Add missing name.
[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 $
fa5e758c 67 * $DragonFly: src/sys/net/rtsock.c,v 1.38 2007/04/21 02:26:47 dillon 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>
77#include <sys/malloc.h>
78#include <sys/mbuf.h>
47e78ae8 79#include <sys/protosw.h>
984263bc
MD
80#include <sys/socket.h>
81#include <sys/socketvar.h>
82#include <sys/domain.h>
a49aa710 83#include <sys/thread2.h>
deffea2e 84
984263bc
MD
85#include <net/if.h>
86#include <net/route.h>
87#include <net/raw_cb.h>
88
78812139
EN
89#ifdef SCTP
90extern void sctp_add_ip_address(struct ifaddr *ifa);
91extern void sctp_delete_ip_address(struct ifaddr *ifa);
92#endif /* SCTP */
93
984263bc
MD
94MALLOC_DEFINE(M_RTABLE, "routetbl", "routing tables");
95
f23061d4
JH
96static struct route_cb {
97 int ip_count;
98 int ip6_count;
99 int ipx_count;
100 int ns_count;
101 int any_count;
102} route_cb;
103
590b8cd4 104static const struct sockaddr route_src = { 2, PF_ROUTE, };
984263bc
MD
105
106struct walkarg {
107 int w_tmemsize;
108 int w_op, w_arg;
8e63efec 109 void *w_tmem;
984263bc
MD
110 struct sysctl_req *w_req;
111};
112
113static struct mbuf *
590b8cd4
JH
114 rt_msg_mbuf (int, struct rt_addrinfo *);
115static void rt_msg_buffer (int, struct rt_addrinfo *, void *buf, int len);
116static int rt_msgsize (int type, struct rt_addrinfo *rtinfo);
ef87f48d 117static int rt_xaddrs (char *, char *, struct rt_addrinfo *);
158abb01
RG
118static int sysctl_dumpentry (struct radix_node *rn, void *vw);
119static int sysctl_iflist (int af, struct walkarg *w);
deffea2e 120static int route_output(struct mbuf *, struct socket *, ...);
47e78ae8 121static void rt_setmetrics (u_long, struct rt_metrics *,
f23061d4 122 struct rt_metrics *);
984263bc 123
984263bc
MD
124/*
125 * It really doesn't make any sense at all for this code to share much
126 * with raw_usrreq.c, since its functionality is so restricted. XXX
127 */
128static int
129rts_abort(struct socket *so)
130{
a49aa710 131 int error;
f23061d4 132
a49aa710 133 crit_enter();
984263bc 134 error = raw_usrreqs.pru_abort(so);
a49aa710 135 crit_exit();
984263bc
MD
136 return error;
137}
138
139/* pru_accept is EOPNOTSUPP */
140
141static int
e4700d00 142rts_attach(struct socket *so, int proto, struct pru_attach_info *ai)
984263bc
MD
143{
144 struct rawcb *rp;
a49aa710 145 int error;
984263bc 146
ef87f48d 147 if (sotorawcb(so) != NULL)
984263bc 148 return EISCONN; /* XXX panic? */
f23061d4 149
efda3bd0 150 rp = kmalloc(sizeof *rp, M_PCB, M_WAITOK | M_ZERO);
ef87f48d 151 if (rp == NULL)
984263bc
MD
152 return ENOBUFS;
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
338static void
339rts_input(struct mbuf *m, sa_family_t family)
340{
341 static const struct sockaddr route_dst = { 2, PF_ROUTE, };
342 struct sockproto route_proto = { PF_ROUTE, family };
343
344 raw_input(m, &route_proto, &route_src, &route_dst);
345}
346
347static void *
348reallocbuf(void *ptr, size_t len, size_t olen)
349{
350 void *newptr;
351
efda3bd0 352 newptr = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
590b8cd4
JH
353 if (newptr == NULL)
354 return NULL;
355 bcopy(ptr, newptr, olen);
efda3bd0 356 kfree(ptr, M_RTABLE);
590b8cd4
JH
357 return (newptr);
358}
359
8e63efec
JH
360/*
361 * Internal helper routine for route_output().
362 */
590b8cd4
JH
363static int
364fillrtmsg(struct rt_msghdr **prtm, struct rtentry *rt,
365 struct rt_addrinfo *rtinfo)
366{
367 int msglen;
368 struct rt_msghdr *rtm = *prtm;
369
370 /* Fill in rt_addrinfo for call to rt_msg_buffer(). */
371 rtinfo->rti_dst = rt_key(rt);
372 rtinfo->rti_gateway = rt->rt_gateway;
373 rtinfo->rti_netmask = rt_mask(rt); /* might be NULL */
374 rtinfo->rti_genmask = rt->rt_genmask; /* might be NULL */
375 if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) {
376 if (rt->rt_ifp != NULL) {
377 rtinfo->rti_ifpaddr =
378 TAILQ_FIRST(&rt->rt_ifp->if_addrhead)->ifa_addr;
379 rtinfo->rti_ifaaddr = rt->rt_ifa->ifa_addr;
380 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT)
381 rtinfo->rti_bcastaddr = rt->rt_ifa->ifa_dstaddr;
382 rtm->rtm_index = rt->rt_ifp->if_index;
383 } else {
384 rtinfo->rti_ifpaddr = NULL;
385 rtinfo->rti_ifaaddr = NULL;
386 }
387 }
388
389 msglen = rt_msgsize(rtm->rtm_type, rtinfo);
390 if (rtm->rtm_msglen < msglen) {
391 rtm = reallocbuf(rtm, msglen, rtm->rtm_msglen);
392 if (rtm == NULL)
393 return (ENOBUFS);
394 *prtm = rtm;
395 }
396 rt_msg_buffer(rtm->rtm_type, rtinfo, rtm, msglen);
397
398 rtm->rtm_flags = rt->rt_flags;
399 rtm->rtm_rmx = rt->rt_rmx;
400 rtm->rtm_addrs = rtinfo->rti_addrs;
401
402 return (0);
403}
404
ecdefdda
MD
405static void route_output_add_callback(int, int, struct rt_addrinfo *,
406 struct rtentry *, void *);
407static void route_output_delete_callback(int, int, struct rt_addrinfo *,
408 struct rtentry *, void *);
409static void route_output_change_callback(int, int, struct rt_addrinfo *,
410 struct rtentry *, void *);
411static void route_output_lock_callback(int, int, struct rt_addrinfo *,
412 struct rtentry *, void *);
413
984263bc
MD
414/*ARGSUSED*/
415static int
deffea2e 416route_output(struct mbuf *m, struct socket *so, ...)
