2 * Copyright (c) 1982, 1986, 1988, 1993
3 * The Regents of the University of California. All rights reserved.
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
8 * 1. Redistributions of source code must retain the above copyright
9 * notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 * notice, this list of conditions and the following disclaimer in the
12 * documentation and/or other materials provided with the distribution.
13 * 3. Neither the name of the University nor the names of its contributors
14 * may be used to endorse or promote products derived from this software
15 * without specific prior written permission.
17 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 * @(#)raw_ip.c 8.7 (Berkeley) 5/15/95
30 * $FreeBSD: src/sys/netinet/raw_ip.c,v 1.64.2.16 2003/08/24 08:24:38 hsu Exp $
33 #include "opt_inet6.h"
36 #include <sys/param.h>
37 #include <sys/systm.h>
38 #include <sys/kernel.h>
40 #include <sys/malloc.h>
44 #include <sys/protosw.h>
45 #include <sys/socket.h>
46 #include <sys/socketvar.h>
47 #include <sys/sysctl.h>
49 #include <sys/thread2.h>
50 #include <sys/socketvar2.h>
51 #include <sys/msgport2.h>
53 #include <machine/stdarg.h>
57 #include <net/if_types.h>
59 #include <net/route.h>
62 #include <netinet/in.h>
63 #include <netinet/in_systm.h>
64 #include <netinet/ip.h>
65 #include <netinet/in_pcb.h>
66 #include <netinet/in_var.h>
67 #include <netinet/ip_var.h>
69 #include <net/ip_mroute/ip_mroute.h>
70 #include <net/ipfw/ip_fw.h>
71 #include <net/ipfw3/ip_fw.h>
72 #include <net/dummynet/ip_dummynet.h>
73 #include <net/dummynet3/ip_dummynet3.h>
75 struct inpcbinfo ripcbinfo;
76 struct inpcbportinfo ripcbportinfo;
79 * hooks for multicast routing. They all default to NULL,
80 * so leave them not initialized and rely on BSS being set to 0.
83 /* The socket used to communicate with the multicast routing daemon. */
84 struct socket *ip_mrouter;
86 /* The various mrouter and rsvp functions */
87 int (*ip_mrouter_set)(struct socket *, struct sockopt *);
88 int (*ip_mrouter_get)(struct socket *, struct sockopt *);
89 int (*ip_mrouter_done)(void);
90 int (*ip_mforward)(struct ip *, struct ifnet *, struct mbuf *,
91 struct ip_moptions *);
92 int (*mrt_ioctl)(u_long, caddr_t);
93 int (*legal_vif_num)(int);
94 u_long (*ip_mcast_src)(int);
96 int (*rsvp_input_p)(struct mbuf **, int *, int);
97 int (*ip_rsvp_vif)(struct socket *, struct sockopt *);
98 void (*ip_rsvp_force_done)(struct socket *);
101 * Nominal space allocated to a raw ip socket.
107 * Raw interface to IP protocol.
111 * Initialize raw connection block queue.
116 in_pcbinfo_init(&ripcbinfo, 0, FALSE);
117 in_pcbportinfo_init(&ripcbportinfo, 1, 0);
119 * XXX We don't use the hash list for raw IP, but it's easier
120 * to allocate a one entry hash list than it is to check all
121 * over the place for hashbase == NULL.
