2 * Copyright (c) 2004, 2005 The DragonFly Project. All rights reserved.
4 * This code is derived from software contributed to The DragonFly Project
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
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.
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
34 * Copyright (c) 1988, 1991, 1993
35 * The Regents of the University of California. All rights reserved.
37 * Redistribution and use in source and binary forms, with or without
38 * modification, are permitted provided that the following conditions
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.
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
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 $
67 * $DragonFly: src/sys/net/rtsock.c,v 1.37 2007/03/04 18:51:59 swildner Exp $
72 #include <sys/param.h>
73 #include <sys/systm.h>
74 #include <sys/kernel.h>
75 #include <sys/sysctl.h>
77 #include <sys/malloc.h>
79 #include <sys/protosw.h>
80 #include <sys/socket.h>
81 #include <sys/socketvar.h>
82 #include <sys/domain.h>
83 #include <sys/thread2.h>
86 #include <net/route.h>
87 #include <net/raw_cb.h>
90 extern void sctp_add_ip_address(struct ifaddr *ifa);
91 extern void sctp_delete_ip_address(struct ifaddr *ifa);
94 MALLOC_DEFINE(M_RTABLE, "routetbl", "routing tables");
96 static struct route_cb {
104 static const struct sockaddr route_src = { 2, PF_ROUTE, };
110 struct sysctl_req *w_req;
114 rt_msg_mbuf (int, struct rt_addrinfo *);
115 static void rt_msg_buffer (int, struct rt_addrinfo *, void *buf, int len);
116 static int rt_msgsize (int type, struct rt_addrinfo *rtinfo);
117 static int rt_xaddrs (char *, char *, struct rt_addrinfo *);
118 static int sysctl_dumpentry (struct radix_node *rn, void *vw);
119 static int sysctl_iflist (int af, struct walkarg *w);
120 static int route_output(struct mbuf *, struct socket *, ...);
121 static void rt_setmetrics (u_long, struct rt_metrics *,
122 struct rt_metrics *);
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
129 rts_abort(struct socket *so)
134 error = raw_usrreqs.pru_abort(so);
139 /* pru_accept is EOPNOTSUPP */
142 rts_attach(struct socket *so, int proto, struct pru_attach_info *ai)
147 if (sotorawcb(so) != NULL)
148 return EISCONN; /* XXX panic? */
150 rp = kmalloc(sizeof *rp, M_PCB, M_WAITOK | M_ZERO);
155 * The critical section is necessary to block protocols from sending
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
159 * eliminate the critical section.
163 error = raw_attach(so, proto, ai->sb_rlimit);
170 switch(rp->rcb_proto.sp_protocol) {
175 route_cb.ip6_count++;
178 route_cb.ipx_count++;
184 rp->rcb_faddr = &route_src;
185 route_cb.any_count++;
187 so->so_options |= SO_USELOOPBACK;
193 rts_bind(struct socket *so, struct sockaddr *nam, struct thread *td)
198 error = raw_usrreqs.pru_bind(so, nam, td); /* xxx just EINVAL */
204 rts_connect(struct socket *so, struct sockaddr *nam, struct thread *td)
209 error = raw_usrreqs.pru_connect(so, nam, td); /* XXX just EINVAL */
214 /* pru_connect2 is EOPNOTSUPP */
215 /* pru_control is EOPNOTSUPP */
218 rts_detach(struct socket *so)
220 struct rawcb *rp = sotorawcb(so);
225 switch(rp->rcb_proto.sp_protocol) {
230 route_cb.ip6_count--;
233 route_cb.ipx_count--;
239 route_cb.any_count--;
241 error = raw_usrreqs.pru_detach(so);
247 rts_disconnect(struct socket *so)
252 error = raw_usrreqs.pru_disconnect(so);
257 /* pru_listen is EOPNOTSUPP */
260 rts_peeraddr(struct socket *so, struct sockaddr **nam)
265 error = raw_usrreqs.pru_peeraddr(so, nam);
270 /* pru_rcvd is EOPNOTSUPP */
271 /* pru_rcvoob is EOPNOTSUPP */
274 rts_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam,
275 struct mbuf *control, struct thread *td)
280 error = raw_usrreqs.pru_send(so, flags, m, nam, control, td);
285 /* pru_sense is null */
288 rts_shutdown(struct socket *so)
293 error = raw_usrreqs.pru_shutdown(so);
299 rts_sockaddr(struct socket *so, struct sockaddr **nam)
304 error = raw_usrreqs.pru_sockaddr(so, nam);
309 static 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
317 static __inline sa_family_t
318 familyof(struct sockaddr *sa)
320 return (sa != NULL ? sa->sa_family : 0);
324 rts_input(struct mbuf *m, sa_family_t family)
326 static const struct sockaddr route_dst = { 2, PF_ROUTE, };
327 struct sockproto route_proto = { PF_ROUTE, family };
329 raw_input(m, &route_proto, &route_src, &route_dst);
333 reallocbuf(void *ptr, size_t len, size_t olen)
337 newptr = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
340 bcopy(ptr, newptr, olen);
341 kfree(ptr, M_RTABLE);
346 * Internal helper routine for route_output().