984263bc 417{
2e9572df 418 struct rt_msghdr *rtm = NULL;
ecdefdda 419 struct rtentry *rt;
984263bc 420 struct radix_node_head *rnh;
f23061d4 421 struct rawcb *rp = NULL;
deffea2e 422 struct pr_output_info *oi;
590b8cd4 423 struct rt_addrinfo rtinfo;
ef87f48d 424 int len, error = 0;
deffea2e
JS
425 __va_list ap;
426
427 __va_start(ap, so);
428 oi = __va_arg(ap, struct pr_output_info *);
429 __va_end(ap);
984263bc 430
ef87f48d 431#define gotoerr(e) { error = e; goto flush;}
590b8cd4
JH
432
433 if (m == NULL ||
434 (m->m_len < sizeof(long) &&
435 (m = m_pullup(m, sizeof(long))) == NULL))
984263bc 436 return (ENOBUFS);
ef87f48d 437 if (!(m->m_flags & M_PKTHDR))
984263bc
MD
438 panic("route_output");
439 len = m->m_pkthdr.len;
590b8cd4 440 if (len < sizeof(struct rt_msghdr) ||
984263bc 441 len != mtod(m, struct rt_msghdr *)->rtm_msglen) {
590b8cd4 442 rtinfo.rti_dst = NULL;
ef87f48d 443 gotoerr(EINVAL);
984263bc 444 }
efda3bd0 445 rtm = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
ef87f48d 446 if (rtm == NULL) {
590b8cd4 447 rtinfo.rti_dst = NULL;
ef87f48d 448 gotoerr(ENOBUFS);
984263bc
MD
449 }
450 m_copydata(m, 0, len, (caddr_t)rtm);
451 if (rtm->rtm_version != RTM_VERSION) {
590b8cd4 452 rtinfo.rti_dst = NULL;
ef87f48d 453 gotoerr(EPROTONOSUPPORT);
984263bc 454 }
47e78ae8 455 rtm->rtm_pid = oi->p_pid;
590b8cd4
JH
456 bzero(&rtinfo, sizeof(struct rt_addrinfo));
457 rtinfo.rti_addrs = rtm->rtm_addrs;
458 if (rt_xaddrs((char *)(rtm + 1), (char *)rtm + len, &rtinfo) != 0) {
459 rtinfo.rti_dst = NULL;
ef87f48d 460 gotoerr(EINVAL);
984263bc 461 }
590b8cd4
JH
462 rtinfo.rti_flags = rtm->rtm_flags;
463 if (rtinfo.rti_dst == NULL || rtinfo.rti_dst->sa_family >= AF_MAX ||
464 (rtinfo.rti_gateway && rtinfo.rti_gateway->sa_family >= AF_MAX))
ef87f48d
JH
465 gotoerr(EINVAL);
466
590b8cd4
JH
467 if (rtinfo.rti_genmask != NULL) {
468 struct radix_node *n;
ef87f48d 469
590b8cd4
JH
470#define clen(s) (*(u_char *)(s))
471 n = rn_addmask((char *)rtinfo.rti_genmask, TRUE, 1);
472 if (n != NULL &&
473 rtinfo.rti_genmask->sa_len >= clen(n->rn_key) &&
474 bcmp((char *)rtinfo.rti_genmask + 1,
475 (char *)n->rn_key + 1, clen(n->rn_key) - 1) == 0)
476 rtinfo.rti_genmask = (struct sockaddr *)n->rn_key;
984263bc 477 else
ef87f48d 478 gotoerr(ENOBUFS);
984263bc
MD
479 }
480
481 /*
482 * Verify that the caller has the appropriate privilege; RTM_GET
483 * is the only operation the non-superuser is allowed.
484 */
dadab5e9 485 if (rtm->rtm_type != RTM_GET && suser_cred(so->so_cred, 0) != 0)
ef87f48d 486 gotoerr(EPERM);
984263bc
MD
487
488 switch (rtm->rtm_type) {
984263bc 489 case RTM_ADD:
ecdefdda
MD
490 if (rtinfo.rti_gateway == NULL) {
491 error = EINVAL;
492 } else {
493 error = rtrequest1_global(RTM_ADD, &rtinfo,
494 route_output_add_callback, rtm);
984263bc
MD
495 }
496 break;
984263bc 497 case RTM_DELETE:
ecdefdda
MD
498 /*
499 * note: &rtm passed as argument so 'rtm' can be replaced.
