Make 'last mounted on' reporting by fsck work again. Add a new option
[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 $
0b30dad7 67 * $DragonFly: src/sys/net/rtsock.c,v 1.37 2007/03/04 18:51:59 swildner 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 = {
310 rts_abort, pru_accept_notsupp, rts_attach, rts_bind, rts_connect,
311 pru_connect2_notsupp, pru_control_notsupp, rts_detach, rts_disconnect,
312 pru_listen_notsupp, rts_peeraddr, pru_rcvd_notsupp, pru_rcvoob_notsupp,
313 rts_send, pru_sense_null, rts_shutdown, rts_sockaddr,
314 sosend, soreceive, sopoll
315};
316
590b8cd4
JH
317static __inline sa_family_t
318familyof(struct sockaddr *sa)
319{
320 return (sa != NULL ? sa->sa_family : 0);
321}
322
323static void
324rts_input(struct mbuf *m, sa_family_t family)
325{
326 static const struct sockaddr route_dst = { 2, PF_ROUTE, };
327 struct sockproto route_proto = { PF_ROUTE, family };
328
329 raw_input(m, &route_proto, &route_src, &route_dst);
330}
331
332static void *
333reallocbuf(void *ptr, size_t len, size_t olen)
334{
335 void *newptr;
336
efda3bd0 337 newptr = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
590b8cd4
JH
338 if (newptr == NULL)
339 return NULL;
340 bcopy(ptr, newptr, olen);
efda3bd0 341 kfree(ptr, M_RTABLE);
590b8cd4
JH
342 return (newptr);
343}
344
8e63efec
JH
345/*
346 * Internal helper routine for route_output().
347 */
590b8cd4
JH
348static int
349fillrtmsg(struct rt_msghdr **prtm, struct rtentry *rt,
350 struct rt_addrinfo *rtinfo)
351{
352 int msglen;
353 struct rt_msghdr *rtm = *prtm;
354
355 /* Fill in rt_addrinfo for call to rt_msg_buffer(). */
356 rtinfo->rti_dst = rt_key(rt);
357 rtinfo->rti_gateway = rt->rt_gateway;
358 rtinfo->rti_netmask = rt_mask(rt); /* might be NULL */
359 rtinfo->rti_genmask = rt->rt_genmask; /* might be NULL */
360 if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) {
361 if (rt->rt_ifp != NULL) {
362 rtinfo->rti_ifpaddr =
363 TAILQ_FIRST(&rt->rt_ifp->if_addrhead)->ifa_addr;
364 rtinfo->rti_ifaaddr = rt->rt_ifa->ifa_addr;
365 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT)
366 rtinfo->rti_bcastaddr = rt->rt_ifa->ifa_dstaddr;
367 rtm->rtm_index = rt->rt_ifp->if_index;
368 } else {
369 rtinfo->rti_ifpaddr = NULL;
370 rtinfo->rti_ifaaddr = NULL;
371 }
372 }
373
374 msglen = rt_msgsize(rtm->rtm_type, rtinfo);
375 if (rtm->rtm_msglen < msglen) {
376 rtm = reallocbuf(rtm, msglen, rtm->rtm_msglen);
377 if (rtm == NULL)
378 return (ENOBUFS);
379 *prtm = rtm;
380 }
381 rt_msg_buffer(rtm->rtm_type, rtinfo, rtm, msglen);
382
383 rtm->rtm_flags = rt->rt_flags;
384 rtm->rtm_rmx = rt->rt_rmx;
385 rtm->rtm_addrs = rtinfo->rti_addrs;
386
387 return (0);
388}
389
ecdefdda
MD
390static void route_output_add_callback(int, int, struct rt_addrinfo *,
391 struct rtentry *, void *);
392static void route_output_delete_callback(int, int, struct rt_addrinfo *,
393 struct rtentry *, void *);
394static void route_output_change_callback(int, int, struct rt_addrinfo *,
395 struct rtentry *, void *);
396static void route_output_lock_callback(int, int, struct rt_addrinfo *,
397 struct rtentry *, void *);
398
984263bc
MD
399/*ARGSUSED*/
400static int
deffea2e 401route_output(struct mbuf *m, struct socket *so, ...)
