RTTVAR, rmx_rttvar);
732 setmetric(RTV_HOPCOUNT, rmx_hopcount);
733 setmetric(RTV_MTU, rmx_mtu);
734 setmetric(RTV_EXPIRE, rmx_expire);
735#undef setmetric
984263bc
MD
736}
737
738#define ROUNDUP(a) \
739 ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long))
984263bc 740
984263bc
MD
741/*
742 * Extract the addresses of the passed sockaddrs.
743 * Do a little sanity checking so as to avoid bad memory references.
744 * This data is derived straight from userland.
745 */
746static int
ef87f48d 747rt_xaddrs(char *cp, char *cplim, struct rt_addrinfo *rtinfo)
984263bc 748{
82ed7fc2
RG
749 struct sockaddr *sa;
750 int i;
984263bc
MD
751
752 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) {
753 if ((rtinfo->rti_addrs & (1 << i)) == 0)
754 continue;
755 sa = (struct sockaddr *)cp;
756 /*
757 * It won't fit.
758 */
0c3c561c 759 if ((cp + sa->sa_len) > cplim) {
984263bc
MD
760 return (EINVAL);
761 }
762
763 /*
ef87f48d 764 * There are no more... Quit now.
984263bc 765 * If there are more bits, they are in error.
ef87f48d 766 * I've seen this. route(1) can evidently generate these.
984263bc 767 * This causes kernel to core dump.
ef87f48d 768 * For compatibility, if we see this, point to a safe address.
984263bc
MD
769 */
770 if (sa->sa_len == 0) {
590b8cd4
JH
771 static struct sockaddr sa_zero = {
772 sizeof sa_zero, AF_INET,
773 };
774
984263bc 775 rtinfo->rti_info[i] = &sa_zero;
9b42cabe 776 kprintf("rtsock: received more addr bits than sockaddrs.\n");
984263bc
MD
777 return (0); /* should be EINVAL but for compat */
778 }
779
ef87f48d 780 /* Accept the sockaddr. */
984263bc 781 rtinfo->rti_info[i] = sa;
590b8cd4 782 cp += ROUNDUP(sa->sa_len);
984263bc
MD
783 }
784 return (0);
785}
786
590b8cd4
JH
787static int
788rt_msghdrsize(int type)
984263bc 789{
984263bc 790 switch (type) {
984263bc
MD
791 case RTM_DELADDR:
792 case RTM_NEWADDR:
590b8cd4 793 return sizeof(struct ifa_msghdr);
984263bc
MD
794 case RTM_DELMADDR:
795 case RTM_NEWMADDR:
590b8cd4 796 return sizeof(struct ifma_msghdr);
984263bc 797 case RTM_IFINFO:
590b8cd4 798 return sizeof(struct if_msghdr);
984263bc 799 case RTM_IFANNOUNCE:
841ab66c 800 case RTM_IEEE80211:
590b8cd4 801 return sizeof(struct if_announcemsghdr);
984263bc 802 default:
590b8cd4 803 return sizeof(struct rt_msghdr);
984263bc 804 }
590b8cd4
JH
805}
806
807static int
808rt_msgsize(int type, struct rt_addrinfo *rtinfo)
809{
810 int len, i;
811
812 len = rt_msghdrsize(type);
813 for (i = 0; i < RTAX_MAX; i++) {
814 if (rtinfo->rti_info[i] != NULL)
815 len += ROUNDUP(rtinfo->rti_info[i]->sa_len);
984263bc 816 }
590b8cd4
JH
817 len = ALIGN(len);
818 return len;
819}
820
821/*
822 * Build a routing message in a buffer.
823 * Copy the addresses in the rtinfo->rti_info[] sockaddr array
824 * to the end of the buffer after the message header.
825 *
826 * Set the rtinfo->rti_addrs bitmask of addresses present in rtinfo->rti_info[].