500 */
501 error = rtrequest1_global(RTM_DELETE, &rtinfo,
502 route_output_delete_callback, &rtm);
984263bc 503 break;
984263bc 504 case RTM_GET:
ecdefdda
MD
505 rnh = rt_tables[mycpuid][rtinfo.rti_dst->sa_family];
506 if (rnh == NULL) {
507 error = EAFNOSUPPORT;
508 break;
509 }
590b8cd4
JH
510 rt = (struct rtentry *)
511 rnh->rnh_lookup((char *)rtinfo.rti_dst,
512 (char *)rtinfo.rti_netmask, rnh);
ecdefdda
MD
513 if (rt == NULL) {
514 error = ESRCH;
984263bc
MD
515 break;
516 }
ecdefdda
MD
517 rt->rt_refcnt++;
518 if (fillrtmsg(&rtm, rt, &rtinfo) != 0)
519 gotoerr(ENOBUFS);
520 --rt->rt_refcnt;
521 break;
522 case RTM_CHANGE:
523 error = rtrequest1_global(RTM_GET, &rtinfo,
524 route_output_change_callback, rtm);
525 break;
526 case RTM_LOCK:
527 error = rtrequest1_global(RTM_GET, &rtinfo,
528 route_output_lock_callback, rtm);
590b8cd4 529 break;
984263bc 530 default:
ecdefdda
MD
531 error = EOPNOTSUPP;
532 break;
984263bc
MD
533 }
534
535flush:
590b8cd4
JH
536 if (rtm != NULL) {
537 if (error != 0)
984263bc
MD
538 rtm->rtm_errno = error;
539 else
540 rtm->rtm_flags |= RTF_DONE;
541 }
ecdefdda 542
984263bc
MD
543 /*
544 * Check to see if we don't want our own messages.
545 */
ef87f48d 546 if (!(so->so_options & SO_USELOOPBACK)) {
984263bc 547 if (route_cb.any_count <= 1) {
590b8cd4 548 if (rtm != NULL)
efda3bd0 549 kfree(rtm, M_RTABLE);
984263bc
MD
550 m_freem(m);
551 return (error);
552 }
553 /* There is another listener, so construct message */
554 rp = sotorawcb(so);
555 }
590b8cd4 556 if (rtm != NULL) {
984263bc
MD
557 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm);
558 if (m->m_pkthdr.len < rtm->rtm_msglen) {
559 m_freem(m);
560 m = NULL;
561 } else if (m->m_pkthdr.len > rtm->rtm_msglen)
562 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len);
efda3bd0 563 kfree(rtm, M_RTABLE);
984263bc 564 }
ef87f48d 565 if (rp != NULL)
984263bc 566 rp->rcb_proto.sp_family = 0; /* Avoid us */
ef87f48d 567 if (m != NULL)
590b8cd4 568 rts_input(m, familyof(rtinfo.rti_dst));
ef87f48d 569 if (rp != NULL)
984263bc 570 rp->rcb_proto.sp_family = PF_ROUTE;
984263bc
MD
571 return (error);
572}
573
ecdefdda
MD
574static void
575route_output_add_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
576 struct rtentry *rt, void *arg)
577{
578 struct rt_msghdr *rtm = arg;
579
580 if (error == 0 && rt != NULL) {
581 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx,
582 &rt->rt_rmx);
583 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits);
584 rt->rt_rmx.rmx_locks |=
585 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks);
586 rt->rt_genmask = rtinfo->rti_genmask;
587 }
588}
589
590static void
591route_output_delete_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
592 struct rtentry *rt, void *arg)
593{
594 struct rt_msghdr **rtm = arg;
595
596 if (error == 0 && rt) {
597 ++rt->rt_refcnt;
598 if (fillrtmsg(rtm, rt, rtinfo) != 0) {
599 error = ENOBUFS;
600 /* XXX no way to return the error */
601 }
602 --rt->rt_refcnt;
603 }
604}
605
606static void
607route_output_change_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
608 struct rtentry *rt, void *arg)
609{
610 struct rt_msghdr *rtm = arg;
611 struct ifaddr *ifa;
612
613 if (error)
614 goto done;
615
616 /*
617 * new gateway could require new ifaddr, ifp;
618 * flags may also be different; ifp may be specified
619 * by ll sockaddr when protocol address is ambiguous
620 */
621 if (((rt->rt_flags & RTF_GATEWAY) && rtinfo->rti_gateway != NULL) ||
622 rtinfo->rti_ifpaddr != NULL || (rtinfo->rti_ifaaddr != NULL &&
623 sa_equal(rtinfo->rti_ifaaddr, rt->rt_ifa->ifa_addr))
624 ) {
625 error = rt_getifa(rtinfo);
626 if (error != 0)
627 goto done;
628 }
629 if (rtinfo->rti_gateway != NULL) {
630 error = rt_setgate(rt, rt_key(rt), rtinfo->rti_gateway);
631 if (error != 0)
632 goto done;
633 }
634 if ((ifa = rtinfo->rti_ifa) != NULL) {
635 struct ifaddr *oifa = rt->rt_ifa;
636
637 if (oifa != ifa) {
638 if (oifa && oifa->ifa_rtrequest)
639 oifa->ifa_rtrequest(RTM_DELETE, rt, rtinfo);
640 IFAFREE(rt->rt_ifa);
641 IFAREF(ifa);
642 rt->rt_ifa = ifa;
643 rt->rt_ifp = rtinfo->rti_ifp;
644 }
645 }
646 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, &rt->rt_rmx);
647 if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest)
648 rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, rtinfo);
649 if (rtinfo->rti_genmask != NULL)
650 rt->rt_genmask = rtinfo->rti_genmask;
651done:
652 /* XXX no way to return error */
653 ;
654}
655
656static void
657route_output_lock_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
658 struct rtentry *rt, void *arg)
659{
660 struct rt_msghdr *rtm = arg;
661
662 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits);
663 rt->rt_rmx.rmx_locks |=
664 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks);
665}
666
984263bc 667static void
ef87f48d 668rt_setmetrics(u_long which, struct rt_metrics *in, struct rt_metrics *out)
984263bc 669{
2e9572df
JH
670#define setmetric(flag, elt) if (which & (flag)) out->elt = in->elt;
671 setmetric(RTV_RPIPE, rmx_recvpipe);
672 setmetric(RTV_SPIPE, rmx_sendpipe);
673 setmetric(RTV_SSTHRESH, rmx_ssthresh);
674 setmetric(RTV_RTT, rmx_rtt);
675 setmetric(RTV_RTTVAR, rmx_rttvar);
676 setmetric(RTV_HOPCOUNT, rmx_hopcount);
677 setmetric(RTV_MTU, rmx_mtu);
678 setmetric(RTV_EXPIRE, rmx_expire);
679#undef setmetric
984263bc
MD
680}
681
682#define ROUNDUP(a) \
683 ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long))
984263bc 684
984263bc
MD
685/*
686 * Extract the addresses of the passed sockaddrs.