984263bc 402{
2e9572df 403 struct rt_msghdr *rtm = NULL;
ecdefdda 404 struct rtentry *rt;
984263bc 405 struct radix_node_head *rnh;
f23061d4 406 struct rawcb *rp = NULL;
deffea2e 407 struct pr_output_info *oi;
590b8cd4 408 struct rt_addrinfo rtinfo;
ef87f48d 409 int len, error = 0;
deffea2e
JS
410 __va_list ap;
411
412 __va_start(ap, so);
413 oi = __va_arg(ap, struct pr_output_info *);
414 __va_end(ap);
984263bc 415
ef87f48d 416#define gotoerr(e) { error = e; goto flush;}
590b8cd4
JH
417
418 if (m == NULL ||
419 (m->m_len < sizeof(long) &&
420 (m = m_pullup(m, sizeof(long))) == NULL))
984263bc 421 return (ENOBUFS);
ef87f48d 422 if (!(m->m_flags & M_PKTHDR))
984263bc
MD
423 panic("route_output");
424 len = m->m_pkthdr.len;
590b8cd4 425 if (len < sizeof(struct rt_msghdr) ||
984263bc 426 len != mtod(m, struct rt_msghdr *)->rtm_msglen) {
590b8cd4 427 rtinfo.rti_dst = NULL;
ef87f48d 428 gotoerr(EINVAL);
984263bc 429 }
efda3bd0 430 rtm = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
ef87f48d 431 if (rtm == NULL) {
590b8cd4 432 rtinfo.rti_dst = NULL;
ef87f48d 433 gotoerr(ENOBUFS);
984263bc
MD
434 }
435 m_copydata(m, 0, len, (caddr_t)rtm);
436 if (rtm->rtm_version != RTM_VERSION) {
590b8cd4 437 rtinfo.rti_dst = NULL;
ef87f48d 438 gotoerr(EPROTONOSUPPORT);
984263bc 439 }
47e78ae8 440 rtm->rtm_pid = oi->p_pid;
590b8cd4
JH
441 bzero(&rtinfo, sizeof(struct rt_addrinfo));
442 rtinfo.rti_addrs = rtm->rtm_addrs;
443 if (rt_xaddrs((char *)(rtm + 1), (char *)rtm + len, &rtinfo) != 0) {
444 rtinfo.rti_dst = NULL;
ef87f48d 445 gotoerr(EINVAL);
984263bc 446 }
590b8cd4
JH
447 rtinfo.rti_flags = rtm->rtm_flags;
448 if (rtinfo.rti_dst == NULL || rtinfo.rti_dst->sa_family >= AF_MAX ||
449 (rtinfo.rti_gateway && rtinfo.rti_gateway->sa_family >= AF_MAX))
ef87f48d
JH
450 gotoerr(EINVAL);
451
590b8cd4
JH
452 if (rtinfo.rti_genmask != NULL) {
453 struct radix_node *n;
ef87f48d 454
590b8cd4
JH
455#define clen(s) (*(u_char *)(s))
456 n = rn_addmask((char *)rtinfo.rti_genmask, TRUE, 1);
457 if (n != NULL &&
458 rtinfo.rti_genmask->sa_len >= clen(n->rn_key) &&
459 bcmp((char *)rtinfo.rti_genmask + 1,
460 (char *)n->rn_key + 1, clen(n->rn_key) - 1) == 0)
461 rtinfo.rti_genmask = (struct sockaddr *)n->rn_key;
984263bc 462 else
ef87f48d 463 gotoerr(ENOBUFS);
984263bc
MD
464 }
465
466 /*
467 * Verify that the caller has the appropriate privilege; RTM_GET
468 * is the only operation the non-superuser is allowed.
469 */
dadab5e9 470 if (rtm->rtm_type != RTM_GET && suser_cred(so->so_cred, 0) != 0)
ef87f48d 471 gotoerr(EPERM);
984263bc
MD
472
473 switch (rtm->rtm_type) {
984263bc 474 case RTM_ADD:
ecdefdda
MD
475 if (rtinfo.rti_gateway == NULL) {
476 error = EINVAL;
477 } else {
478 error = rtrequest1_global(RTM_ADD, &rtinfo,
479 route_output_add_callback, rtm);
984263bc
MD
480 }
481 break;
984263bc 482 case RTM_DELETE:
ecdefdda
MD
483 /*
484 * note: &rtm passed as argument so 'rtm' can be replaced.
485 */
486 error = rtrequest1_global(RTM_DELETE, &rtinfo,
487 route_output_delete_callback, &rtm);
984263bc 488 break;
984263bc 489 case RTM_GET:
ecdefdda
MD
490 rnh = rt_tables[mycpuid][rtinfo.rti_dst->sa_family];
491 if (rnh == NULL) {
492 error = EAFNOSUPPORT;
493 break;
494 }
590b8cd4
JH
495 rt = (struct rtentry *)
496 rnh->rnh_lookup((char *)rtinfo.rti_dst,
497 (char *)rtinfo.rti_netmask, rnh);
ecdefdda
MD
498 if (rt == NULL) {
499 error = ESRCH;
984263bc
MD
500 break;
501 }
ecdefdda
MD
502 rt->rt_refcnt++;
503 if (fillrtmsg(&rtm, rt, &rtinfo) != 0)
504 gotoerr(ENOBUFS);
505 --rt->rt_refcnt;
506 break;
507 case RTM_CHANGE:
508 error = rtrequest1_global(RTM_GET, &rtinfo,
509 route_output_change_callback, rtm);
510 break;
511 case RTM_LOCK:
512 error = rtrequest1_global(RTM_GET, &rtinfo,
513 route_output_lock_callback, rtm);
590b8cd4 514 break;
984263bc 515 default:
ecdefdda
MD
516 error = EOPNOTSUPP;
517 break;
984263bc
MD
518 }
519
520flush:
590b8cd4
JH
521 if (rtm != NULL) {
522 if (error != 0)
984263bc
MD
523 rtm->rtm_errno = error;
524 else
525 rtm->rtm_flags |= RTF_DONE;
526 }
ecdefdda 527
984263bc
MD
528 /*
529 * Check to see if we don't want our own messages.