349 fillrtmsg(struct rt_msghdr **prtm, struct rtentry *rt,
350 struct rt_addrinfo *rtinfo)
353 struct rt_msghdr *rtm = *prtm;
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;
369 rtinfo->rti_ifpaddr = NULL;
370 rtinfo->rti_ifaaddr = NULL;
374 msglen = rt_msgsize(rtm->rtm_type, rtinfo);
375 if (rtm->rtm_msglen < msglen) {
376 rtm = reallocbuf(rtm, msglen, rtm->rtm_msglen);
381 rt_msg_buffer(rtm->rtm_type, rtinfo, rtm, msglen);
383 rtm->rtm_flags = rt->rt_flags;
384 rtm->rtm_rmx = rt->rt_rmx;
385 rtm->rtm_addrs = rtinfo->rti_addrs;
390 static void route_output_add_callback(int, int, struct rt_addrinfo *,
391 struct rtentry *, void *);
392 static void route_output_delete_callback(int, int, struct rt_addrinfo *,
393 struct rtentry *, void *);
394 static void route_output_change_callback(int, int, struct rt_addrinfo *,
395 struct rtentry *, void *);
396 static void route_output_lock_callback(int, int, struct rt_addrinfo *,
397 struct rtentry *, void *);
401 route_output(struct mbuf *m, struct socket *so, ...)
403 struct rt_msghdr *rtm = NULL;
405 struct radix_node_head *rnh;
406 struct rawcb *rp = NULL;
407 struct pr_output_info *oi;
408 struct rt_addrinfo rtinfo;
413 oi = __va_arg(ap, struct pr_output_info *);
416 #define gotoerr(e) { error = e; goto flush;}
419 (m->m_len < sizeof(long) &&
420 (m = m_pullup(m, sizeof(long))) == NULL))
422 if (!(m->m_flags & M_PKTHDR))
423 panic("route_output");
424 len = m->m_pkthdr.len;
425 if (len < sizeof(struct rt_msghdr) ||
426 len != mtod(m, struct rt_msghdr *)->rtm_msglen) {
427 rtinfo.rti_dst = NULL;
430 rtm = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
432 rtinfo.rti_dst = NULL;
435 m_copydata(m, 0, len, (caddr_t)rtm);
436 if (rtm->rtm_version != RTM_VERSION) {
437 rtinfo.rti_dst = NULL;
438 gotoerr(EPROTONOSUPPORT);
440 rtm->rtm_pid = oi->p_pid;
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;
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))
452 if (rtinfo.rti_genmask != NULL) {
453 struct radix_node *n;
455 #define clen(s) (*(u_char *)(s))
456 n = rn_addmask((char *)rtinfo.rti_genmask, TRUE, 1);
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;
467 * Verify that the caller has the appropriate privilege; RTM_GET
468 * is the only operation the non-superuser is allowed.
470 if (rtm->rtm_type != RTM_GET && suser_cred(so->so_cred, 0) != 0)
473 switch (rtm->rtm_type) {
475 if (rtinfo.rti_gateway == NULL) {
478 error = rtrequest1_global(RTM_ADD, &rtinfo,
479 route_output_add_callback, rtm);
484 * note: &rtm passed as argument so 'rtm' can be replaced.