827 * This side-effect can be avoided if we reorder the addrs bitmask field in all
828 * the route messages to line up so we can set it here instead of back in the
829 * calling routine.
830 */
831static void
832rt_msg_buffer(int type, struct rt_addrinfo *rtinfo, void *buf, int msglen)
833{
834 struct rt_msghdr *rtm;
835 char *cp;
836 int dlen, i;
837
838 rtm = (struct rt_msghdr *) buf;
839 rtm->rtm_version = RTM_VERSION;
840 rtm->rtm_type = type;
841 rtm->rtm_msglen = msglen;
842
843 cp = (char *)buf + rt_msghdrsize(type);
844 rtinfo->rti_addrs = 0;
984263bc 845 for (i = 0; i < RTAX_MAX; i++) {
590b8cd4
JH
846 struct sockaddr *sa;
847
984263bc
MD
848 if ((sa = rtinfo->rti_info[i]) == NULL)
849 continue;
850 rtinfo->rti_addrs |= (1 << i);
851 dlen = ROUNDUP(sa->sa_len);
590b8cd4
JH
852 bcopy(sa, cp, dlen);
853 cp += dlen;
984263bc 854 }
984263bc
MD
855}
856
590b8cd4
JH
857/*
858 * Build a routing message in a mbuf chain.
859 * Copy the addresses in the rtinfo->rti_info[] sockaddr array
860 * to the end of the mbuf after the message header.
861 *
862 * Set the rtinfo->rti_addrs bitmask of addresses present in rtinfo->rti_info[].
863 * This side-effect can be avoided if we reorder the addrs bitmask field in all
864 * the route messages to line up so we can set it here instead of back in the
865 * calling routine.
866 */
867static struct mbuf *
868rt_msg_mbuf(int type, struct rt_addrinfo *rtinfo)
984263bc 869{
590b8cd4
JH
870 struct mbuf *m;
871 struct rt_msghdr *rtm;
872 int hlen, len;
82ed7fc2 873 int i;
984263bc 874
590b8cd4
JH
875 hlen = rt_msghdrsize(type);
876 KASSERT(hlen <= MCLBYTES, ("rt_msg_mbuf: hlen %d doesn't fit", hlen));
984263bc 877
50503f0f 878 m = m_getl(hlen, MB_DONTWAIT, MT_DATA, M_PKTHDR, NULL);
590b8cd4
JH
879 if (m == NULL)
880 return (NULL);
e9fa4b60 881 mbuftrackid(m, 32);
590b8cd4
JH
882 m->m_pkthdr.len = m->m_len = hlen;
883 m->m_pkthdr.rcvif = NULL;
884 rtinfo->rti_addrs = 0;
885 len = hlen;
984263bc 886 for (i = 0; i < RTAX_MAX; i++) {
82ed7fc2 887 struct sockaddr *sa;
590b8cd4 888 int dlen;
984263bc 889
ef87f48d 890 if ((sa = rtinfo->rti_info[i]) == NULL)
984263bc
MD
891 continue;
892 rtinfo->rti_addrs |= (1 << i);
893 dlen = ROUNDUP(sa->sa_len);
590b8cd4 894 m_copyback(m, len, dlen, (caddr_t)sa); /* can grow mbuf chain */
984263bc
MD
895 len += dlen;
896 }
590b8cd4
JH
897 if (m->m_pkthdr.len != len) { /* one of the m_copyback() calls failed */
898 m_freem(m);
899 return (NULL);
984263bc 900 }
590b8cd4
JH
901 rtm = mtod(m, struct rt_msghdr *);
902 bzero(rtm, hlen);
903 rtm->rtm_msglen = len;
904 rtm->rtm_version = RTM_VERSION;
905 rtm->rtm_type = type;
906 return (m);
984263bc
MD
907}
908
909/*
910 * This routine is called to generate a message from the routing
f23061d4 911 * socket indicating that a redirect has occurred, a routing lookup
984263bc
MD
912 * has failed, or that a protocol has detected timeouts to a particular
913 * destination.
