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 1.38 2007/04/21 02:26:47 dillon 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 .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,
332 static __inline sa_family_t
333 familyof(struct sockaddr *sa)
335 return (sa != NULL ? sa->sa_family : 0);
339 rts_input(struct mbuf *m, sa_family_t family)
341 static const struct sockaddr route_dst = { 2, PF_ROUTE, };
342 struct sockproto route_proto = { PF_ROUTE, family };
344 raw_input(m, &route_proto, &route_src, &route_dst);
348 reallocbuf(void *ptr, size_t len, size_t olen)
352 newptr = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
355 bcopy(ptr, newptr, olen);
356 kfree(ptr, M_RTABLE);
361 * Internal helper routine for route_output().
364 fillrtmsg(struct rt_msghdr **prtm, struct rtentry *rt,
365 struct rt_addrinfo *rtinfo)
368 struct rt_msghdr *rtm = *prtm;
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;
384 rtinfo->rti_ifpaddr = NULL;
385 rtinfo->rti_ifaaddr = NULL;
389 msglen = rt_msgsize(rtm->rtm_type, rtinfo);
390 if (rtm->rtm_msglen < msglen) {
391 rtm = reallocbuf(rtm, msglen, rtm->rtm_msglen);
396 rt_msg_buffer(rtm->rtm_type, rtinfo, rtm, msglen);
398 rtm->rtm_flags = rt->rt_flags;
399 rtm->rtm_rmx = rt->rt_rmx;
400 rtm->rtm_addrs = rtinfo->rti_addrs;
405 static void route_output_add_callback(int, int, struct rt_addrinfo *,
406 struct rtentry *, void *);
407 static void route_output_delete_callback(int, int, struct rt_addrinfo *,
408 struct rtentry *, void *);
409 static void route_output_change_callback(int, int, struct rt_addrinfo *,
410 struct rtentry *, void *);
411 static void route_output_lock_callback(int, int, struct rt_addrinfo *,
412 struct rtentry *, void *);
416 route_output(struct mbuf *m, struct socket *so, ...)
418 struct rt_msghdr *rtm = NULL;
420 struct radix_node_head *rnh;
421 struct rawcb *rp = NULL;
422 struct pr_output_info *oi;
423 struct rt_addrinfo rtinfo;
428 oi = __va_arg(ap, struct pr_output_info *);
431 #define gotoerr(e) { error = e; goto flush;}
434 (m->m_len < sizeof(long) &&
435 (m = m_pullup(m, sizeof(long))) == NULL))
437 if (!(m->m_flags & M_PKTHDR))
438 panic("route_output");
439 len = m->m_pkthdr.len;
440 if (len < sizeof(struct rt_msghdr) ||
441 len != mtod(m, struct rt_msghdr *)->rtm_msglen) {
442 rtinfo.rti_dst = NULL;
445 rtm = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
447 rtinfo.rti_dst = NULL;
450 m_copydata(m, 0, len, (caddr_t)rtm);
451 if (rtm->rtm_version != RTM_VERSION) {
452 rtinfo.rti_dst = NULL;
453 gotoerr(EPROTONOSUPPORT);
455 rtm->rtm_pid = oi->p_pid;
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;
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))
467 if (rtinfo.rti_genmask != NULL) {
468 struct radix_node *n;
470 #define clen(s) (*(u_char *)(s))
471 n = rn_addmask((char *)rtinfo.rti_genmask, TRUE, 1);
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;
482 * Verify that the caller has the appropriate privilege; RTM_GET
483 * is the only operation the non-superuser is allowed.
485 if (rtm->rtm_type != RTM_GET && suser_cred(so->so_cred, 0) != 0)
488 switch (rtm->rtm_type) {
490 if (rtinfo.rti_gateway == NULL) {
493 error = rtrequest1_global(RTM_ADD, &rtinfo,
494 route_output_add_callback, rtm);
499 * note: &rtm passed as argument so 'rtm' can be replaced.