486 error = rtrequest1_global(RTM_DELETE, &rtinfo,
487 route_output_delete_callback, &rtm);
490 rnh = rt_tables[mycpuid][rtinfo.rti_dst->sa_family];
492 error = EAFNOSUPPORT;
495 rt = (struct rtentry *)
496 rnh->rnh_lookup((char *)rtinfo.rti_dst,
497 (char *)rtinfo.rti_netmask, rnh);
503 if (fillrtmsg(&rtm, rt, &rtinfo) != 0)
508 error = rtrequest1_global(RTM_GET, &rtinfo,
509 route_output_change_callback, rtm);
512 error = rtrequest1_global(RTM_GET, &rtinfo,
513 route_output_lock_callback, rtm);
523 rtm->rtm_errno = error;
525 rtm->rtm_flags |= RTF_DONE;
529 * Check to see if we don't want our own messages.
531 if (!(so->so_options & SO_USELOOPBACK)) {
532 if (route_cb.any_count <= 1) {
534 kfree(rtm, M_RTABLE);
538 /* There is another listener, so construct message */
542 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm);
543 if (m->m_pkthdr.len < rtm->rtm_msglen) {
546 } else if (m->m_pkthdr.len > rtm->rtm_msglen)
547 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len);
548 kfree(rtm, M_RTABLE);
551 rp->rcb_proto.sp_family = 0; /* Avoid us */
553 rts_input(m, familyof(rtinfo.rti_dst));
555 rp->rcb_proto.sp_family = PF_ROUTE;
560 route_output_add_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
561 struct rtentry *rt, void *arg)
563 struct rt_msghdr *rtm = arg;
565 if (error == 0 && rt != NULL) {
566 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_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;
576 route_output_delete_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
577 struct rtentry *rt, void *arg)
579 struct rt_msghdr **rtm = arg;
581 if (error == 0 && rt) {
583 if (fillrtmsg(rtm, rt, rtinfo) != 0) {
585 /* XXX no way to return the error */
592 route_output_change_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
593 struct rtentry *rt, void *arg)
595 struct rt_msghdr *rtm = arg;
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
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))
610 error = rt_getifa(rtinfo);
614 if (rtinfo->rti_gateway != NULL) {
615 error = rt_setgate(rt, rt_key(rt), rtinfo->rti_gateway);
619 if ((ifa = rtinfo->rti_ifa) != NULL) {
620 struct ifaddr *oifa = rt->rt_ifa;
623 if (oifa && oifa->ifa_rtrequest)
624 oifa->ifa_rtrequest(RTM_DELETE, rt, rtinfo);
628 rt->rt_ifp = rtinfo->rti_ifp;
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;
637 /* XXX no way to return error */
642 route_output_lock_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
643 struct rtentry *rt, void *arg)
645 struct rt_msghdr *rtm = arg;
647 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits);
648 rt->rt_rmx.rmx_locks |=
649 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks);
653 rt_setmetrics(u_long which, struct rt_metrics *in, struct rt_metrics *out)
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);
668 ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long))
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.
676 rt_xaddrs(char *cp, char *cplim, struct rt_addrinfo *rtinfo)
681 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) {
682 if ((rtinfo->rti_addrs & (1 << i)) == 0)
684 sa = (struct sockaddr *)cp;
688 if ((cp + sa->sa_len) > cplim) {
693 * There are no more... Quit now.
694 * If there are more bits, they are in error.
695 * I've seen this. route(1) can evidently generate these.
696 * This causes kernel to core dump.
697 * For compatibility, if we see this, point to a safe address.
699 if (sa->sa_len == 0) {
700 static struct sockaddr sa_zero = {
701 sizeof sa_zero, AF_INET,
704 rtinfo->rti_info[i] = &sa_zero;
705 return (0); /* should be EINVAL but for compat */
708 /* Accept the sockaddr. */
709 rtinfo->rti_info[i] = sa;
710 cp += ROUNDUP(sa->sa_len);
716 rt_msghdrsize(int type)
721 return sizeof(struct ifa_msghdr);
724 return sizeof(struct ifma_msghdr);
726 return sizeof(struct if_msghdr);
729 return sizeof(struct if_announcemsghdr);
731 return sizeof(struct rt_msghdr);
736 rt_msgsize(int type, struct rt_addrinfo *rtinfo)
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);
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.