530 */
ef87f48d 531 if (!(so->so_options & SO_USELOOPBACK)) {
984263bc 532 if (route_cb.any_count <= 1) {
590b8cd4 533 if (rtm != NULL)
efda3bd0 534 kfree(rtm, M_RTABLE);
984263bc
MD
535 m_freem(m);
536 return (error);
537 }
538 /* There is another listener, so construct message */
539 rp = sotorawcb(so);
540 }
590b8cd4 541 if (rtm != NULL) {
984263bc
MD
542 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm);
543 if (m->m_pkthdr.len < rtm->rtm_msglen) {
544 m_freem(m);
545 m = NULL;
546 } else if (m->m_pkthdr.len > rtm->rtm_msglen)
547 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len);
efda3bd0 548 kfree(rtm, M_RTABLE);
984263bc 549 }
ef87f48d 550 if (rp != NULL)
984263bc 551 rp->rcb_proto.sp_family = 0; /* Avoid us */
ef87f48d 552 if (m != NULL)
590b8cd4 553 rts_input(m, familyof(rtinfo.rti_dst));
ef87f48d 554 if (rp != NULL)
984263bc 555 rp->rcb_proto.sp_family = PF_ROUTE;
984263bc
MD
556 return (error);
557}
558
ecdefdda
MD
559static void
560route_output_add_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
561 struct rtentry *rt, void *arg)
562{
563 struct rt_msghdr *rtm = arg;
564
565 if (error == 0 && rt != NULL) {
566 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx,
567 &rt->rt_rmx);
568 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits);
569 rt->rt_rmx.rmx_locks |=
570 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks);
571 rt->rt_genmask = rtinfo->rti_genmask;
572 }
573}
574
575static void
576route_output_delete_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
577 struct rtentry *rt, void *arg)
578{
579 struct rt_msghdr **rtm = arg;
580
581 if (error == 0 && rt) {
582 ++rt->rt_refcnt;
583 if (fillrtmsg(rtm, rt, rtinfo) != 0) {
584 error = ENOBUFS;
585 /* XXX no way to return the error */
586 }
587 --rt->rt_refcnt;
588 }
589}
590
591static void
592route_output_change_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
593 struct rtentry *rt, void *arg)
594{
595 struct rt_msghdr *rtm = arg;
596 struct ifaddr *ifa;
597
598 if (error)
599 goto done;
600
601 /*
602 * new gateway could require new ifaddr, ifp;
603 * flags may also be different; ifp may be specified
604 * by ll sockaddr when protocol address is ambiguous
605 */
606 if (((rt->rt_flags & RTF_GATEWAY) && rtinfo->rti_gateway != NULL) ||
607 rtinfo->rti_ifpaddr != NULL || (rtinfo->rti_ifaaddr != NULL &&
608 sa_equal(rtinfo->rti_ifaaddr, rt->rt_ifa->ifa_addr))
609 ) {
610 error = rt_getifa(rtinfo);
611 if (error != 0)
612 goto done;
613 }
614 if (rtinfo->rti_gateway != NULL) {
615 error = rt_setgate(rt, rt_key(rt), rtinfo->rti_gateway);
616 if (error != 0)
617 goto done;
618 }
619 if ((ifa = rtinfo->rti_ifa) != NULL) {
620 struct ifaddr *oifa = rt->rt_ifa;
621
622 if (oifa != ifa) {
623 if (oifa && oifa->ifa_rtrequest)
624 oifa->ifa_rtrequest(RTM_DELETE, rt, rtinfo);
625 IFAFREE(rt->rt_ifa);
626 IFAREF(ifa);
627 rt->rt_ifa = ifa;
628 rt->rt_ifp = rtinfo->rti_ifp;
629 }
630 }
631 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, &rt->rt_rmx);
632 if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest)
633 rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, rtinfo);
634 if (rtinfo->rti_genmask != NULL)
635 rt->rt_genmask = rtinfo->rti_genmask;
636done:
637 /* XXX no way to return error */
638 ;
639}
640
641static void
642route_output_lock_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
643 struct rtentry *rt, void *arg)
644{
645 struct rt_msghdr *rtm = arg;
646
647 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits);
648 rt->rt_rmx.rmx_locks |=
649 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks);
650}
651
984263bc 652static void
ef87f48d 653rt_setmetrics(u_long which, struct rt_metrics *in, struct rt_metrics *out)
984263bc 654{
2e9572df
JH
655#define setmetric(flag, elt) if (which & (flag)) out->elt = in->elt;
656 setmetric(RTV_RPIPE, rmx_recvpipe);
657 setmetric(RTV_SPIPE, rmx_sendpipe);
658 setmetric(RTV_SSTHRESH, rmx_ssthresh);
659 setmetric(RTV_RTT, rmx_rtt);
660 setmetric(RTV_RTTVAR, rmx_rttvar);
661 setmetric(RTV_HOPCOUNT, rmx_hopcount);
662 setmetric(RTV_MTU, rmx_mtu);
663 setmetric(RTV_EXPIRE, rmx_expire);
664#undef setmetric
984263bc
MD
665}
666
667#define ROUNDUP(a) \
668 ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long))
984263bc 669
984263bc
MD
670/*
671 * Extract the addresses of the passed sockaddrs.
672 * Do a little sanity checking so as to avoid bad memory references.
673 * This data is derived straight from userland.
674 */
675static int
ef87f48d 676rt_xaddrs(char *cp, char *cplim, struct rt_addrinfo *rtinfo)
984263bc 677{
82ed7fc2
RG
678 struct sockaddr *sa;
679 int i;
984263bc
MD
680
681 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) {
682 if ((rtinfo->rti_addrs & (1 << i)) == 0)
683 continue;
684 sa = (struct sockaddr *)cp;
685 /*
686 * It won't fit.
687 */
0c3c561c 688 if ((cp + sa->sa_len) > cplim) {
984263bc
MD
689 return (EINVAL);
690 }
691
692 /*
ef87f48d 693 * There are no more... Quit now.
984263bc 694 * If there are more bits, they are in error.
ef87f48d 695 * I've seen this. route(1) can evidently generate these.
984263bc 696 * This causes kernel to core dump.
ef87f48d 697 * For compatibility, if we see this, point to a safe address.