123 ripcbinfo.hashbase = hashinit(1, M_PCB, &ripcbinfo.hashmask);
124 in_pcbportinfo_set(&ripcbinfo, &ripcbportinfo, 1);
125 ripcbinfo.wildcardhashbase = hashinit(1, M_PCB,
126 &ripcbinfo.wildcardhashmask);
127 ripcbinfo.ipi_size = sizeof(struct inpcb);
131 * Setup generic address and protocol structures
132 * for raw_input routine, then pass them along with
136 rip_input(struct mbuf **mp, int *offp, int proto)
138 struct sockaddr_in ripsrc = { sizeof ripsrc, AF_INET };
139 struct mbuf *m = *mp;
140 struct ip *ip = mtod(m, struct ip *);
142 struct inpcb *last = NULL;
143 struct mbuf *opts = NULL;
149 ripsrc.sin_addr = ip->ip_src;
150 LIST_FOREACH(inp, &ripcbinfo.pcblisthead, inp_list) {
151 if (inp->inp_flags & INP_PLACEMARKER)
154 if (!INP_ISIPV4(inp))
157 if (inp->inp_ip_p && inp->inp_ip_p != proto)
159 if (inp->inp_laddr.s_addr != INADDR_ANY &&
160 inp->inp_laddr.s_addr != ip->ip_dst.s_addr)
162 if (inp->inp_faddr.s_addr != INADDR_ANY &&
163 inp->inp_faddr.s_addr != ip->ip_src.s_addr)
166 struct mbuf *n = m_copypacket(m, M_NOWAIT);
169 lwkt_gettoken(&last->inp_socket->so_rcv.ssb_token);
170 if (last->inp_flags & INP_CONTROLOPTS ||
171 last->inp_socket->so_options & SO_TIMESTAMP)
172 ip_savecontrol(last, &opts, ip, n);
173 if (ssb_appendaddr(&last->inp_socket->so_rcv,
174 (struct sockaddr *)&ripsrc, n,
176 /* should notify about lost packet */
181 sorwakeup(last->inp_socket);
183 lwkt_reltoken(&last->inp_socket->so_rcv.ssb_token);
189 /* Check the minimum TTL for socket. */
190 if (last && ip->ip_ttl < last->inp_ip_minttl) {
192 ipstat.ips_delivered--;
194 if (last->inp_flags & INP_CONTROLOPTS ||
195 last->inp_socket->so_options & SO_TIMESTAMP)
196 ip_savecontrol(last, &opts, ip, m);
197 lwkt_gettoken(&last->inp_socket->so_rcv.ssb_token);
198 if (ssb_appendaddr(&last->inp_socket->so_rcv,
199 (struct sockaddr *)&ripsrc, m, opts) == 0) {
204 sorwakeup(last->inp_socket);
206 lwkt_reltoken(&last->inp_socket->so_rcv.ssb_token);
209 ipstat.ips_noproto++;
210 ipstat.ips_delivered--;
212 return(IPPROTO_DONE);
216 * Generate IP header and pass packet to ip_output.
217 * Tack on options user may have setup with control call.
220 rip_output(struct mbuf *m, struct socket *so, ...)
223 struct inpcb *inp = so->so_pcb;
225 int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST;
231 dst = __va_arg(ap, u_long);
235 * If the user handed us a complete IP packet, use it.
236 * Otherwise, allocate an mbuf for a header and fill it in.
238 if ((inp->inp_flags & INP_HDRINCL) == 0) {
239 if (m->m_pkthdr.len + sizeof(struct ip) > IP_MAXPACKET) {
243 M_PREPEND(m, sizeof(struct ip), M_WAITOK);
246 ip = mtod(m, struct ip *);
247 ip->ip_tos = inp->inp_ip_tos;
249 ip->ip_p = inp->inp_ip_p;
250 ip->ip_len = m->m_pkthdr.len;
251 ip->ip_src = inp->inp_laddr;
252 ip->ip_dst.s_addr = dst;
253 ip->ip_ttl = inp->inp_ip_ttl;
257 if (m->m_pkthdr.len > IP_MAXPACKET) {
261 if (m->m_len < sizeof(struct ip)) {
262 m = m_pullup(m, sizeof(struct ip));
266 ip = mtod(m, struct ip *);
267 hlen = IP_VHL_HL(ip->ip_vhl) << 2;
269 /* Don't allow header length less than the minimum. */
270 if (hlen < sizeof(struct ip)) {
276 * Don't allow both user specified and setsockopt options.
277 * Don't allow packet length sizes that will crash.
279 if ((hlen != sizeof(struct ip) && inp->inp_options) ||
280 ip->ip_len > m->m_pkthdr.len || ip->ip_len < hlen) {
285 ip->ip_id = ip_newid();
287 /* Prevent ip_output from overwriting header fields */
288 flags |= IP_RAWOUTPUT;
292 return ip_output(m, inp->inp_options, &inp->inp_route, flags,
293 inp->inp_moptions, inp);
297 * Raw IP socket option processing.