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
760 rt_msg_buffer(int type, struct rt_addrinfo *rtinfo, void *buf, int msglen)
762 struct rt_msghdr *rtm;
766 rtm = (struct rt_msghdr *) buf;
767 rtm->rtm_version = RTM_VERSION;
768 rtm->rtm_type = type;
769 rtm->rtm_msglen = msglen;
771 cp = (char *)buf + rt_msghdrsize(type);
772 rtinfo->rti_addrs = 0;
773 for (i = 0; i < RTAX_MAX; i++) {
776 if ((sa = rtinfo->rti_info[i]) == NULL)
778 rtinfo->rti_addrs |= (1 << i);
779 dlen = ROUNDUP(sa->sa_len);
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.
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
796 rt_msg_mbuf(int type, struct rt_addrinfo *rtinfo)
799 struct rt_msghdr *rtm;
803 hlen = rt_msghdrsize(type);
804 KASSERT(hlen <= MCLBYTES, ("rt_msg_mbuf: hlen %d doesn't fit", hlen));
806 m = m_getl(hlen, MB_DONTWAIT, MT_DATA, M_PKTHDR, NULL);
809 m->m_pkthdr.len = m->m_len = hlen;
810 m->m_pkthdr.rcvif = NULL;
811 rtinfo->rti_addrs = 0;
813 for (i = 0; i < RTAX_MAX; i++) {
817 if ((sa = rtinfo->rti_info[i]) == NULL)
819 rtinfo->rti_addrs |= (1 << i);
820 dlen = ROUNDUP(sa->sa_len);
821 m_copyback(m, len, dlen, (caddr_t)sa); /* can grow mbuf chain */
824 if (m->m_pkthdr.len != len) { /* one of the m_copyback() calls failed */
828 rtm = mtod(m, struct rt_msghdr *);
830 rtm->rtm_msglen = len;
831 rtm->rtm_version = RTM_VERSION;
832 rtm->rtm_type = type;
837 * This routine is called to generate a message from the routing
838 * socket indicating that a redirect has occurred, a routing lookup
839 * has failed, or that a protocol has detected timeouts to a particular
843 rt_missmsg(int type, struct rt_addrinfo *rtinfo, int flags, int error)
845 struct sockaddr *dst = rtinfo->rti_info[RTAX_DST];
846 struct rt_msghdr *rtm;
849 if (route_cb.any_count == 0)
851 m = rt_msg_mbuf(type, rtinfo);
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;
858 rts_input(m, familyof(dst));
862 rt_dstmsg(int type, struct sockaddr *dst, int error)
864 struct rt_msghdr *rtm;
865 struct rt_addrinfo addrs;
868 if (route_cb.any_count == 0)
870 bzero(&addrs, sizeof(struct rt_addrinfo));
871 addrs.rti_info[RTAX_DST] = dst;
872 m = rt_msg_mbuf(type, &addrs);
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;
879 rts_input(m, familyof(dst));
883 * This routine is called to generate a message from the routing
884 * socket indicating that the status of a network interface has changed.
887 rt_ifmsg(struct ifnet *ifp)
889 struct if_msghdr *ifm;
891 struct rt_addrinfo rtinfo;
893 if (route_cb.any_count == 0)
895 bzero(&rtinfo, sizeof(struct rt_addrinfo));
896 m = rt_msg_mbuf(RTM_IFINFO, &rtinfo);
899 ifm = mtod(m, struct if_msghdr *);
900 ifm->ifm_index = ifp->if_index;
901 ifm->ifm_flags = ifp->if_flags;
902 ifm->ifm_data = ifp->if_data;
908 rt_ifamsg(int cmd, struct ifaddr *ifa)
910 struct ifa_msghdr *ifam;
911 struct rt_addrinfo rtinfo;
913 struct ifnet *ifp = ifa->ifa_ifp;
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;
921 m = rt_msg_mbuf(cmd, &rtinfo);
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;
929 ifam->ifam_addrs = rtinfo.rti_addrs;
931 rts_input(m, familyof(ifa->ifa_addr));
935 rt_rtmsg(int cmd, struct rtentry *rt, struct ifnet *ifp, int error)
937 struct rt_msghdr *rtm;
938 struct rt_addrinfo rtinfo;
940 struct sockaddr *dst;
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);
950 rtinfo.rti_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr;
951 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr;
953 m = rt_msg_mbuf(cmd, &rtinfo);
957 rtm = mtod(m, struct rt_msghdr *);
959 rtm->rtm_index = ifp->if_index;
960 rtm->rtm_flags |= rt->rt_flags;
961 rtm->rtm_errno = error;
962 rtm->rtm_addrs = rtinfo.rti_addrs;
964 rts_input(m, familyof(dst));
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
976 rt_newaddrmsg(int cmd, struct ifaddr *ifa, int error, struct rtentry *rt)
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...