984263bc
MD
698 */
699 if (sa->sa_len == 0) {
590b8cd4
JH
700 static struct sockaddr sa_zero = {
701 sizeof sa_zero, AF_INET,
702 };
703
984263bc
MD
704 rtinfo->rti_info[i] = &sa_zero;
705 return (0); /* should be EINVAL but for compat */
706 }
707
ef87f48d 708 /* Accept the sockaddr. */
984263bc 709 rtinfo->rti_info[i] = sa;
590b8cd4 710 cp += ROUNDUP(sa->sa_len);
984263bc
MD
711 }
712 return (0);
713}
714
590b8cd4
JH
715static int
716rt_msghdrsize(int type)
984263bc 717{
984263bc 718 switch (type) {
984263bc
MD
719 case RTM_DELADDR:
720 case RTM_NEWADDR:
590b8cd4 721 return sizeof(struct ifa_msghdr);
984263bc
MD
722 case RTM_DELMADDR:
723 case RTM_NEWMADDR:
590b8cd4 724 return sizeof(struct ifma_msghdr);
984263bc 725 case RTM_IFINFO:
590b8cd4 726 return sizeof(struct if_msghdr);
984263bc 727 case RTM_IFANNOUNCE:
841ab66c 728 case RTM_IEEE80211:
590b8cd4 729 return sizeof(struct if_announcemsghdr);
984263bc 730 default:
590b8cd4 731 return sizeof(struct rt_msghdr);
984263bc 732 }
590b8cd4
JH
733}
734
735static int
736rt_msgsize(int type, struct rt_addrinfo *rtinfo)
737{
738 int len, i;
739
740 len = rt_msghdrsize(type);
741 for (i = 0; i < RTAX_MAX; i++) {
742 if (rtinfo->rti_info[i] != NULL)
743 len += ROUNDUP(rtinfo->rti_info[i]->sa_len);
984263bc 744 }
590b8cd4
JH
745 len = ALIGN(len);
746 return len;
747}
748
749/*
750 * Build a routing message in a buffer.
751 * Copy the addresses in the rtinfo->rti_info[] sockaddr array
752 * to the end of the buffer after the message header.
753 *
754 * Set the rtinfo->rti_addrs bitmask of addresses present in rtinfo->rti_info[].
755 * This side-effect can be avoided if we reorder the addrs bitmask field in all
756 * the route messages to line up so we can set it here instead of back in the
757 * calling routine.
758 */
759static void
760rt_msg_buffer(int type, struct rt_addrinfo *rtinfo, void *buf, int msglen)
761{
762 struct rt_msghdr *rtm;
763 char *cp;
764 int dlen, i;
765
766 rtm = (struct rt_msghdr *) buf;
767 rtm->rtm_version = RTM_VERSION;
768 rtm->rtm_type = type;
769 rtm->rtm_msglen = msglen;
770
771 cp = (char *)buf + rt_msghdrsize(type);
772 rtinfo->rti_addrs = 0;
984263bc 773 for (i = 0; i < RTAX_MAX; i++) {
590b8cd4
JH
774 struct sockaddr *sa;
775
984263bc
MD
776 if ((sa = rtinfo->rti_info[i]) == NULL)
777 continue;
778 rtinfo->rti_addrs |= (1 << i);
779 dlen = ROUNDUP(sa->sa_len);
590b8cd4
JH
780 bcopy(sa, cp, dlen);
781 cp += dlen;
984263bc 782 }
984263bc
MD
783}
784
590b8cd4
JH
785/*
786 * Build a routing message in a mbuf chain.
787 * Copy the addresses in the rtinfo->rti_info[] sockaddr array
788 * to the end of the mbuf after the message header.
789 *
790 * Set the rtinfo->rti_addrs bitmask of addresses present in rtinfo->rti_info[].
791 * This side-effect can be avoided if we reorder the addrs bitmask field in all
792 * the route messages to line up so we can set it here instead of back in the
793 * calling routine.
794 */
795static struct mbuf *
796rt_msg_mbuf(int type, struct rt_addrinfo *rtinfo)
984263bc 797{
590b8cd4
JH
798 struct mbuf *m;
799 struct rt_msghdr *rtm;
800 int hlen, len;
82ed7fc2 801 int i;
984263bc 802
590b8cd4
JH
803 hlen = rt_msghdrsize(type);
804 KASSERT(hlen <= MCLBYTES, ("rt_msg_mbuf: hlen %d doesn't fit", hlen));
984263bc 805
50503f0f 806 m = m_getl(hlen, MB_DONTWAIT, MT_DATA, M_PKTHDR, NULL);
590b8cd4
JH
807 if (m == NULL)
808 return (NULL);
590b8cd4
JH
809 m->m_pkthdr.len = m->m_len = hlen;
810 m->m_pkthdr.rcvif = NULL;
811 rtinfo->rti_addrs = 0;
812 len = hlen;
984263bc 813 for (i = 0; i < RTAX_MAX; i++) {
82ed7fc2 814 struct sockaddr *sa;
590b8cd4 815 int dlen;
984263bc 816
ef87f48d 817 if ((sa = rtinfo->rti_info[i]) == NULL)
984263bc
MD
818 continue;
819 rtinfo->rti_addrs |= (1 << i);
820 dlen = ROUNDUP(sa->sa_len);
590b8cd4 821 m_copyback(m, len, dlen, (caddr_t)sa); /* can grow mbuf chain */
984263bc
MD
822 len += dlen;
823 }
590b8cd4
JH
824 if (m->m_pkthdr.len != len) { /* one of the m_copyback() calls failed */
825 m_freem(m);
826 return (NULL);
984263bc 827 }
590b8cd4
JH
828 rtm = mtod(m, struct rt_msghdr *);
829 bzero(rtm, hlen);
830 rtm->rtm_msglen = len;
831 rtm->rtm_version = RTM_VERSION;
832 rtm->rtm_type = type;
833 return (m);
984263bc
MD
834}
835
836/*
837 * This routine is called to generate a message from the routing
f23061d4 838 * socket indicating that a redirect has occurred, a routing lookup
984263bc
MD
839 * has failed, or that a protocol has detected timeouts to a particular
840 * destination.