914 */
915void
f23061d4 916rt_missmsg(int type, struct rt_addrinfo *rtinfo, int flags, int error)
984263bc 917{
0c3c561c 918 struct sockaddr *dst = rtinfo->rti_info[RTAX_DST];
82ed7fc2
RG
919 struct rt_msghdr *rtm;
920 struct mbuf *m;
984263bc
MD
921
922 if (route_cb.any_count == 0)
923 return;
590b8cd4 924 m = rt_msg_mbuf(type, rtinfo);
ef87f48d 925 if (m == NULL)
984263bc
MD
926 return;
927 rtm = mtod(m, struct rt_msghdr *);
928 rtm->rtm_flags = RTF_DONE | flags;
929 rtm->rtm_errno = error;
930 rtm->rtm_addrs = rtinfo->rti_addrs;
590b8cd4 931 rts_input(m, familyof(dst));
984263bc
MD
932}
933
f3ed2586
JH
934void
935rt_dstmsg(int type, struct sockaddr *dst, int error)
936{
937 struct rt_msghdr *rtm;
938 struct rt_addrinfo addrs;
939 struct mbuf *m;
940
941 if (route_cb.any_count == 0)
942 return;
943 bzero(&addrs, sizeof(struct rt_addrinfo));
944 addrs.rti_info[RTAX_DST] = dst;
590b8cd4 945 m = rt_msg_mbuf(type, &addrs);
f3ed2586
JH
946 if (m == NULL)
947 return;
948 rtm = mtod(m, struct rt_msghdr *);
949 rtm->rtm_flags = RTF_DONE;
950 rtm->rtm_errno = error;
951 rtm->rtm_addrs = addrs.rti_addrs;
590b8cd4 952 rts_input(m, familyof(dst));
f3ed2586
JH
953}
954
984263bc
MD
955/*
956 * This routine is called to generate a message from the routing
957 * socket indicating that the status of a network interface has changed.
958 */
959void
0c3c561c 960rt_ifmsg(struct ifnet *ifp)
984263bc 961{
82ed7fc2 962 struct if_msghdr *ifm;
984263bc 963 struct mbuf *m;
590b8cd4 964 struct rt_addrinfo rtinfo;
984263bc
MD
965
966 if (route_cb.any_count == 0)
967 return;
590b8cd4
JH
968 bzero(&rtinfo, sizeof(struct rt_addrinfo));
969 m = rt_msg_mbuf(RTM_IFINFO, &rtinfo);
ef87f48d 970 if (m == NULL)
984263bc
MD
971 return;
972 ifm = mtod(m, struct if_msghdr *);
973 ifm->ifm_index = ifp->if_index;
9c095379 974 ifm->ifm_flags = ifp->if_flags;
984263bc 975 ifm->ifm_data = ifp->if_data;
590b8cd4
JH
976 ifm->ifm_addrs = 0;
977 rts_input(m, 0);
984263bc
MD
978}
979
372316d9
JH
980static void
981rt_ifamsg(int cmd, struct ifaddr *ifa)
982{
983 struct ifa_msghdr *ifam;
590b8cd4 984 struct rt_addrinfo rtinfo;
372316d9 985 struct mbuf *m;
372316d9
JH
986 struct ifnet *ifp = ifa->ifa_ifp;
987
590b8cd4
JH
988 bzero(&rtinfo, sizeof(struct rt_addrinfo));
989 rtinfo.rti_ifaaddr = ifa->ifa_addr;
b2632176
SZ
990 rtinfo.rti_ifpaddr =
991 TAILQ_FIRST(&ifp->if_addrheads[mycpuid])->ifa->ifa_addr;
590b8cd4
JH
992 rtinfo.rti_netmask = ifa->ifa_netmask;
993 rtinfo.rti_bcastaddr = ifa->ifa_dstaddr;
f23061d4 994
590b8cd4 995 m = rt_msg_mbuf(cmd, &rtinfo);
f23061d4 996 if (m == NULL)
372316d9 997 return;
f23061d4 998
372316d9
JH
999 ifam = mtod(m, struct ifa_msghdr *);
1000 ifam->ifam_index = ifp->if_index;
1001 ifam->ifam_metric = ifa->ifa_metric;
1002 ifam->ifam_flags = ifa->ifa_flags;