985 sctp_add_ip_address(ifa);
986 else if (cmd == RTM_DELETE)
987 sctp_delete_ip_address(ifa);
990 if (route_cb.any_count == 0)
993 if (cmd == RTM_ADD) {
994 rt_ifamsg(RTM_NEWADDR, ifa);
995 rt_rtmsg(RTM_ADD, rt, ifa->ifa_ifp, error);
997 KASSERT((cmd == RTM_DELETE), ("unknown cmd %d", cmd));
998 rt_rtmsg(RTM_DELETE, rt, ifa->ifa_ifp, error);
999 rt_ifamsg(RTM_DELADDR, ifa);
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.
1009 rt_newmaddrmsg(int cmd, struct ifmultiaddr *ifma)
1011 struct rt_addrinfo rtinfo;
1012 struct mbuf *m = NULL;
1013 struct ifnet *ifp = ifma->ifma_ifp;
1014 struct ifma_msghdr *ifmam;
1016 if (route_cb.any_count == 0)
1019 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1020 rtinfo.rti_ifaaddr = ifma->ifma_addr;
1021 if (ifp != NULL && !TAILQ_EMPTY(&ifp->if_addrhead))
1022 rtinfo.rti_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr;
1024 * If a link-layer address is present, present it as a ``gateway''
1025 * (similarly to how ARP entries, e.g., are presented).
1027 rtinfo.rti_gateway = ifma->ifma_lladdr;
1029 m = rt_msg_mbuf(cmd, &rtinfo);
1033 ifmam = mtod(m, struct ifma_msghdr *);
1034 ifmam->ifmam_index = ifp->if_index;
1035 ifmam->ifmam_addrs = rtinfo.rti_addrs;
1037 rts_input(m, familyof(ifma->ifma_addr));
1040 static struct mbuf *
1041 rt_makeifannouncemsg(struct ifnet *ifp, int type, int what,
1042 struct rt_addrinfo *info)
1044 struct if_announcemsghdr *ifan;
1047 if (route_cb.any_count == 0)
1050 bzero(info, sizeof(*info));
1051 m = rt_msg_mbuf(type, info);
1055 ifan = mtod(m, struct if_announcemsghdr *);
1056 ifan->ifan_index = ifp->if_index;
1057 strlcpy(ifan->ifan_name, ifp->if_xname, sizeof ifan->ifan_name);
1058 ifan->ifan_what = what;
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.
1068 rt_ieee80211msg(struct ifnet *ifp, int what, void *data, size_t data_len)
1070 struct rt_addrinfo info;
1073 m = rt_makeifannouncemsg(ifp, RTM_IEEE80211, what, &info);
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.
1082 * NB: we assume m is a single mbuf.
1084 if (data_len > M_TRAILINGSPACE(m)) {
1085 struct mbuf *n = m_get(MB_DONTWAIT, MT_DATA);
1090 bcopy(data, mtod(n, void *), data_len);
1091 n->m_len = data_len;
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;
1097 if (m->m_flags & M_PKTHDR)
1098 m->m_pkthdr.len += data_len;
1099 mtod(m, struct if_announcemsghdr *)->ifan_msglen += data_len;
1104 * This is called to generate routing socket messages indicating
1105 * network interface arrival and departure.
1108 rt_ifannouncemsg(struct ifnet *ifp, int what)
1110 struct rt_addrinfo addrinfo;
1113 m = rt_makeifannouncemsg(ifp, RTM_IFANNOUNCE, what, &addrinfo);
1119 resizewalkarg(struct walkarg *w, int len)
1123 newptr = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
1126 if (w->w_tmem != NULL)
1127 kfree(w->w_tmem, M_RTABLE);
1129 w->w_tmemsize = len;
1134 * This is used in dumping the kernel table via sysctl().