841 */
842void
f23061d4 843rt_missmsg(int type, struct rt_addrinfo *rtinfo, int flags, int error)
984263bc 844{
0c3c561c 845 struct sockaddr *dst = rtinfo->rti_info[RTAX_DST];
82ed7fc2
RG
846 struct rt_msghdr *rtm;
847 struct mbuf *m;
984263bc
MD
848
849 if (route_cb.any_count == 0)
850 return;
590b8cd4 851 m = rt_msg_mbuf(type, rtinfo);
ef87f48d 852 if (m == NULL)
984263bc
MD
853 return;
854 rtm = mtod(m, struct rt_msghdr *);
855 rtm->rtm_flags = RTF_DONE | flags;
856 rtm->rtm_errno = error;
857 rtm->rtm_addrs = rtinfo->rti_addrs;
590b8cd4 858 rts_input(m, familyof(dst));
984263bc
MD
859}
860
f3ed2586
JH
861void
862rt_dstmsg(int type, struct sockaddr *dst, int error)
863{
864 struct rt_msghdr *rtm;
865 struct rt_addrinfo addrs;
866 struct mbuf *m;
867
868 if (route_cb.any_count == 0)
869 return;
870 bzero(&addrs, sizeof(struct rt_addrinfo));
871 addrs.rti_info[RTAX_DST] = dst;
590b8cd4 872 m = rt_msg_mbuf(type, &addrs);
f3ed2586
JH
873 if (m == NULL)
874 return;
875 rtm = mtod(m, struct rt_msghdr *);
876 rtm->rtm_flags = RTF_DONE;
877 rtm->rtm_errno = error;
878 rtm->rtm_addrs = addrs.rti_addrs;
590b8cd4 879 rts_input(m, familyof(dst));
f3ed2586
JH
880}
881
984263bc
MD
882/*
883 * This routine is called to generate a message from the routing
884 * socket indicating that the status of a network interface has changed.
885 */
886void
0c3c561c 887rt_ifmsg(struct ifnet *ifp)
984263bc 888{
82ed7fc2 889 struct if_msghdr *ifm;
984263bc 890 struct mbuf *m;
590b8cd4 891 struct rt_addrinfo rtinfo;
984263bc
MD
892
893 if (route_cb.any_count == 0)
894 return;
590b8cd4
JH
895 bzero(&rtinfo, sizeof(struct rt_addrinfo));
896 m = rt_msg_mbuf(RTM_IFINFO, &rtinfo);
ef87f48d 897 if (m == NULL)
984263bc
MD
898 return;
899 ifm = mtod(m, struct if_msghdr *);
900 ifm->ifm_index = ifp->if_index;
9c095379 901 ifm->ifm_flags = ifp->if_flags;
984263bc 902 ifm->ifm_data = ifp->if_data;
590b8cd4
JH
903 ifm->ifm_addrs = 0;
904 rts_input(m, 0);
984263bc
MD
905}
906
372316d9
JH
907static void
908rt_ifamsg(int cmd, struct ifaddr *ifa)
909{
910 struct ifa_msghdr *ifam;
590b8cd4 911 struct rt_addrinfo rtinfo;
372316d9 912 struct mbuf *m;
372316d9
JH
913 struct ifnet *ifp = ifa->ifa_ifp;
914
590b8cd4
JH
915 bzero(&rtinfo, sizeof(struct rt_addrinfo));
916 rtinfo.rti_ifaaddr = ifa->ifa_addr;
917 rtinfo.rti_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr;
918 rtinfo.rti_netmask = ifa->ifa_netmask;
919 rtinfo.rti_bcastaddr = ifa->ifa_dstaddr;
f23061d4 920
590b8cd4 921 m = rt_msg_mbuf(cmd, &rtinfo);
f23061d4 922 if (m == NULL)
372316d9 923 return;
f23061d4 924
372316d9
JH
925 ifam = mtod(m, struct ifa_msghdr *);
926 ifam->ifam_index = ifp->if_index;
927 ifam->ifam_metric = ifa->ifa_metric;
928 ifam->ifam_flags = ifa->ifa_flags;
590b8cd4 929 ifam->ifam_addrs = rtinfo.rti_addrs;
f23061d4 930
590b8cd4 931 rts_input(m, familyof(ifa->ifa_addr));
372316d9
JH
932}
933
f3ed2586
JH
934void
935rt_rtmsg(int cmd, struct rtentry *rt, struct ifnet *ifp, int error)
372316d9
JH
936{
937 struct rt_msghdr *rtm;
590b8cd4 938 struct rt_addrinfo rtinfo;
372316d9 939 struct mbuf *m;
f3ed2586 940 struct sockaddr *dst;
372316d9
JH
941
942 if (rt == NULL)
943 return;
f23061d4 944
590b8cd4
JH
945 bzero(&rtinfo, sizeof(struct rt_addrinfo));
946 rtinfo.rti_dst = dst = rt_key(rt);
947 rtinfo.rti_gateway = rt->rt_gateway;
948 rtinfo.rti_netmask = rt_mask(rt);
f3ed2586 949 if (ifp != NULL)
590b8cd4
JH
950 rtinfo.rti_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr;
951 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr;
f23061d4 952
590b8cd4 953 m = rt_msg_mbuf(cmd, &rtinfo);
f23061d4 954 if (m == NULL)
372316d9 955 return;
f23061d4 956
372316d9 957 rtm = mtod(m, struct rt_msghdr *);
f3ed2586
JH
958 if (ifp != NULL)
959 rtm->rtm_index = ifp->if_index;
372316d9
JH
960 rtm->rtm_flags |= rt->rt_flags;
961 rtm->rtm_errno = error;
590b8cd4 962 rtm->rtm_addrs = rtinfo.rti_addrs;
f23061d4 963
590b8cd4 964 rts_input(m, familyof(dst));
372316d9
JH
965}
966
984263bc
MD
967/*
968 * This is called to generate messages from the routing socket
969 * indicating a network interface has had addresses associated with it.