687 * Do a little sanity checking so as to avoid bad memory references.
688 * This data is derived straight from userland.
689 */
690static int
ef87f48d 691rt_xaddrs(char *cp, char *cplim, struct rt_addrinfo *rtinfo)
984263bc 692{
82ed7fc2
RG
693 struct sockaddr *sa;
694 int i;
984263bc
MD
695
696 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) {
697 if ((rtinfo->rti_addrs & (1 << i)) == 0)
698 continue;
699 sa = (struct sockaddr *)cp;
700 /*
701 * It won't fit.
702 */
0c3c561c 703 if ((cp + sa->sa_len) > cplim) {
984263bc
MD
704 return (EINVAL);
705 }
706
707 /*
ef87f48d 708 * There are no more... Quit now.
984263bc 709 * If there are more bits, they are in error.
ef87f48d 710 * I've seen this. route(1) can evidently generate these.
984263bc 711 * This causes kernel to core dump.
ef87f48d 712 * For compatibility, if we see this, point to a safe address.
984263bc
MD
713 */
714 if (sa->sa_len == 0) {
590b8cd4
JH
715 static struct sockaddr sa_zero = {
716 sizeof sa_zero, AF_INET,
717 };
718
984263bc
MD
719 rtinfo->rti_info[i] = &sa_zero;
720 return (0); /* should be EINVAL but for compat */
721 }
722
ef87f48d 723 /* Accept the sockaddr. */
984263bc 724 rtinfo->rti_info[i] = sa;
590b8cd4 725 cp += ROUNDUP(sa->sa_len);
984263bc
MD
726 }
727 return (0);
728}
729
590b8cd4
JH
730static int
731rt_msghdrsize(int type)
984263bc 732{
984263bc 733 switch (type) {
984263bc
MD
734 case RTM_DELADDR:
735 case RTM_NEWADDR:
590b8cd4 736 return sizeof(struct ifa_msghdr);
984263bc
MD
737 case RTM_DELMADDR:
738 case RTM_NEWMADDR:
590b8cd4 739 return sizeof(struct ifma_msghdr);
984263bc 740 case RTM_IFINFO:
590b8cd4 741 return sizeof(struct if_msghdr);
984263bc 742 case RTM_IFANNOUNCE:
841ab66c 743 case RTM_IEEE80211:
590b8cd4 744 return sizeof(struct if_announcemsghdr);
984263bc 745 default:
590b8cd4 746 return sizeof(struct rt_msghdr);
984263bc 747 }
590b8cd4
JH
748}
749
750static int
751rt_msgsize(int type, struct rt_addrinfo *rtinfo)
752{
753 int len, i;
754
755 len = rt_msghdrsize(type);
756 for (i = 0; i < RTAX_MAX; i++) {
757 if (rtinfo->rti_info[i] != NULL)
758 len += ROUNDUP(rtinfo->rti_info[i]->sa_len);
984263bc 759 }
590b8cd4
JH
760 len = ALIGN(len);
761 return len;
762}
763
764/*
765 * Build a routing message in a buffer.
766 * Copy the addresses in the rtinfo->rti_info[] sockaddr array
767 * to the end of the buffer after the message header.
768 *
769 * Set the rtinfo->rti_addrs bitmask of addresses present in rtinfo->rti_info[].
770 * This side-effect can be avoided if we reorder the addrs bitmask field in all
771 * the route messages to line up so we can set it here instead of back in the
772 * calling routine.
773 */
774static void
775rt_msg_buffer(int type, struct rt_addrinfo *rtinfo, void *buf, int msglen)
776{
777 struct rt_msghdr *rtm;
778 char *cp;
779 int dlen, i;
780
781 rtm = (struct rt_msghdr *) buf;
782 rtm->rtm_version = RTM_VERSION;
783 rtm->rtm_type = type;
784 rtm->rtm_msglen = msglen;
785
786 cp = (char *)buf + rt_msghdrsize(type);
787 rtinfo->rti_addrs = 0;
984263bc 788 for (i = 0; i < RTAX_MAX; i++) {
590b8cd4
JH
789 struct sockaddr *sa;
790
984263bc
MD
791 if ((sa = rtinfo->rti_info[i]) == NULL)
792 continue;
793 rtinfo->rti_addrs |= (1 << i);
794 dlen = ROUNDUP(sa->sa_len);
590b8cd4
JH
795 bcopy(sa, cp, dlen);
796 cp += dlen;
984263bc 797 }
984263bc
MD
798}
799
590b8cd4
JH
800/*
801 * Build a routing message in a mbuf chain.
802 * Copy the addresses in the rtinfo->rti_info[] sockaddr array
803 * to the end of the mbuf after the message header.
804 *
805 * Set the rtinfo->rti_addrs bitmask of addresses present in rtinfo->rti_info[].
806 * This side-effect can be avoided if we reorder the addrs bitmask field in all
807 * the route messages to line up so we can set it here instead of back in the
808 * calling routine.