590b8cd4 1003 ifam->ifam_addrs = rtinfo.rti_addrs;
f23061d4 1004
590b8cd4 1005 rts_input(m, familyof(ifa->ifa_addr));
372316d9
JH
1006}
1007
f3ed2586
JH
1008void
1009rt_rtmsg(int cmd, struct rtentry *rt, struct ifnet *ifp, int error)
372316d9
JH
1010{
1011 struct rt_msghdr *rtm;
590b8cd4 1012 struct rt_addrinfo rtinfo;
372316d9 1013 struct mbuf *m;
f3ed2586 1014 struct sockaddr *dst;
372316d9
JH
1015
1016 if (rt == NULL)
1017 return;
f23061d4 1018
590b8cd4
JH
1019 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1020 rtinfo.rti_dst = dst = rt_key(rt);
1021 rtinfo.rti_gateway = rt->rt_gateway;
1022 rtinfo.rti_netmask = rt_mask(rt);
b2632176
SZ
1023 if (ifp != NULL) {
1024 rtinfo.rti_ifpaddr =
1025 TAILQ_FIRST(&ifp->if_addrheads[mycpuid])->ifa->ifa_addr;
1026 }
590b8cd4 1027 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr;
f23061d4 1028
590b8cd4 1029 m = rt_msg_mbuf(cmd, &rtinfo);
f23061d4 1030 if (m == NULL)
372316d9 1031 return;
f23061d4 1032
372316d9 1033 rtm = mtod(m, struct rt_msghdr *);
f3ed2586
JH
1034 if (ifp != NULL)
1035 rtm->rtm_index = ifp->if_index;
372316d9
JH
1036 rtm->rtm_flags |= rt->rt_flags;
1037 rtm->rtm_errno = error;
590b8cd4 1038 rtm->rtm_addrs = rtinfo.rti_addrs;
f23061d4 1039
590b8cd4 1040 rts_input(m, familyof(dst));
372316d9
JH
1041}
1042
984263bc
MD
1043/*
1044 * This is called to generate messages from the routing socket
1045 * indicating a network interface has had addresses associated with it.
1046 * if we ever reverse the logic and replace messages TO the routing
1047 * socket indicate a request to configure interfaces, then it will
1048 * be unnecessary as the routing socket will automatically generate
1049 * copies of it.
1050 */
1051void
0c3c561c 1052rt_newaddrmsg(int cmd, struct ifaddr *ifa, int error, struct rtentry *rt)
984263bc 1053{
78812139
EN
1054#ifdef SCTP
1055 /*
1056 * notify the SCTP stack
1057 * this will only get called when an address is added/deleted
1058 * XXX pass the ifaddr struct instead if ifa->ifa_addr...
1059 */
1060 if (cmd == RTM_ADD)
1061 sctp_add_ip_address(ifa);
1062 else if (cmd == RTM_DELETE)
1063 sctp_delete_ip_address(ifa);
1064#endif /* SCTP */
1065
984263bc
MD
1066 if (route_cb.any_count == 0)
1067 return;
984263bc 1068
372316d9
JH
1069 if (cmd == RTM_ADD) {
1070 rt_ifamsg(RTM_NEWADDR, ifa);
f3ed2586 1071 rt_rtmsg(RTM_ADD, rt, ifa->ifa_ifp, error);
372316d9
JH
1072 } else {
1073 KASSERT((cmd == RTM_DELETE), ("unknown cmd %d", cmd));
f3ed2586 1074 rt_rtmsg(RTM_DELETE, rt, ifa->ifa_ifp, error);
372316d9 1075 rt_ifamsg(RTM_DELADDR, ifa);
984263bc
MD
1076 }
1077}
1078
1079/*
1080 * This is the analogue to the rt_newaddrmsg which performs the same
1081 * function but for multicast group memberhips. This is easier since
1082 * there is no route state to worry about.