970 * if we ever reverse the logic and replace messages TO the routing
971 * socket indicate a request to configure interfaces, then it will
972 * be unnecessary as the routing socket will automatically generate
973 * copies of it.
974 */
975void
0c3c561c 976rt_newaddrmsg(int cmd, struct ifaddr *ifa, int error, struct rtentry *rt)
984263bc 977{
78812139
EN
978#ifdef SCTP
979 /*
980 * notify the SCTP stack
981 * this will only get called when an address is added/deleted
982 * XXX pass the ifaddr struct instead if ifa->ifa_addr...
983 */
984 if (cmd == RTM_ADD)
985 sctp_add_ip_address(ifa);
986 else if (cmd == RTM_DELETE)
987 sctp_delete_ip_address(ifa);
988#endif /* SCTP */
989
984263bc
MD
990 if (route_cb.any_count == 0)
991 return;
984263bc 992
372316d9
JH
993 if (cmd == RTM_ADD) {
994 rt_ifamsg(RTM_NEWADDR, ifa);
f3ed2586 995 rt_rtmsg(RTM_ADD, rt, ifa->ifa_ifp, error);
372316d9
JH
996 } else {
997 KASSERT((cmd == RTM_DELETE), ("unknown cmd %d", cmd));
f3ed2586 998 rt_rtmsg(RTM_DELETE, rt, ifa->ifa_ifp, error);
372316d9 999 rt_ifamsg(RTM_DELADDR, ifa);
984263bc
MD
1000 }
1001}
1002
1003/*
1004 * This is the analogue to the rt_newaddrmsg which performs the same
1005 * function but for multicast group memberhips. This is easier since
1006 * there is no route state to worry about.
1007 */
1008void
0c3c561c 1009rt_newmaddrmsg(int cmd, struct ifmultiaddr *ifma)
984263bc 1010{
590b8cd4 1011 struct rt_addrinfo rtinfo;
ef87f48d 1012 struct mbuf *m = NULL;
984263bc
MD
1013 struct ifnet *ifp = ifma->ifma_ifp;
1014 struct ifma_msghdr *ifmam;
1015
1016 if (route_cb.any_count == 0)
1017 return;
1018
590b8cd4
JH
1019 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1020 rtinfo.rti_ifaaddr = ifma->ifma_addr;
1782116d 1021 if (ifp != NULL && !TAILQ_EMPTY(&ifp->if_addrhead))
590b8cd4 1022 rtinfo.rti_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr;
984263bc
MD
1023 /*
1024 * If a link-layer address is present, present it as a ``gateway''
1025 * (similarly to how ARP entries, e.g., are presented).
1026 */
590b8cd4 1027 rtinfo.rti_gateway = ifma->ifma_lladdr;
f23061d4 1028
590b8cd4 1029 m = rt_msg_mbuf(cmd, &rtinfo);
f23061d4 1030 if (m == NULL)
984263bc 1031 return;
f23061d4 1032
984263bc
MD
1033 ifmam = mtod(m, struct ifma_msghdr *);
1034 ifmam->ifmam_index = ifp->if_index;
590b8cd4 1035 ifmam->ifmam_addrs = rtinfo.rti_addrs;
f23061d4 1036
590b8cd4 1037 rts_input(m, familyof(ifma->ifma_addr));
984263bc
MD
1038}
1039
841ab66c
SZ
1040static struct mbuf *
1041rt_makeifannouncemsg(struct ifnet *ifp, int type, int what,
1042 struct rt_addrinfo *info)
984263bc 1043{
590b8cd4 1044 struct if_announcemsghdr *ifan;
841ab66c 1045 struct mbuf *m;
984263bc
MD
1046
1047 if (route_cb.any_count == 0)
841ab66c 1048 return NULL;
f23061d4 1049
841ab66c
SZ
1050 bzero(info, sizeof(*info));
1051 m = rt_msg_mbuf(type, info);
984263bc 1052 if (m == NULL)
841ab66c 1053 return NULL;
f23061d4 1054
984263bc
MD
1055 ifan = mtod(m, struct if_announcemsghdr *);
1056 ifan->ifan_index = ifp->if_index;
5fe66e68 1057 strlcpy(ifan->ifan_name, ifp->if_xname, sizeof ifan->ifan_name);
984263bc 1058 ifan->ifan_what = what;
841ab66c
SZ
1059 return m;
1060}
1061
1062/*
1063 * This is called to generate routing socket messages indicating
1064 * IEEE80211 wireless events.
1065 * XXX we piggyback on the RTM_IFANNOUNCE msg format in a clumsy way.
1066 */
1067void
1068rt_ieee80211msg(struct ifnet *ifp, int what, void *data, size_t data_len)
1069{
1070 struct rt_addrinfo info;
1071 struct mbuf *m;
f23061d4 1072
841ab66c
SZ
1073 m = rt_makeifannouncemsg(ifp, RTM_IEEE80211, what, &info);
1074 if (m == NULL)
1075 return;
1076
1077 /*
1078 * Append the ieee80211 data. Try to stick it in the
1079 * mbuf containing the ifannounce msg; otherwise allocate
1080 * a new mbuf and append.