809 */
810static struct mbuf *
811rt_msg_mbuf(int type, struct rt_addrinfo *rtinfo)
984263bc 812{
590b8cd4
JH
813 struct mbuf *m;
814 struct rt_msghdr *rtm;
815 int hlen, len;
82ed7fc2 816 int i;
984263bc 817
590b8cd4
JH
818 hlen = rt_msghdrsize(type);
819 KASSERT(hlen <= MCLBYTES, ("rt_msg_mbuf: hlen %d doesn't fit", hlen));
984263bc 820
50503f0f 821 m = m_getl(hlen, MB_DONTWAIT, MT_DATA, M_PKTHDR, NULL);
590b8cd4
JH
822 if (m == NULL)
823 return (NULL);
590b8cd4
JH
824 m->m_pkthdr.len = m->m_len = hlen;
825 m->m_pkthdr.rcvif = NULL;
826 rtinfo->rti_addrs = 0;
827 len = hlen;
984263bc 828 for (i = 0; i < RTAX_MAX; i++) {
82ed7fc2 829 struct sockaddr *sa;
590b8cd4 830 int dlen;
984263bc 831
ef87f48d 832 if ((sa = rtinfo->rti_info[i]) == NULL)
984263bc
MD
833 continue;
834 rtinfo->rti_addrs |= (1 << i);
835 dlen = ROUNDUP(sa->sa_len);
590b8cd4 836 m_copyback(m, len, dlen, (caddr_t)sa); /* can grow mbuf chain */
984263bc
MD
837 len += dlen;
838 }
590b8cd4
JH
839 if (m->m_pkthdr.len != len) { /* one of the m_copyback() calls failed */
840 m_freem(m);
841 return (NULL);
984263bc 842 }
590b8cd4
JH
843 rtm = mtod(m, struct rt_msghdr *);
844 bzero(rtm, hlen);
845 rtm->rtm_msglen = len;
846 rtm->rtm_version = RTM_VERSION;
847 rtm->rtm_type = type;
848 return (m);
984263bc
MD
849}
850
851/*
852 * This routine is called to generate a message from the routing
f23061d4 853 * socket indicating that a redirect has occurred, a routing lookup
984263bc
MD
854 * has failed, or that a protocol has detected timeouts to a particular
855 * destination.
856 */
857void
f23061d4 858rt_missmsg(int type, struct rt_addrinfo *rtinfo, int flags, int error)
984263bc 859{
0c3c561c 860 struct sockaddr *dst = rtinfo->rti_info[RTAX_DST];
82ed7fc2
RG
861 struct rt_msghdr *rtm;
862 struct mbuf *m;
984263bc
MD
863
864 if (route_cb.any_count == 0)
865 return;
590b8cd4 866 m = rt_msg_mbuf(type, rtinfo);
ef87f48d 867 if (m == NULL)
984263bc
MD
868 return;
869 rtm = mtod(m, struct rt_msghdr *);
870 rtm->rtm_flags = RTF_DONE | flags;
871 rtm->rtm_errno = error;
872 rtm->rtm_addrs = rtinfo->rti_addrs;
590b8cd4 873 rts_input(m, familyof(dst));
984263bc
MD
874}
875
f3ed2586
JH
876void
877rt_dstmsg(int type, struct sockaddr *dst, int error)
878{
879 struct rt_msghdr *rtm;
880 struct rt_addrinfo addrs;
881 struct mbuf *m;
882
883 if (route_cb.any_count == 0)
884 return;
885 bzero(&addrs, sizeof(struct rt_addrinfo));
886 addrs.rti_info[RTAX_DST] = dst;
590b8cd4 887 m = rt_msg_mbuf(type, &addrs);
f3ed2586
JH
888 if (m == NULL)
889 return;
890 rtm = mtod(m, struct rt_msghdr *);
891 rtm->rtm_flags = RTF_DONE;
892 rtm->rtm_errno = error;
893 rtm->rtm_addrs = addrs.rti_addrs;
590b8cd4 894 rts_input(m, familyof(dst));
f3ed2586
JH
895}
896
984263bc
MD
897/*
898 * This routine is called to generate a message from the routing
899 * socket indicating that the status of a network interface has changed.
900 */
901void
0c3c561c 902rt_ifmsg(struct ifnet *ifp)
984263bc 903{
82ed7fc2 904 struct if_msghdr *ifm;
984263bc 905 struct mbuf *m;
590b8cd4 906 struct rt_addrinfo rtinfo;
984263bc
MD
907
908 if (route_cb.any_count == 0)
909 return;
590b8cd4
JH
910 bzero(&rtinfo, sizeof(struct rt_addrinfo));
911 m = rt_msg_mbuf(RTM_IFINFO, &rtinfo);
ef87f48d 912 if (m == NULL)
984263bc
MD
913 return;
914 ifm = mtod(m, struct if_msghdr *);
915 ifm->ifm_index = ifp->if_index;
9c095379 916 ifm->ifm_flags = ifp->if_flags;
984263bc 917 ifm->ifm_data = ifp->if_data;
590b8cd4
JH
918 ifm->ifm_addrs = 0;
919 rts_input(m, 0);
984263bc
MD
920}
921
372316d9
JH
922static void
923rt_ifamsg(int cmd, struct ifaddr *ifa)
924{
925 struct ifa_msghdr *ifam;
590b8cd4 926 struct rt_addrinfo rtinfo;
372316d9 927 struct mbuf *m;
372316d9
JH
928 struct ifnet *ifp = ifa->ifa_ifp;
929
590b8cd4
JH
930 bzero(&rtinfo, sizeof(struct rt_addrinfo));
931 rtinfo.rti_ifaaddr = ifa->ifa_addr;
932 rtinfo.rti_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr;
933 rtinfo.rti_netmask = ifa->ifa_netmask;
934 rtinfo.rti_bcastaddr = ifa->ifa_dstaddr;
f23061d4 935
590b8cd4 936 m = rt_msg_mbuf(cmd, &rtinfo);
f23061d4 937 if (m == NULL)
372316d9 938 return;
f23061d4 939
372316d9
JH
940 ifam = mtod(m, struct ifa_msghdr *);
941 ifam->ifam_index = ifp->if_index;
942 ifam->ifam_metric = ifa->ifa_metric;
943 ifam->ifam_flags = ifa->ifa_flags;
590b8cd4 944 ifam->ifam_addrs = rtinfo.rti_addrs;
f23061d4 945
590b8cd4 946 rts_input(m, familyof(ifa->ifa_addr));
372316d9
JH
947}
948
f3ed2586
JH
949void
950rt_rtmsg(int cmd, struct rtentry *rt, struct ifnet *ifp, int error)
372316d9
JH
951{
952 struct rt_msghdr *rtm;
590b8cd4 953 struct rt_addrinfo rtinfo;
372316d9 954 struct mbuf *m;
f3ed2586 955 struct sockaddr *dst;
372316d9
JH
956
957 if (rt == NULL)
958 return;
f23061d4 959
590b8cd4
JH
960 bzero(&rtinfo, sizeof(struct rt_addrinfo));
961 rtinfo.rti_dst = dst = rt_key(rt);
962 rtinfo.rti_gateway = rt->rt_gateway;
963 rtinfo.rti_netmask = rt_mask(rt);
f3ed2586 964 if (ifp != NULL)
590b8cd4
JH
965 rtinfo.rti_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr;
966 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr;
f23061d4 967
590b8cd4 968 m = rt_msg_mbuf(cmd, &rtinfo);
f23061d4 969 if (m == NULL)
372316d9 970 return;
f23061d4 971
372316d9 972 rtm = mtod(m, struct rt_msghdr *);
f3ed2586
JH
973 if (ifp != NULL)
974 rtm->rtm_index = ifp->if_index;
372316d9
JH
975 rtm->rtm_flags |= rt->rt_flags;
976 rtm->rtm_errno = error;
590b8cd4 977 rtm->rtm_addrs = rtinfo.rti_addrs;
f23061d4 978
590b8cd4 979 rts_input(m, familyof(dst));
372316d9
JH
980}
981
984263bc
MD
982/*
983 * This is called to generate messages from the routing socket
984 * indicating a network interface has had addresses associated with it.