1083 */
1084void
0c3c561c 1085rt_newmaddrmsg(int cmd, struct ifmultiaddr *ifma)
984263bc 1086{
590b8cd4 1087 struct rt_addrinfo rtinfo;
ef87f48d 1088 struct mbuf *m = NULL;
984263bc
MD
1089 struct ifnet *ifp = ifma->ifma_ifp;
1090 struct ifma_msghdr *ifmam;
1091
1092 if (route_cb.any_count == 0)
1093 return;
1094
590b8cd4
JH
1095 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1096 rtinfo.rti_ifaaddr = ifma->ifma_addr;
b2632176
SZ
1097 if (ifp != NULL && !TAILQ_EMPTY(&ifp->if_addrheads[mycpuid])) {
1098 rtinfo.rti_ifpaddr =
1099 TAILQ_FIRST(&ifp->if_addrheads[mycpuid])->ifa->ifa_addr;
1100 }
984263bc
MD
1101 /*
1102 * If a link-layer address is present, present it as a ``gateway''
1103 * (similarly to how ARP entries, e.g., are presented).
1104 */
590b8cd4 1105 rtinfo.rti_gateway = ifma->ifma_lladdr;
f23061d4 1106
590b8cd4 1107 m = rt_msg_mbuf(cmd, &rtinfo);
f23061d4 1108 if (m == NULL)
984263bc 1109 return;
f23061d4 1110
984263bc
MD
1111 ifmam = mtod(m, struct ifma_msghdr *);
1112 ifmam->ifmam_index = ifp->if_index;
590b8cd4 1113 ifmam->ifmam_addrs = rtinfo.rti_addrs;
f23061d4 1114
590b8cd4 1115 rts_input(m, familyof(ifma->ifma_addr));
984263bc
MD
1116}
1117
841ab66c
SZ
1118static struct mbuf *
1119rt_makeifannouncemsg(struct ifnet *ifp, int type, int what,
1120 struct rt_addrinfo *info)
984263bc 1121{
590b8cd4 1122 struct if_announcemsghdr *ifan;
841ab66c 1123 struct mbuf *m;
984263bc
MD
1124
1125 if (route_cb.any_count == 0)
841ab66c 1126 return NULL;
f23061d4 1127
841ab66c
SZ
1128 bzero(info, sizeof(*info));
1129 m = rt_msg_mbuf(type, info);
984263bc 1130 if (m == NULL)
841ab66c 1131 return NULL;
f23061d4 1132
984263bc
MD
1133 ifan = mtod(m, struct if_announcemsghdr *);
1134 ifan->ifan_index = ifp->if_index;
5fe66e68 1135 strlcpy(ifan->ifan_name, ifp->if_xname, sizeof ifan->ifan_name);
984263bc 1136 ifan->ifan_what = what;
841ab66c
SZ
1137 return m;
1138}
1139
1140/*
1141 * This is called to generate routing socket messages indicating
1142 * IEEE80211 wireless events.
1143 * XXX we piggyback on the RTM_IFANNOUNCE msg format in a clumsy way.
1144 */
1145void
1146rt_ieee80211msg(struct ifnet *ifp, int what, void *data, size_t data_len)
1147{
1148 struct rt_addrinfo info;
1149 struct mbuf *m;
f23061d4 1150
841ab66c
SZ
1151 m = rt_makeifannouncemsg(ifp, RTM_IEEE80211, what, &info);
1152 if (m == NULL)
1153 return;
1154
1155 /*
1156 * Append the ieee80211 data. Try to stick it in the
1157 * mbuf containing the ifannounce msg; otherwise allocate
1158 * a new mbuf and append.