1081 *
1082 * NB: we assume m is a single mbuf.
1083 */
1084 if (data_len > M_TRAILINGSPACE(m)) {
b11e1686 1085 struct mbuf *n = m_get(MB_DONTWAIT, MT_DATA);
841ab66c
SZ
1086 if (n == NULL) {
1087 m_freem(m);
1088 return;
1089 }
1090 bcopy(data, mtod(n, void *), data_len);
1091 n->m_len = data_len;
1092 m->m_next = n;
1093 } else if (data_len > 0) {
1094 bcopy(data, mtod(m, u_int8_t *) + m->m_len, data_len);
1095 m->m_len += data_len;
1096 }
1097 if (m->m_flags & M_PKTHDR)
1098 m->m_pkthdr.len += data_len;
1099 mtod(m, struct if_announcemsghdr *)->ifan_msglen += data_len;
590b8cd4 1100 rts_input(m, 0);
841ab66c
SZ
1101}
1102
1103/*
1104 * This is called to generate routing socket messages indicating
1105 * network interface arrival and departure.
1106 */
1107void
1108rt_ifannouncemsg(struct ifnet *ifp, int what)
1109{
1110 struct rt_addrinfo addrinfo;
1111 struct mbuf *m;
1112
1113 m = rt_makeifannouncemsg(ifp, RTM_IFANNOUNCE, what, &addrinfo);
1114 if (m != NULL)
1115 rts_input(m, 0);
590b8cd4 1116}
f23061d4 1117
590b8cd4
JH
1118static int
1119resizewalkarg(struct walkarg *w, int len)
1120{
1121 void *newptr;
1122
efda3bd0 1123 newptr = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
590b8cd4
JH
1124 if (newptr == NULL)
1125 return (ENOMEM);
1126 if (w->w_tmem != NULL)
efda3bd0 1127 kfree(w->w_tmem, M_RTABLE);
590b8cd4
JH
1128 w->w_tmem = newptr;
1129 w->w_tmemsize = len;
1130 return (0);
1131}
984263bc
MD
1132
1133/*
1134 * This is used in dumping the kernel table via sysctl().
1135 */
1136int
590b8cd4 1137sysctl_dumpentry(struct radix_node *rn, void *vw)
984263bc 1138{
82ed7fc2
RG
1139 struct walkarg *w = vw;
1140 struct rtentry *rt = (struct rtentry *)rn;
590b8cd4
JH
1141 struct rt_addrinfo rtinfo;
1142 int error, msglen;
984263bc
MD
1143
1144 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg))
1145 return 0;
f23061d4 1146
590b8cd4
JH
1147 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1148 rtinfo.rti_dst = rt_key(rt);
1149 rtinfo.rti_gateway = rt->rt_gateway;
1150 rtinfo.rti_netmask = rt_mask(rt);
1151 rtinfo.rti_genmask = rt->rt_genmask;
ef87f48d 1152 if (rt->rt_ifp != NULL) {
590b8cd4 1153 rtinfo.rti_ifpaddr =
ef87f48d 1154 TAILQ_FIRST(&rt->rt_ifp->if_addrhead)->ifa_addr;
590b8cd4 1155 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr;
984263bc 1156 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT)
590b8cd4 1157 rtinfo.rti_bcastaddr = rt->rt_ifa->ifa_dstaddr;
984263bc 1158 }
590b8cd4
JH
1159 msglen = rt_msgsize(RTM_GET, &rtinfo);
1160 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0)
1161 return (ENOMEM);
1162 rt_msg_buffer(RTM_GET, &rtinfo, w->w_tmem, msglen);
1163 if (w->w_req != NULL) {
8e63efec 1164 struct rt_msghdr *rtm = w->w_tmem;
984263bc
MD
1165
1166 rtm->rtm_flags = rt->rt_flags;
1167 rtm->rtm_use = rt->rt_use;
1168 rtm->rtm_rmx = rt->rt_rmx;
1169 rtm->rtm_index = rt->rt_ifp->if_index;
1170 rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0;
590b8cd4
JH
1171 rtm->rtm_addrs = rtinfo.rti_addrs;
1172 error = SYSCTL_OUT(w->w_req, rtm, msglen);
984263bc
MD
1173 return (error);
1174 }
590b8cd4 1175 return (0);
984263bc
MD
1176}
1177
590b8cd4
JH
1178static int
1179sysctl_iflist(int af, struct walkarg *w)
984263bc 1180{
82ed7fc2
RG
1181 struct ifnet *ifp;
1182 struct ifaddr *ifa;
590b8cd4
JH
1183 struct rt_addrinfo rtinfo;
1184 int msglen, error;
984263bc 1185
590b8cd4 1186 bzero(&rtinfo, sizeof(struct rt_addrinfo));
984263bc
MD
1187 TAILQ_FOREACH(ifp, &ifnet, if_link) {
1188 if (w->w_arg && w->w_arg != ifp->if_index)
1189 continue;
1190 ifa = TAILQ_FIRST(&ifp->if_addrhead);
590b8cd4
JH
1191 rtinfo.rti_ifpaddr = ifa->ifa_addr;
1192 msglen = rt_msgsize(RTM_IFINFO, &rtinfo);
1193 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0)
1194 return (ENOMEM);
1195 rt_msg_buffer(RTM_IFINFO, &rtinfo, w->w_tmem, msglen);