985 * if we ever reverse the logic and replace messages TO the routing
986 * socket indicate a request to configure interfaces, then it will
987 * be unnecessary as the routing socket will automatically generate
988 * copies of it.
989 */
990void
0c3c561c 991rt_newaddrmsg(int cmd, struct ifaddr *ifa, int error, struct rtentry *rt)
984263bc 992{
78812139
EN
993#ifdef SCTP
994 /*
995 * notify the SCTP stack
996 * this will only get called when an address is added/deleted
997 * XXX pass the ifaddr struct instead if ifa->ifa_addr...
998 */
999 if (cmd == RTM_ADD)
1000 sctp_add_ip_address(ifa);
1001 else if (cmd == RTM_DELETE)
1002 sctp_delete_ip_address(ifa);
1003#endif /* SCTP */
1004
984263bc
MD
1005 if (route_cb.any_count == 0)
1006 return;
984263bc 1007
372316d9
JH
1008 if (cmd == RTM_ADD) {
1009 rt_ifamsg(RTM_NEWADDR, ifa);
f3ed2586 1010 rt_rtmsg(RTM_ADD, rt, ifa->ifa_ifp, error);
372316d9
JH
1011 } else {
1012 KASSERT((cmd == RTM_DELETE), ("unknown cmd %d", cmd));
f3ed2586 1013 rt_rtmsg(RTM_DELETE, rt, ifa->ifa_ifp, error);
372316d9 1014 rt_ifamsg(RTM_DELADDR, ifa);
984263bc
MD
1015 }
1016}
1017
1018/*
1019 * This is the analogue to the rt_newaddrmsg which performs the same
1020 * function but for multicast group memberhips. This is easier since
1021 * there is no route state to worry about.
1022 */
1023void
0c3c561c 1024rt_newmaddrmsg(int cmd, struct ifmultiaddr *ifma)
984263bc 1025{
590b8cd4 1026 struct rt_addrinfo rtinfo;
ef87f48d 1027 struct mbuf *m = NULL;
984263bc
MD
1028 struct ifnet *ifp = ifma->ifma_ifp;
1029 struct ifma_msghdr *ifmam;
1030
1031 if (route_cb.any_count == 0)
1032 return;
1033
590b8cd4
JH
1034 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1035 rtinfo.rti_ifaaddr = ifma->ifma_addr;
1782116d 1036 if (ifp != NULL && !TAILQ_EMPTY(&ifp->if_addrhead))
590b8cd4 1037 rtinfo.rti_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr;
984263bc
MD
1038 /*
1039 * If a link-layer address is present, present it as a ``gateway''
1040 * (similarly to how ARP entries, e.g., are presented).
1041 */
590b8cd4 1042 rtinfo.rti_gateway = ifma->ifma_lladdr;
f23061d4 1043
590b8cd4 1044 m = rt_msg_mbuf(cmd, &rtinfo);
f23061d4 1045 if (m == NULL)
984263bc 1046 return;
f23061d4 1047
984263bc
MD
1048 ifmam = mtod(m, struct ifma_msghdr *);
1049 ifmam->ifmam_index = ifp->if_index;
590b8cd4 1050 ifmam->ifmam_addrs = rtinfo.rti_addrs;
f23061d4 1051
590b8cd4 1052 rts_input(m, familyof(ifma->ifma_addr));
984263bc
MD
1053}
1054
841ab66c
SZ
1055static struct mbuf *
1056rt_makeifannouncemsg(struct ifnet *ifp, int type, int what,
1057 struct rt_addrinfo *info)
984263bc 1058{
590b8cd4 1059 struct if_announcemsghdr *ifan;
841ab66c 1060 struct mbuf *m;
984263bc
MD
1061
1062 if (route_cb.any_count == 0)
841ab66c 1063 return NULL;
f23061d4 1064
841ab66c
SZ
1065 bzero(info, sizeof(*info));
1066 m = rt_msg_mbuf(type, info);
984263bc 1067 if (m == NULL)
841ab66c 1068 return NULL;
f23061d4 1069
984263bc
MD
1070 ifan = mtod(m, struct if_announcemsghdr *);
1071 ifan->ifan_index = ifp->if_index;
5fe66e68 1072 strlcpy(ifan->ifan_name, ifp->if_xname, sizeof ifan->ifan_name);
984263bc 1073 ifan->ifan_what = what;
841ab66c
SZ
1074 return m;
1075}
1076
1077/*
1078 * This is called to generate routing socket messages indicating
1079 * IEEE80211 wireless events.