1159 *
1160 * NB: we assume m is a single mbuf.
1161 */
1162 if (data_len > M_TRAILINGSPACE(m)) {
b11e1686 1163 struct mbuf *n = m_get(MB_DONTWAIT, MT_DATA);
841ab66c
SZ
1164 if (n == NULL) {
1165 m_freem(m);
1166 return;
1167 }
1168 bcopy(data, mtod(n, void *), data_len);
1169 n->m_len = data_len;
1170 m->m_next = n;
1171 } else if (data_len > 0) {
1172 bcopy(data, mtod(m, u_int8_t *) + m->m_len, data_len);
1173 m->m_len += data_len;
1174 }
e9fa4b60 1175 mbuftrackid(m, 33);
841ab66c
SZ
1176 if (m->m_flags & M_PKTHDR)
1177 m->m_pkthdr.len += data_len;
1178 mtod(m, struct if_announcemsghdr *)->ifan_msglen += data_len;
590b8cd4 1179 rts_input(m, 0);
841ab66c
SZ
1180}
1181
1182/*
1183 * This is called to generate routing socket messages indicating
1184 * network interface arrival and departure.
1185 */
1186void
1187rt_ifannouncemsg(struct ifnet *ifp, int what)
1188{
1189 struct rt_addrinfo addrinfo;
1190 struct mbuf *m;
1191
1192 m = rt_makeifannouncemsg(ifp, RTM_IFANNOUNCE, what, &addrinfo);
1193 if (m != NULL)
1194 rts_input(m, 0);
590b8cd4 1195}
f23061d4 1196
590b8cd4
JH
1197static int
1198resizewalkarg(struct walkarg *w, int len)
1199{
1200 void *newptr;
1201
efda3bd0 1202 newptr = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
590b8cd4
JH
1203 if (newptr == NULL)
1204 return (ENOMEM);
1205 if (w->w_tmem != NULL)
efda3bd0 1206 kfree(w->w_tmem, M_RTABLE);
590b8cd4
JH
1207 w->w_tmem = newptr;
1208 w->w_tmemsize = len;
1209 return (0);
1210}
984263bc
MD
1211
1212/*
1213 * This is used in dumping the kernel table via sysctl().
1214 */
1215int
590b8cd4 1216sysctl_dumpentry(struct radix_node *rn, void *vw)
984263bc 1217{
82ed7fc2
RG
1218 struct walkarg *w = vw;
1219 struct rtentry *rt = (struct rtentry *)rn;
590b8cd4
JH
1220 struct rt_addrinfo rtinfo;
1221 int error, msglen;
984263bc
MD
1222
1223 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg))
1224 return 0;
f23061d4 1225
590b8cd4
JH
1226 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1227 rtinfo.rti_dst = rt_key(rt);
1228 rtinfo.rti_gateway = rt->rt_gateway;
1229 rtinfo.rti_netmask = rt_mask(rt);
1230 rtinfo.rti_genmask = rt->rt_genmask;
ef87f48d 1231 if (rt->rt_ifp != NULL) {
590b8cd4 1232 rtinfo.rti_ifpaddr =
b2632176 1233 TAILQ_FIRST(&rt->rt_ifp->if_addrheads[mycpuid])->ifa->ifa_addr;
590b8cd4 1234 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr;
984263bc 1235 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT)
590b8cd4 1236 rtinfo.rti_bcastaddr = rt->rt_ifa->ifa_dstaddr;
984263bc 1237 }
590b8cd4
JH
1238 msglen = rt_msgsize(RTM_GET, &rtinfo);
1239 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0)