300 rip_ctloutput(netmsg_t msg)
302 struct socket *so = msg->base.nm_so;
303 struct sockopt *sopt = msg->ctloutput.nm_sopt;
304 struct inpcb *inp = so->so_pcb;
311 /* Get socket's owner cpuid hint */
312 if (sopt->sopt_level == SOL_SOCKET &&
313 sopt->sopt_dir == SOPT_GET &&
314 sopt->sopt_name == SO_CPUHINT) {
316 soopt_from_kbuf(sopt, &optval, sizeof(optval));
320 if (sopt->sopt_level != IPPROTO_IP) {
325 switch (sopt->sopt_dir) {
327 switch (sopt->sopt_name) {
329 optval = inp->inp_flags & INP_HDRINCL;
330 soopt_from_kbuf(sopt, &optval, sizeof optval);
334 error = ip_fw3_sockopt(sopt);
337 case IP_FW_ADD: /* ADD actually returns the body... */
340 case IP_FW_TBL_EXPIRE: /* returns # of expired addresses */
341 error = ip_fw_sockopt(sopt);
344 case IP_DUMMYNET_GET:
345 error = ip_dn_sockopt(sopt);
356 case MRT_API_SUPPORT:
358 case MRT_ADD_BW_UPCALL:
359 case MRT_DEL_BW_UPCALL:
360 error = ip_mrouter_get ? ip_mrouter_get(so, sopt) :
366 /* msg invalid now */
372 switch (sopt->sopt_name) {
374 error = soopt_to_kbuf(sopt, &optval, sizeof optval,
379 inp->inp_flags |= INP_HDRINCL;
381 inp->inp_flags &= ~INP_HDRINCL;
385 error = ip_fw3_sockopt(sopt);
393 case IP_FW_TBL_CREATE:
394 case IP_FW_TBL_DESTROY:
397 case IP_FW_TBL_FLUSH:
399 case IP_FW_TBL_EXPIRE:
400 error = ip_fw_sockopt(sopt);
403 case IP_DUMMYNET_CONFIGURE:
404 case IP_DUMMYNET_DEL:
405 case IP_DUMMYNET_FLUSH:
406 error = ip_dn_sockopt(sopt);
410 error = ip_rsvp_init(so);
414 error = ip_rsvp_done();
418 case IP_RSVP_VIF_OFF:
419 error = ip_rsvp_vif ?
420 ip_rsvp_vif(so, sopt) : EINVAL;
431 case MRT_API_SUPPORT:
433 case MRT_ADD_BW_UPCALL:
434 case MRT_DEL_BW_UPCALL:
435 error = ip_mrouter_set ? ip_mrouter_set(so, sopt) :
441 /* msg invalid now */
447 lwkt_replymsg(&msg->lmsg, error);
451 * This function exists solely to receive the PRC_IFDOWN messages which
452 * are sent by if_down(). It looks for an ifaddr whose ifa_addr is sa,
453 * and calls in_ifadown() to remove all routes corresponding to that address.
454 * It also receives the PRC_IFUP messages from if_up() and reinstalls the
458 rip_ctlinput(netmsg_t msg)
460 int cmd = msg->ctlinput.nm_cmd;
461 struct sockaddr *sa = msg->ctlinput.nm_arg;
462 struct in_ifaddr *ia;
463 struct in_ifaddr_container *iac;
472 TAILQ_FOREACH(iac, &in_ifaddrheads[mycpuid], ia_link) {
475 if (ia->ia_ifa.ifa_addr == sa &&
476 (ia->ia_flags & IFA_ROUTE)) {
478 * in_ifscrub kills the interface route.
480 in_ifscrub(ia->ia_ifp, ia);
482 * in_ifadown gets rid of all the rest of
483 * the routes. This is not quite the right
484 * thing to do, but at least if we are running
485 * a routing process they will come back.
487 in_ifadown(&ia->ia_ifa, 0);
495 TAILQ_FOREACH(iac, &in_ifaddrheads[mycpuid], ia_link) {
496 if (iac->ia->ia_ifa.ifa_addr == sa) {
501 if (ia == NULL || (ia->ia_flags & IFA_ROUTE))
504 ifp = ia->ia_ifa.ifa_ifp;
508 * Don't add prefix routes for CARP interfaces.
509 * Prefix routes creation is handled by CARP
510 * interfaces themselves.
512 if (ifp->if_type == IFT_CARP)
516 if ((ifp->if_flags & IFF_LOOPBACK) ||
517 (ifp->if_flags & IFF_POINTOPOINT))
520 err = rtinit(&ia->ia_ifa, RTM_ADD, flags);
522 ia->ia_flags |= IFA_ROUTE;
526 lwkt_replymsg(&msg->lmsg, 0);
529 u_long rip_sendspace = RIPSNDQ;
530 u_long rip_recvspace = RIPRCVQ;
532 SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW,
533 &rip_sendspace, 0, "Maximum outgoing raw IP datagram size");
534 SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW,
535 &rip_recvspace, 0, "Maximum incoming raw IP datagram size");
538 rip_attach(netmsg_t msg)
540 struct socket *so = msg->base.nm_so;
541 int proto = msg->attach.nm_proto;
542 struct pru_attach_info *ai = msg->attach.nm_ai;
551 error = priv_check_cred(ai->p_ucred, PRIV_NETINET_RAW, NULL_CRED_OKAY);
555 error = soreserve(so, rip_sendspace, rip_recvspace, ai->sb_rlimit);
559 error = in_pcballoc(so, &ripcbinfo);
561 inp = (struct inpcb *)so->so_pcb;
562 inp->inp_ip_p = proto;
563 inp->inp_ip_ttl = ip_defttl;
566 lwkt_replymsg(&msg->lmsg, error);
570 rip_detach(netmsg_t msg)
572 struct socket *so = msg->base.nm_so;
580 if (so == ip_mrouter && ip_mrouter_done)
582 if (ip_rsvp_force_done)
583 ip_rsvp_force_done(so);
587 lwkt_replymsg(&msg->lmsg, 0);
591 rip_abort(netmsg_t msg)
594 * Raw socket does not support listen(2),
595 * so this should never be called.