1137 sysctl_dumpentry(struct radix_node *rn, void *vw)
1139 struct walkarg *w = vw;
1140 struct rtentry *rt = (struct rtentry *)rn;
1141 struct rt_addrinfo rtinfo;
1144 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg))
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;
1152 if (rt->rt_ifp != NULL) {
1153 rtinfo.rti_ifpaddr =
1154 TAILQ_FIRST(&rt->rt_ifp->if_addrhead)->ifa_addr;
1155 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr;
1156 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT)
1157 rtinfo.rti_bcastaddr = rt->rt_ifa->ifa_dstaddr;
1159 msglen = rt_msgsize(RTM_GET, &rtinfo);
1160 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0)
1162 rt_msg_buffer(RTM_GET, &rtinfo, w->w_tmem, msglen);
1163 if (w->w_req != NULL) {
1164 struct rt_msghdr *rtm = w->w_tmem;
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;
1171 rtm->rtm_addrs = rtinfo.rti_addrs;
1172 error = SYSCTL_OUT(w->w_req, rtm, msglen);
1179 sysctl_iflist(int af, struct walkarg *w)
1183 struct rt_addrinfo rtinfo;
1186 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1187 TAILQ_FOREACH(ifp, &ifnet, if_link) {
1188 if (w->w_arg && w->w_arg != ifp->if_index)
1190 ifa = TAILQ_FIRST(&ifp->if_addrhead);
1191 rtinfo.rti_ifpaddr = ifa->ifa_addr;
1192 msglen = rt_msgsize(RTM_IFINFO, &rtinfo);
1193 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0)
1195 rt_msg_buffer(RTM_IFINFO, &rtinfo, w->w_tmem, msglen);
1196 rtinfo.rti_ifpaddr = NULL;
1197 if (w->w_req != NULL && w->w_tmem != NULL) {
1198 struct if_msghdr *ifm = w->w_tmem;
1200 ifm->ifm_index = ifp->if_index;
1201 ifm->ifm_flags = ifp->if_flags;
1202 ifm->ifm_data = ifp->if_data;
1203 ifm->ifm_addrs = rtinfo.rti_addrs;
1204 error = SYSCTL_OUT(w->w_req, ifm, msglen);
1208 while ((ifa = TAILQ_NEXT(ifa, ifa_link)) != NULL) {
1209 if (af && af != ifa->ifa_addr->sa_family)
1211 if (curproc->p_ucred->cr_prison &&
1212 prison_if(curproc->p_ucred, ifa->ifa_addr))
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)
1221 rt_msg_buffer(RTM_NEWADDR, &rtinfo, w->w_tmem, msglen);
1222 if (w->w_req != NULL) {
1223 struct ifa_msghdr *ifam = w->w_tmem;
1225 ifam->ifam_index = ifa->ifa_ifp->if_index;
1226 ifam->ifam_flags = ifa->ifa_flags;
1227 ifam->ifam_metric = ifa->ifa_metric;
1228 ifam->ifam_addrs = rtinfo.rti_addrs;
1229 error = SYSCTL_OUT(w->w_req, w->w_tmem, msglen);
1234 rtinfo.rti_netmask = NULL;
1235 rtinfo.rti_ifaaddr = NULL;
1236 rtinfo.rti_bcastaddr = NULL;
1242 sysctl_rtsock(SYSCTL_HANDLER_ARGS)
1244 int *name = (int *)arg1;
1245 u_int namelen = arg2;
1246 struct radix_node_head *rnh;
1247 int i, error = EINVAL;
1256 if (namelen != 3 && namelen != 4)
1259 bzero(&w, sizeof w);
1265 * Optional third argument specifies cpu, used primarily for
1266 * debugging the route table.
1269 if (name[3] < 0 || name[3] >= ncpus)
1272 lwkt_migratecpu(name[3]);
1280 for (i = 1; i <= AF_MAX; i++)
1281 if ((rnh = rt_tables[mycpuid][i]) &&
1282 (af == 0 || af == i) &&
1283 (error = rnh->rnh_walktree(rnh,
1284 sysctl_dumpentry, &w)))
1289 error = sysctl_iflist(af, &w);
1292 if (w.w_tmem != NULL)
1293 kfree(w.w_tmem, M_RTABLE);
1295 lwkt_migratecpu(origcpu);
1299 SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, "");
1302 * Definitions of protocols supported in the ROUTE domain.
1305 static struct domain routedomain; /* or at least forward */
1307 static struct protosw routesw[] = {
1308 { SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR,
1309 0, route_output, raw_ctlinput, 0,
1316 static struct domain routedomain = {
1317 PF_ROUTE, "route", NULL, NULL, NULL,
1318 routesw, &routesw[(sizeof routesw)/(sizeof routesw[0])],