1240 return (ENOMEM);
1241 rt_msg_buffer(RTM_GET, &rtinfo, w->w_tmem, msglen);
1242 if (w->w_req != NULL) {
8e63efec 1243 struct rt_msghdr *rtm = w->w_tmem;
984263bc
MD
1244
1245 rtm->rtm_flags = rt->rt_flags;
1246 rtm->rtm_use = rt->rt_use;
1247 rtm->rtm_rmx = rt->rt_rmx;
1248 rtm->rtm_index = rt->rt_ifp->if_index;
1249 rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0;
590b8cd4
JH
1250 rtm->rtm_addrs = rtinfo.rti_addrs;
1251 error = SYSCTL_OUT(w->w_req, rtm, msglen);
984263bc
MD
1252 return (error);
1253 }
590b8cd4 1254 return (0);
984263bc
MD
1255}
1256
590b8cd4
JH
1257static int
1258sysctl_iflist(int af, struct walkarg *w)
984263bc 1259{
82ed7fc2 1260 struct ifnet *ifp;
590b8cd4
JH
1261 struct rt_addrinfo rtinfo;
1262 int msglen, error;
984263bc 1263
590b8cd4 1264 bzero(&rtinfo, sizeof(struct rt_addrinfo));
984263bc 1265 TAILQ_FOREACH(ifp, &ifnet, if_link) {
b2632176
SZ
1266 struct ifaddr_container *ifac;
1267 struct ifaddr *ifa;
1268
984263bc
MD
1269 if (w->w_arg && w->w_arg != ifp->if_index)
1270 continue;
b2632176
SZ
1271 ifac = TAILQ_FIRST(&ifp->if_addrheads[mycpuid]);
1272 ifa = ifac->ifa;
590b8cd4
JH
1273 rtinfo.rti_ifpaddr = ifa->ifa_addr;
1274 msglen = rt_msgsize(RTM_IFINFO, &rtinfo);
1275 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0)
1276 return (ENOMEM);
1277 rt_msg_buffer(RTM_IFINFO, &rtinfo, w->w_tmem, msglen);
1278 rtinfo.rti_ifpaddr = NULL;
ef87f48d 1279 if (w->w_req != NULL && w->w_tmem != NULL) {
8e63efec 1280 struct if_msghdr *ifm = w->w_tmem;
984263bc 1281
984263bc 1282 ifm->ifm_index = ifp->if_index;
9c095379 1283 ifm->ifm_flags = ifp->if_flags;
984263bc 1284 ifm->ifm_data = ifp->if_data;
590b8cd4
JH
1285 ifm->ifm_addrs = rtinfo.rti_addrs;
1286 error = SYSCTL_OUT(w->w_req, ifm, msglen);
984263bc
MD
1287 if (error)
1288 return (error);
1289 }
b2632176
SZ
1290 while ((ifac = TAILQ_NEXT(ifac, ifa_link)) != NULL) {
1291 ifa = ifac->ifa;
1292
984263bc
MD
1293 if (af && af != ifa->ifa_addr->sa_family)
1294 continue;
590b8cd4 1295 if (curproc->p_ucred->cr_prison &&
87de5057 1296 prison_if(curproc->p_ucred, ifa->ifa_addr))
984263bc 1297 continue;
590b8cd4
JH
1298 rtinfo.rti_ifaaddr = ifa->ifa_addr;
1299 rtinfo.rti_netmask = ifa->ifa_netmask;
1300 rtinfo.rti_bcastaddr = ifa->ifa_dstaddr;
1301 msglen = rt_msgsize(RTM_NEWADDR, &rtinfo);
1302 if (w->w_tmemsize < msglen &&
1303 resizewalkarg(w, msglen) != 0)
1304 return (ENOMEM);
1305 rt_msg_buffer(RTM_NEWADDR, &rtinfo, w->w_tmem, msglen);
1306 if (w->w_req != NULL) {
8e63efec 1307 struct ifa_msghdr *ifam = w->w_tmem;
984263bc 1308
984263bc
MD
1309 ifam->ifam_index = ifa->ifa_ifp->if_index;
1310 ifam->ifam_flags = ifa->ifa_flags;