1196 rtinfo.rti_ifpaddr = NULL;
ef87f48d 1197 if (w->w_req != NULL && w->w_tmem != NULL) {
8e63efec 1198 struct if_msghdr *ifm = w->w_tmem;
984263bc 1199
984263bc 1200 ifm->ifm_index = ifp->if_index;
9c095379 1201 ifm->ifm_flags = ifp->if_flags;
984263bc 1202 ifm->ifm_data = ifp->if_data;
590b8cd4
JH
1203 ifm->ifm_addrs = rtinfo.rti_addrs;
1204 error = SYSCTL_OUT(w->w_req, ifm, msglen);
984263bc
MD
1205 if (error)
1206 return (error);
1207 }
ef87f48d 1208 while ((ifa = TAILQ_NEXT(ifa, ifa_link)) != NULL) {
984263bc
MD
1209 if (af && af != ifa->ifa_addr->sa_family)
1210 continue;
590b8cd4 1211 if (curproc->p_ucred->cr_prison &&
87de5057 1212 prison_if(curproc->p_ucred, ifa->ifa_addr))
984263bc 1213 continue;
590b8cd4
JH
1214 rtinfo.rti_ifaaddr = ifa->ifa_addr;
1215 rtinfo.rti_netmask = ifa->ifa_netmask;
1216 rtinfo.rti_bcastaddr = ifa->ifa_dstaddr;
1217 msglen = rt_msgsize(RTM_NEWADDR, &rtinfo);
1218 if (w->w_tmemsize < msglen &&
1219 resizewalkarg(w, msglen) != 0)
1220 return (ENOMEM);
1221 rt_msg_buffer(RTM_NEWADDR, &rtinfo, w->w_tmem, msglen);
1222 if (w->w_req != NULL) {
8e63efec 1223 struct ifa_msghdr *ifam = w->w_tmem;
984263bc 1224
984263bc
MD
1225 ifam->ifam_index = ifa->ifa_ifp->if_index;
1226 ifam->ifam_flags = ifa->ifa_flags;
1227 ifam->ifam_metric = ifa->ifa_metric;
590b8cd4
JH
1228 ifam->ifam_addrs = rtinfo.rti_addrs;
1229 error = SYSCTL_OUT(w->w_req, w->w_tmem, msglen);
984263bc
MD
1230 if (error)
1231 return (error);
1232 }
1233 }
590b8cd4
JH
1234 rtinfo.rti_netmask = NULL;
1235 rtinfo.rti_ifaaddr = NULL;
1236 rtinfo.rti_bcastaddr = NULL;
984263bc
MD
1237 }
1238 return (0);
1239}
1240
1241static int
1242sysctl_rtsock(SYSCTL_HANDLER_ARGS)
1243{
1244 int *name = (int *)arg1;
1245 u_int namelen = arg2;
82ed7fc2 1246 struct radix_node_head *rnh;
a49aa710 1247 int i, error = EINVAL;
ecdefdda 1248 int origcpu;
984263bc
MD
1249 u_char af;
1250 struct walkarg w;
1251
1252 name ++;
1253 namelen--;
1254 if (req->newptr)
1255 return (EPERM);
ecdefdda 1256 if (namelen != 3 && namelen != 4)
984263bc
MD
1257 return (EINVAL);
1258 af = name[0];
5fe66e68 1259 bzero(&w, sizeof w);
984263bc
MD
1260 w.w_op = name[1];
1261 w.w_arg = name[2];
1262 w.w_req = req;
1263
ecdefdda
MD
1264 /*
1265 * Optional third argument specifies cpu, used primarily for
1266 * debugging the route table.
1267 */
1268 if (namelen == 4) {
1269 if (name[3] < 0 || name[3] >= ncpus)
1270 return (EINVAL);
1271 origcpu = mycpuid;
1272 lwkt_migratecpu(name[3]);
1273 } else {
1274 origcpu = -1;
1275 }
a49aa710 1276 crit_enter();
984263bc 1277 switch (w.w_op) {
984263bc
MD
1278 case NET_RT_DUMP:
1279 case NET_RT_FLAGS:
1280 for (i = 1; i <= AF_MAX; i++)
ecdefdda
MD
1281 if ((rnh = rt_tables[mycpuid][i]) &&
1282 (af == 0 || af == i) &&
984263bc 1283 (error = rnh->rnh_walktree(rnh,
f23061d4 1284 sysctl_dumpentry, &w)))
984263bc
MD
1285 break;
1286 break;
1287
1288 case NET_RT_IFLIST:
1289 error = sysctl_iflist(af, &w);
1290 }
a49aa710 1291 crit_exit();
590b8cd4 1292 if (w.w_tmem != NULL)
efda3bd0 1293 kfree(w.w_tmem, M_RTABLE);
ecdefdda
MD
1294 if (origcpu >= 0)
1295 lwkt_migratecpu(origcpu);
984263bc
MD
1296 return (error);
1297}
1298
1299SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, "");
1300
1301/*
1302 * Definitions of protocols supported in the ROUTE domain.
1303 */
1304
ba4257b8 1305static struct domain routedomain; /* or at least forward */
984263bc
MD
1306
1307static struct protosw routesw[] = {
1308{ SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR,
1309 0, route_output, raw_ctlinput, 0,
9eeaa8a9 1310 cpu0_soport,
984263bc
MD
1311 raw_init, 0, 0, 0,
1312 &route_usrreqs
1313}
1314};
1315
ba4257b8 1316static struct domain routedomain = {
9c70fe43
JH
1317 PF_ROUTE, "route", NULL, NULL, NULL,
1318 routesw, &routesw[(sizeof routesw)/(sizeof routesw[0])],
1319};
984263bc
MD
1320
1321DOMAIN_SET(route);