501 error = rtrequest1_global(RTM_DELETE, &rtinfo,
502 route_output_delete_callback, &rtm);
505 rnh = rt_tables[mycpuid][rtinfo.rti_dst->sa_family];
507 error = EAFNOSUPPORT;
510 rt = (struct rtentry *)
511 rnh->rnh_lookup((char *)rtinfo.rti_dst,
512 (char *)rtinfo.rti_netmask, rnh);
518 if (fillrtmsg(&rtm, rt, &rtinfo) != 0)
523 error = rtrequest1_global(RTM_GET, &rtinfo,
524 route_output_change_callback, rtm);
527 error = rtrequest1_global(RTM_GET, &rtinfo,
528 route_output_lock_callback, rtm);
538 rtm->rtm_errno = error;
540 rtm->rtm_flags |= RTF_DONE;
544 * Check to see if we don't want our own messages.
546 if (!(so->so_options & SO_USELOOPBACK)) {
547 if (route_cb.any_count <= 1) {
549 kfree(rtm, M_RTABLE);
553 /* There is another listener, so construct message */
557 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm);
558 if (m->m_pkthdr.len < rtm->rtm_msglen) {
561 } else if (m->m_pkthdr.len > rtm->rtm_msglen)
562 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len);
563 kfree(rtm, M_RTABLE);
566 rp->rcb_proto.sp_family = 0; /* Avoid us */
568 rts_input(m, familyof(rtinfo.rti_dst));
570 rp->rcb_proto.sp_family = PF_ROUTE;
575 route_output_add_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
576 struct rtentry *rt, void *arg)
578 struct rt_msghdr *rtm = arg;
580 if (error == 0 && rt != NULL) {
581 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_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;
591 route_output_delete_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
592 struct rtentry *rt, void *arg)
594 struct rt_msghdr **rtm = arg;
596 if (error == 0 && rt) {
598 if (fillrtmsg(rtm, rt, rtinfo) != 0) {
600 /* XXX no way to return the error */
607 route_output_change_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
608 struct rtentry *rt, void *arg)
610 struct rt_msghdr *rtm = arg;
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
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))
625 error = rt_getifa(rtinfo);
629 if (rtinfo->rti_gateway != NULL) {
630 error = rt_setgate(rt, rt_key(rt), rtinfo->rti_gateway);
634 if ((ifa = rtinfo->rti_ifa) != NULL) {
635 struct ifaddr *oifa = rt->rt_ifa;
638 if (oifa && oifa->ifa_rtrequest)
639 oifa->ifa_rtrequest(RTM_DELETE, rt, rtinfo);
643 rt->rt_ifp = rtinfo->rti_ifp;
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;
652 /* XXX no way to return error */
657 route_output_lock_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
658 struct rtentry *rt, void *arg)
660 struct rt_msghdr *rtm = arg;
662 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits);
663 rt->rt_rmx.rmx_locks |=
664 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks);
668 rt_setmetrics(u_long which, struct rt_metrics *in, struct rt_metrics *out)
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);
683 ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long))
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.
691 rt_xaddrs(char *cp, char *cplim, struct rt_addrinfo *rtinfo)
696 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) {
697 if ((rtinfo->rti_addrs & (1 << i)) == 0)
699 sa = (struct sockaddr *)cp;
703 if ((cp + sa->sa_len) > cplim) {
708 * There are no more... Quit now.
709 * If there are more bits, they are in error.
710 * I've seen this. route(1) can evidently generate these.
711 * This causes kernel to core dump.
712 * For compatibility, if we see this, point to a safe address.
714 if (sa->sa_len == 0) {
715 static struct sockaddr sa_zero = {
716 sizeof sa_zero, AF_INET,
719 rtinfo->rti_info[i] = &sa_zero;
720 return (0); /* should be EINVAL but for compat */
723 /* Accept the sockaddr. */
724 rtinfo->rti_info[i] = sa;
725 cp += ROUNDUP(sa->sa_len);
731 rt_msghdrsize(int type)
736 return sizeof(struct ifa_msghdr);
739 return sizeof(struct ifma_msghdr);
741 return sizeof(struct if_msghdr);
744 return sizeof(struct if_announcemsghdr);
746 return sizeof(struct rt_msghdr);
751 rt_msgsize(int type, struct rt_addrinfo *rtinfo)
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);
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.
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
775 rt_msg_buffer(int type, struct rt_addrinfo *rtinfo, void *buf, int msglen)
777 struct rt_msghdr *rtm;
781 rtm = (struct rt_msghdr *) buf;
782 rtm->rtm_version = RTM_VERSION;
783 rtm->rtm_type = type;
784 rtm->rtm_msglen = msglen;
786 cp = (char *)buf + rt_msghdrsize(type);
787 rtinfo->rti_addrs = 0;
788 for (i = 0; i < RTAX_MAX; i++) {
791 if ((sa = rtinfo->rti_info[i]) == NULL)
793 rtinfo->rti_addrs |= (1 << i);
794 dlen = ROUNDUP(sa->sa_len);
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.