1080 * XXX we piggyback on the RTM_IFANNOUNCE msg format in a clumsy way.
1081 */
1082void
1083rt_ieee80211msg(struct ifnet *ifp, int what, void *data, size_t data_len)
1084{
1085 struct rt_addrinfo info;
1086 struct mbuf *m;
f23061d4 1087
841ab66c
SZ
1088 m = rt_makeifannouncemsg(ifp, RTM_IEEE80211, what, &info);
1089 if (m == NULL)
1090 return;
1091
1092 /*
1093 * Append the ieee80211 data. Try to stick it in the
1094 * mbuf containing the ifannounce msg; otherwise allocate
1095 * a new mbuf and append.
1096 *
1097 * NB: we assume m is a single mbuf.
1098 */
1099 if (data_len > M_TRAILINGSPACE(m)) {
b11e1686 1100 struct mbuf *n = m_get(MB_DONTWAIT, MT_DATA);
841ab66c
SZ
1101 if (n == NULL) {
1102 m_freem(m);
1103 return;
1104 }
1105 bcopy(data, mtod(n, void *), data_len);
1106 n->m_len = data_len;
1107 m->m_next = n;
1108 } else if (data_len > 0) {
1109 bcopy(data, mtod(m, u_int8_t *) + m->m_len, data_len);
1110 m->m_len += data_len;
1111 }
1112 if (m->m_flags & M_PKTHDR)
1113 m->m_pkthdr.len += data_len;
1114 mtod(m, struct if_announcemsghdr *)->ifan_msglen += data_len;
590b8cd4 1115 rts_input(m, 0);
841ab66c
SZ
1116}
1117
1118/*
1119 * This is called to generate routing socket messages indicating
1120 * network interface arrival and departure.
1121 */
1122void
1123rt_ifannouncemsg(struct ifnet *ifp, int what)
1124{
1125 struct rt_addrinfo addrinfo;
1126 struct mbuf *m;
1127
1128 m = rt_makeifannouncemsg(ifp, RTM_IFANNOUNCE, what, &addrinfo);
1129 if (m != NULL)
1130 rts_input(m, 0);
590b8cd4 1131}
f23061d4 1132
590b8cd4
JH
1133static int
1134resizewalkarg(struct walkarg *w, int len)
1135{
1136 void *newptr;
1137
efda3bd0 1138 newptr = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
590b8cd4
JH
1139 if (newptr == NULL)
1140 return (ENOMEM);
1141 if (w->w_tmem != NULL)
efda3bd0 1142 kfree(w->w_tmem, M_RTABLE);
590b8cd4
JH
1143 w->w_tmem = newptr;
1144 w->w_tmemsize = len;
1145 return (0);
1146}
984263bc
MD
1147
1148/*
1149 * This is used in dumping the kernel table via sysctl().
1150 */
1151int
590b8cd4 1152sysctl_dumpentry(struct radix_node *rn, void *vw)
984263bc 1153{
82ed7fc2
RG
1154 struct walkarg *w = vw;
1155 struct rtentry *rt = (struct rtentry *)rn;
590b8cd4
JH
1156 struct rt_addrinfo rtinfo;
1157 int error, msglen;
984263bc
MD
1158
1159 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg))
1160 return 0;
f23061d4 1161
590b8cd4
JH
1162 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1163 rtinfo.rti_dst = rt_key(rt);
1164 rtinfo.rti_gateway = rt->rt_gateway;
1165 rtinfo.rti_netmask = rt_mask(rt);
1166 rtinfo.rti_genmask = rt->rt_genmask;
ef87f48d 1167 if (rt->rt_ifp != NULL) {
590b8cd4 1168 rtinfo.rti_ifpaddr =
ef87f48d 1169 TAILQ_FIRST(&rt->rt_ifp->if_addrhead)->ifa_addr;
590b8cd4 1170 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr;
984263bc 1171 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT)
590b8cd4 1172 rtinfo.rti_bcastaddr = rt->rt_ifa->ifa_dstaddr;
984263bc 1173 }
590b8cd4
JH
1174 msglen = rt_msgsize(RTM_GET, &rtinfo);
1175 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0)
1176 return (ENOMEM);
1177 rt_msg_buffer(RTM_GET, &rtinfo, w->w_tmem, msglen);
1178 if (w->w_req != NULL) {
8e63efec 1179 struct rt_msghdr *rtm = w->w_tmem;
984263bc
MD
1180
1181 rtm->rtm_flags = rt->rt_flags;
1182 rtm->rtm_use = rt->rt_use;
1183 rtm->rtm_rmx = rt->rt_rmx;
1184 rtm->rtm_index = rt->rt_ifp->if_index;
1185 rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0;
590b8cd4
JH
1186 rtm->rtm_addrs = rtinfo.rti_addrs;
1187 error = SYSCTL_OUT(w->w_req, rtm, msglen);
984263bc
MD
1188 return (error);
1189 }
590b8cd4 1190 return (0);
984263bc
MD
1191}
1192
590b8cd4
JH
1193static int
1194sysctl_iflist(int af, struct walkarg *w)
984263bc 1195{
82ed7fc2
RG
1196 struct ifnet *ifp;
1197 struct ifaddr *ifa;
590b8cd4
JH
1198 struct rt_addrinfo rtinfo;
1199 int msglen, error;
984263bc 1200
590b8cd4 1201 bzero(&rtinfo, sizeof(struct rt_addrinfo));
984263bc
MD
1202 TAILQ_FOREACH(ifp, &ifnet, if_link) {
1203 if (w->w_arg && w->w_arg != ifp->if_index)
1204 continue;
1205 ifa = TAILQ_FIRST(&ifp->if_addrhead);