597 panic("rip_abort is called");
601 rip_disconnect(netmsg_t msg)
603 struct socket *so = msg->base.nm_so;
608 if (so->so_state & SS_ISCONNECTED) {
609 soisdisconnected(so);
614 lwkt_replymsg(&msg->lmsg, error);
618 rip_bind(netmsg_t msg)
620 struct socket *so = msg->base.nm_so;
621 struct sockaddr *nam = msg->bind.nm_nam;
622 struct inpcb *inp = so->so_pcb;
623 struct sockaddr_in *addr = (struct sockaddr_in *)nam;
628 if (nam->sa_len == sizeof(*addr)) {
629 if (ifnet_array_isempty() ||
630 ((addr->sin_family != AF_INET) &&
631 (addr->sin_family != AF_IMPLINK)) ||
632 (addr->sin_addr.s_addr != INADDR_ANY &&
633 ifa_ifwithaddr((struct sockaddr *)addr) == 0)) {
634 error = EADDRNOTAVAIL;
636 inp->inp_laddr = addr->sin_addr;
642 lwkt_replymsg(&msg->lmsg, error);
646 rip_connect(netmsg_t msg)
648 struct socket *so = msg->base.nm_so;
649 struct sockaddr *nam = msg->connect.nm_nam;
650 struct inpcb *inp = so->so_pcb;
651 struct sockaddr_in *addr = (struct sockaddr_in *)nam;
656 if (nam->sa_len != sizeof(*addr)) {
658 } else if (ifnet_array_isempty()) {
659 error = EADDRNOTAVAIL;
661 if ((addr->sin_family != AF_INET) &&
662 (addr->sin_family != AF_IMPLINK)) {
663 error = EAFNOSUPPORT;
665 inp->inp_faddr = addr->sin_addr;
670 lwkt_replymsg(&msg->lmsg, error);
674 rip_shutdown(netmsg_t msg)
678 socantsendmore(msg->base.nm_so);
679 lwkt_replymsg(&msg->lmsg, 0);
683 rip_send(netmsg_t msg)
685 struct socket *so = msg->base.nm_so;
686 struct mbuf *m = msg->send.nm_m;
687 /*struct mbuf *control = msg->send.nm_control;*/
688 struct sockaddr *nam = msg->send.nm_addr;
689 /*int flags = msg->send.nm_flags;*/
690 struct inpcb *inp = so->so_pcb;
696 if (so->so_state & SS_ISCONNECTED) {
701 dst = inp->inp_faddr.s_addr;
702 error = rip_output(m, so, dst);
709 dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr;
710 error = rip_output(m, so, dst);
713 lwkt_replymsg(&msg->lmsg, error);
716 SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, &ripcbinfo, 1,
717 in_pcblist_range, "S,xinpcb", "List of active raw IP sockets");
719 struct pr_usrreqs rip_usrreqs = {
720 .pru_abort = rip_abort,
721 .pru_accept = pr_generic_notsupp,
722 .pru_attach = rip_attach,
723 .pru_bind = rip_bind,
724 .pru_connect = rip_connect,
725 .pru_connect2 = pr_generic_notsupp,
726 .pru_control = in_control_dispatch,
727 .pru_detach = rip_detach,
728 .pru_disconnect = rip_disconnect,
729 .pru_listen = pr_generic_notsupp,
730 .pru_peeraddr = in_setpeeraddr_dispatch,
731 .pru_rcvd = pr_generic_notsupp,
732 .pru_rcvoob = pr_generic_notsupp,
733 .pru_send = rip_send,
734 .pru_sense = pru_sense_null,
735 .pru_shutdown = rip_shutdown,
736 .pru_sockaddr = in_setsockaddr_dispatch,
737 .pru_sosend = sosend,
738 .pru_soreceive = soreceive