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
811 rt_msg_mbuf(int type, struct rt_addrinfo *rtinfo)
814 struct rt_msghdr *rtm;
818 hlen = rt_msghdrsize(type);
819 KASSERT(hlen <= MCLBYTES, ("rt_msg_mbuf: hlen %d doesn't fit", hlen));
821 m = m_getl(hlen, MB_DONTWAIT, MT_DATA, M_PKTHDR, NULL);
824 m->m_pkthdr.len = m->m_len = hlen;
825 m->m_pkthdr.rcvif = NULL;
826 rtinfo->rti_addrs = 0;
828 for (i = 0; i < RTAX_MAX; i++) {
832 if ((sa = rtinfo->rti_info[i]) == NULL)
834 rtinfo->rti_addrs |= (1 << i);
835 dlen = ROUNDUP(sa->sa_len);
836 m_copyback(m, len, dlen, (caddr_t)sa); /* can grow mbuf chain */
839 if (m->m_pkthdr.len != len) { /* one of the m_copyback() calls failed */
843 rtm = mtod(m, struct rt_msghdr *);
845 rtm->rtm_msglen = len;
846 rtm->rtm_version = RTM_VERSION;
847 rtm->rtm_type = type;
852 * This routine is called to generate a message from the routing
853 * socket indicating that a redirect has occurred, a routing lookup
854 * has failed, or that a protocol has detected timeouts to a particular
858 rt_missmsg(int type, struct rt_addrinfo *rtinfo, int flags, int error)
860 struct sockaddr *dst = rtinfo->rti_info[RTAX_DST];
861 struct rt_msghdr *rtm;
864 if (route_cb.any_count == 0)
866 m = rt_msg_mbuf(type, rtinfo);
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;
873 rts_input(m, familyof(dst));
877 rt_dstmsg(int type, struct sockaddr *dst, int error)
879 struct rt_msghdr *rtm;
880 struct rt_addrinfo addrs;
883 if (route_cb.any_count == 0)
885 bzero(&addrs, sizeof(struct rt_addrinfo));
886 addrs.rti_info[RTAX_DST] = dst;
887 m = rt_msg_mbuf(type, &addrs);
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;
894 rts_input(m, familyof(dst));
898 * This routine is called to generate a message from the routing
899 * socket indicating that the status of a network interface has changed.
902 rt_ifmsg(struct ifnet *ifp)
904 struct if_msghdr *ifm;
906 struct rt_addrinfo rtinfo;
908 if (route_cb.any_count == 0)
910 bzero(&rtinfo, sizeof(struct rt_addrinfo));
911 m = rt_msg_mbuf(RTM_IFINFO, &rtinfo);
914 ifm = mtod(m, struct if_msghdr *);
915 ifm->ifm_index = ifp->if_index;
916 ifm->ifm_flags = ifp->if_flags;
917 ifm->ifm_data = ifp->if_data;
923 rt_ifamsg(int cmd, struct ifaddr *ifa)
925 struct ifa_msghdr *ifam;
926 struct rt_addrinfo rtinfo;
928 struct ifnet *ifp = ifa->ifa_ifp;
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;
936 m = rt_msg_mbuf(cmd, &rtinfo);
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;
944 ifam->ifam_addrs = rtinfo.rti_addrs;
946 rts_input(m, familyof(ifa->ifa_addr));
950 rt_rtmsg(int cmd, struct rtentry *rt, struct ifnet *ifp, int error)
952 struct rt_msghdr *rtm;
953 struct rt_addrinfo rtinfo;
955 struct sockaddr *dst;
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);
965 rtinfo.rti_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr;
966 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr;
968 m = rt_msg_mbuf(cmd, &rtinfo);
972 rtm = mtod(m, struct rt_msghdr *);
974 rtm->rtm_index = ifp->if_index;
975 rtm->rtm_flags |= rt->rt_flags;
976 rtm->rtm_errno = error;
977 rtm->rtm_addrs = rtinfo.rti_addrs;
979 rts_input(m, familyof(dst));
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
991 rt_newaddrmsg(int cmd, struct ifaddr *ifa, int error, struct rtentry *rt)
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...