590b8cd4
JH
1206 rtinfo.rti_ifpaddr = ifa->ifa_addr;
1207 msglen = rt_msgsize(RTM_IFINFO, &rtinfo);
1208 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0)
1209 return (ENOMEM);
1210 rt_msg_buffer(RTM_IFINFO, &rtinfo, w->w_tmem, msglen);
1211 rtinfo.rti_ifpaddr = NULL;
ef87f48d 1212 if (w->w_req != NULL && w->w_tmem != NULL) {
8e63efec 1213 struct if_msghdr *ifm = w->w_tmem;
984263bc 1214
984263bc 1215 ifm->ifm_index = ifp->if_index;
9c095379 1216 ifm->ifm_flags = ifp->if_flags;
984263bc 1217 ifm->ifm_data = ifp->if_data;
590b8cd4
JH
1218 ifm->ifm_addrs = rtinfo.rti_addrs;
1219 error = SYSCTL_OUT(w->w_req, ifm, msglen);
984263bc
MD
1220 if (error)
1221 return (error);
1222 }
ef87f48d 1223 while ((ifa = TAILQ_NEXT(ifa, ifa_link)) != NULL) {
984263bc
MD
1224 if (af && af != ifa->ifa_addr->sa_family)
1225 continue;
590b8cd4 1226 if (curproc->p_ucred->cr_prison &&
87de5057 1227 prison_if(curproc->p_ucred, ifa->ifa_addr))
984263bc 1228 continue;
590b8cd4
JH
1229 rtinfo.rti_ifaaddr = ifa->ifa_addr;
1230 rtinfo.rti_netmask = ifa->ifa_netmask;
1231 rtinfo.rti_bcastaddr = ifa->ifa_dstaddr;
1232 msglen = rt_msgsize(RTM_NEWADDR, &rtinfo);
1233 if (w->w_tmemsize < msglen &&
1234 resizewalkarg(w, msglen) != 0)
1235 return (ENOMEM);
1236 rt_msg_buffer(RTM_NEWADDR, &rtinfo, w->w_tmem, msglen);
1237 if (w->w_req != NULL) {
8e63efec 1238 struct ifa_msghdr *ifam = w->w_tmem;
984263bc 1239
984263bc
MD
1240 ifam->ifam_index = ifa->ifa_ifp->if_index;
1241 ifam->ifam_flags = ifa->ifa_flags;
1242 ifam->ifam_metric = ifa->ifa_metric;
590b8cd4
JH
1243 ifam->ifam_addrs = rtinfo.rti_addrs;
1244 error = SYSCTL_OUT(w->w_req, w->w_tmem, msglen);
984263bc
MD
1245 if (error)
1246 return (error);
1247 }
1248 }
590b8cd4
JH
1249 rtinfo.rti_netmask = NULL;
1250 rtinfo.rti_ifaaddr = NULL;
1251 rtinfo.rti_bcastaddr = NULL;
984263bc
MD
1252 }
1253 return (0);
1254}
1255
1256static int
1257sysctl_rtsock(SYSCTL_HANDLER_ARGS)
1258{
1259 int *name = (int *)arg1;
1260 u_int namelen = arg2;
82ed7fc2 1261 struct radix_node_head *rnh;
a49aa710 1262 int i, error = EINVAL;
ecdefdda 1263 int origcpu;
984263bc
MD
1264 u_char af;
1265 struct walkarg w;
1266
1267 name ++;
1268 namelen--;
1269 if (req->newptr)
1270 return (EPERM);
ecdefdda 1271 if (namelen != 3 && namelen != 4)
984263bc
MD
1272 return (EINVAL);
1273 af = name[0];
5fe66e68 1274 bzero(&w, sizeof w);
984263bc
MD
1275 w.w_op = name[1];
1276 w.w_arg = name[2];
1277 w.w_req = req;
1278
ecdefdda
MD
1279 /*
1280 * Optional third argument specifies cpu, used primarily for
1281 * debugging the route table.
1282 */
1283 if (namelen == 4) {
1284 if (name[3] < 0 || name[3] >= ncpus)
1285 return (EINVAL);
1286 origcpu = mycpuid;
1287 lwkt_migratecpu(name[3]);
1288 } else {
1289 origcpu = -1;
1290 }
a49aa710 1291 crit_enter();
984263bc 1292 switch (w.w_op) {
984263bc
MD
1293 case NET_RT_DUMP:
1294 case NET_RT_FLAGS:
1295 for (i = 1; i <= AF_MAX; i++)
ecdefdda
MD
1296 if ((rnh = rt_tables[mycpuid][i]) &&
1297 (af == 0 || af == i) &&
984263bc 1298 (error = rnh->rnh_walktree(rnh,
f23061d4 1299 sysctl_dumpentry, &w)))
984263bc
MD
1300 break;
1301 break;
1302
1303 case NET_RT_IFLIST:
1304 error = sysctl_iflist(af, &w);
1305 }
a49aa710 1306 crit_exit();
590b8cd4 1307 if (w.w_tmem != NULL)
efda3bd0 1308 kfree(w.w_tmem, M_RTABLE);
ecdefdda
MD
1309 if (origcpu >= 0)
1310 lwkt_migratecpu(origcpu);
984263bc
MD
1311 return (error);
1312}
1313
1314SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, "");
1315
1316/*
1317 * Definitions of protocols supported in the ROUTE domain.
1318 */
1319
ba4257b8 1320static struct domain routedomain; /* or at least forward */
984263bc
MD
1321
1322static struct protosw routesw[] = {
1323{ SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR,
1324 0, route_output, raw_ctlinput, 0,
9eeaa8a9 1325 cpu0_soport,
984263bc
MD
1326 raw_init, 0, 0, 0,
1327 &route_usrreqs
1328}
1329};
1330
ba4257b8 1331static struct domain routedomain = {
9c70fe43
JH
1332 PF_ROUTE, "route", NULL, NULL, NULL,
1333 routesw, &routesw[(sizeof routesw)/(sizeof routesw[0])],
1334};
984263bc
MD
1335
1336DOMAIN_SET(route);