1311 ifam->ifam_metric = ifa->ifa_metric;
590b8cd4
JH
1312 ifam->ifam_addrs = rtinfo.rti_addrs;
1313 error = SYSCTL_OUT(w->w_req, w->w_tmem, msglen);
984263bc
MD
1314 if (error)
1315 return (error);
1316 }
1317 }
590b8cd4
JH
1318 rtinfo.rti_netmask = NULL;
1319 rtinfo.rti_ifaaddr = NULL;
1320 rtinfo.rti_bcastaddr = NULL;
984263bc
MD
1321 }
1322 return (0);
1323}
1324
1325static int
1326sysctl_rtsock(SYSCTL_HANDLER_ARGS)
1327{
1328 int *name = (int *)arg1;
1329 u_int namelen = arg2;
82ed7fc2 1330 struct radix_node_head *rnh;
a49aa710 1331 int i, error = EINVAL;
ecdefdda 1332 int origcpu;
984263bc
MD
1333 u_char af;
1334 struct walkarg w;
1335
1336 name ++;
1337 namelen--;
1338 if (req->newptr)
1339 return (EPERM);
ecdefdda 1340 if (namelen != 3 && namelen != 4)
984263bc
MD
1341 return (EINVAL);
1342 af = name[0];
5fe66e68 1343 bzero(&w, sizeof w);
984263bc
MD
1344 w.w_op = name[1];
1345 w.w_arg = name[2];
1346 w.w_req = req;
1347
ecdefdda
MD
1348 /*
1349 * Optional third argument specifies cpu, used primarily for
1350 * debugging the route table.
1351 */
1352 if (namelen == 4) {
1353 if (name[3] < 0 || name[3] >= ncpus)
1354 return (EINVAL);
1355 origcpu = mycpuid;
1356 lwkt_migratecpu(name[3]);
1357 } else {
1358 origcpu = -1;
1359 }
a49aa710 1360 crit_enter();
984263bc 1361 switch (w.w_op) {
984263bc
MD
1362 case NET_RT_DUMP:
1363 case NET_RT_FLAGS:
1364 for (i = 1; i <= AF_MAX; i++)
ecdefdda
MD
1365 if ((rnh = rt_tables[mycpuid][i]) &&
1366 (af == 0 || af == i) &&
984263bc 1367 (error = rnh->rnh_walktree(rnh,
f23061d4 1368 sysctl_dumpentry, &w)))
984263bc
MD
1369 break;
1370 break;
1371
1372 case NET_RT_IFLIST:
1373 error = sysctl_iflist(af, &w);
1374 }
a49aa710 1375 crit_exit();
590b8cd4 1376 if (w.w_tmem != NULL)
efda3bd0 1377 kfree(w.w_tmem, M_RTABLE);
ecdefdda
MD
1378 if (origcpu >= 0)
1379 lwkt_migratecpu(origcpu);
984263bc
MD
1380 return (error);
1381}
1382
1383SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, "");
1384
1385/*
1386 * Definitions of protocols supported in the ROUTE domain.
1387 */
1388
ba4257b8 1389static struct domain routedomain; /* or at least forward */
984263bc
MD
1390
1391static struct protosw routesw[] = {
1392{ SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR,
1393 0, route_output, raw_ctlinput, 0,
e3873585 1394 cpu0_soport, cpu0_ctlport,
984263bc
MD
1395 raw_init, 0, 0, 0,
1396 &route_usrreqs
1397}
1398};
1399
ba4257b8 1400static struct domain routedomain = {
9c70fe43
JH
1401 PF_ROUTE, "route", NULL, NULL, NULL,
1402 routesw, &routesw[(sizeof routesw)/(sizeof routesw[0])],
1403};
984263bc
MD
1404
1405DOMAIN_SET(route);
47db7f9b 1406