1000 sctp_add_ip_address(ifa);
1001 else if (cmd == RTM_DELETE)
1002 sctp_delete_ip_address(ifa);
1005 if (route_cb.any_count == 0)
1008 if (cmd == RTM_ADD) {
1009 rt_ifamsg(RTM_NEWADDR, ifa);
1010 rt_rtmsg(RTM_ADD, rt, ifa->ifa_ifp, error);
1012 KASSERT((cmd == RTM_DELETE), ("unknown cmd %d", cmd));
1013 rt_rtmsg(RTM_DELETE, rt, ifa->ifa_ifp, error);
1014 rt_ifamsg(RTM_DELADDR, ifa);
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.
1024 rt_newmaddrmsg(int cmd, struct ifmultiaddr *ifma)
1026 struct rt_addrinfo rtinfo;
1027 struct mbuf *m = NULL;
1028 struct ifnet *ifp = ifma->ifma_ifp;
1029 struct ifma_msghdr *ifmam;
1031 if (route_cb.any_count == 0)
1034 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1035 rtinfo.rti_ifaaddr = ifma->ifma_addr;
1036 if (ifp != NULL && !TAILQ_EMPTY(&ifp->if_addrhead))
1037 rtinfo.rti_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr;
1039 * If a link-layer address is present, present it as a ``gateway''
1040 * (similarly to how ARP entries, e.g., are presented).
1042 rtinfo.rti_gateway = ifma->ifma_lladdr;
1044 m = rt_msg_mbuf(cmd, &rtinfo);
1048 ifmam = mtod(m, struct ifma_msghdr *);
1049 ifmam->ifmam_index = ifp->if_index;
1050 ifmam->ifmam_addrs = rtinfo.rti_addrs;
1052 rts_input(m, familyof(ifma->ifma_addr));
1055 static struct mbuf *
1056 rt_makeifannouncemsg(struct ifnet *ifp, int type, int what,
1057 struct rt_addrinfo *info)
1059 struct if_announcemsghdr *ifan;
1062 if (route_cb.any_count == 0)
1065 bzero(info, sizeof(*info));
1066 m = rt_msg_mbuf(type, info);
1070 ifan = mtod(m, struct if_announcemsghdr *);
1071 ifan->ifan_index = ifp->if_index;
1072 strlcpy(ifan->ifan_name, ifp->if_xname, sizeof ifan->ifan_name);
1073 ifan->ifan_what = what;
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.
1083 rt_ieee80211msg(struct ifnet *ifp, int what, void *data, size_t data_len)
1085 struct rt_addrinfo info;
1088 m = rt_makeifannouncemsg(ifp, RTM_IEEE80211, what, &info);
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.
1097 * NB: we assume m is a single mbuf.
1099 if (data_len > M_TRAILINGSPACE(m)) {
1100 struct mbuf *n = m_get(MB_DONTWAIT, MT_DATA);
1105 bcopy(data, mtod(n, void *), data_len);
1106 n->m_len = data_len;
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;
1112 if (m->m_flags & M_PKTHDR)
1113 m->m_pkthdr.len += data_len;
1114 mtod(m, struct if_announcemsghdr *)->ifan_msglen += data_len;
1119 * This is called to generate routing socket messages indicating
1120 * network interface arrival and departure.
1123 rt_ifannouncemsg(struct ifnet *ifp, int what)
1125 struct rt_addrinfo addrinfo;
1128 m = rt_makeifannouncemsg(ifp, RTM_IFANNOUNCE, what, &addrinfo);
1134 resizewalkarg(struct walkarg *w, int len)
1138 newptr = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
1141 if (w->w_tmem != NULL)
1142 kfree(w->w_tmem, M_RTABLE);
1144 w->w_tmemsize = len;
1149 * This is used in dumping the kernel table via sysctl().
1152 sysctl_dumpentry(struct radix_node *rn, void *vw)
1154 struct walkarg *w = vw;
1155 struct rtentry *rt = (struct rtentry *)rn;
1156 struct rt_addrinfo rtinfo;
1159 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg))
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;
1167 if (rt->rt_ifp != NULL) {
1168 rtinfo.rti_ifpaddr =
1169 TAILQ_FIRST(&rt->rt_ifp->if_addrhead)->ifa_addr;
1170 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr;
1171 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT)
1172 rtinfo.rti_bcastaddr = rt->rt_ifa->ifa_dstaddr;
1174 msglen = rt_msgsize(RTM_GET, &rtinfo);
1175 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0)
1177 rt_msg_buffer(RTM_GET, &rtinfo, w->w_tmem, msglen);
1178 if (w->w_req != NULL) {
1179 struct rt_msghdr *rtm = w->w_tmem;
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;
1186 rtm->rtm_addrs = rtinfo.rti_addrs;
1187 error = SYSCTL_OUT(w->w_req, rtm, msglen);
1194 sysctl_iflist(int af, struct walkarg *w)
1198 struct rt_addrinfo rtinfo;
1201 bzero(&rtinfo, sizeof(struct rt_addrinfo));
1202 TAILQ_FOREACH(ifp, &ifnet, if_link) {
1203 if (w->w_arg && w->w_arg != ifp->if_index)
1205 ifa = TAILQ_FIRST(&ifp->if_addrhead);
1206 rtinfo.rti_ifpaddr = ifa->ifa_addr;
1207 msglen = rt_msgsize(RTM_IFINFO, &rtinfo);
1208 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0)
1210 rt_msg_buffer(RTM_IFINFO, &rtinfo, w->w_tmem, msglen);
1211 rtinfo.rti_ifpaddr = NULL;
1212 if (w->w_req != NULL && w->w_tmem != NULL) {
1213 struct if_msghdr *ifm = w->w_tmem;
1215 ifm->ifm_index = ifp->if_index;
1216 ifm->ifm_flags = ifp->if_flags;
1217 ifm->ifm_data = ifp->if_data;
1218 ifm->ifm_addrs = rtinfo.rti_addrs;
1219 error = SYSCTL_OUT(w->w_req, ifm, msglen);
1223 while ((ifa = TAILQ_NEXT(ifa, ifa_link)) != NULL) {
1224 if (af && af != ifa->ifa_addr->sa_family)
1226 if (curproc->p_ucred->cr_prison &&
1227 prison_if(curproc->p_ucred, ifa->ifa_addr))
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)
1236 rt_msg_buffer(RTM_NEWADDR, &rtinfo, w->w_tmem, msglen);
1237 if (w->w_req != NULL) {
1238 struct ifa_msghdr *ifam = w->w_tmem;
1240 ifam->ifam_index = ifa->ifa_ifp->if_index;
1241 ifam->ifam_flags = ifa->ifa_flags;
1242 ifam->ifam_metric = ifa->ifa_metric;
1243 ifam->ifam_addrs = rtinfo.rti_addrs;
1244 error = SYSCTL_OUT(w->w_req, w->w_tmem, msglen);
1249 rtinfo.rti_netmask = NULL;
1250 rtinfo.rti_ifaaddr = NULL;
1251 rtinfo.rti_bcastaddr = NULL;
1257 sysctl_rtsock(SYSCTL_HANDLER_ARGS)
1259 int *name = (int *)arg1;
1260 u_int namelen = arg2;
1261 struct radix_node_head *rnh;
1262 int i, error = EINVAL;
1271 if (namelen != 3 && namelen != 4)
1274 bzero(&w, sizeof w);
1280 * Optional third argument specifies cpu, used primarily for
1281 * debugging the route table.
1284 if (name[3] < 0 || name[3] >= ncpus)
1287 lwkt_migratecpu(name[3]);
1295 for (i = 1; i <= AF_MAX; i++)
1296 if ((rnh = rt_tables[mycpuid][i]) &&
1297 (af == 0 || af == i) &&
1298 (error = rnh->rnh_walktree(rnh,
1299 sysctl_dumpentry, &w)))
1304 error = sysctl_iflist(af, &w);
1307 if (w.w_tmem != NULL)
1308 kfree(w.w_tmem, M_RTABLE);
1310 lwkt_migratecpu(origcpu);
1314 SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, "");
1317 * Definitions of protocols supported in the ROUTE domain.
1320 static struct domain routedomain; /* or at least forward */
1322 static struct protosw routesw[] = {
1323 { SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR,
1324 0, route_output, raw_ctlinput, 0,
1331 static struct domain routedomain = {
1332 PF_ROUTE, "route", NULL, NULL, NULL,
1333 routesw, &routesw[(sizeof routesw)/(sizeof routesw[0])],