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"
34 #include "opt_ipsec.h"
37 #include <sys/param.h>
38 #include <sys/systm.h>
39 #include <sys/kernel.h>
41 #include <sys/malloc.h>
45 #include <sys/protosw.h>
46 #include <sys/socket.h>
47 #include <sys/socketvar.h>
48 #include <sys/sysctl.h>
50 #include <sys/thread2.h>
51 #include <sys/socketvar2.h>
52 #include <sys/msgport2.h>
54 #include <machine/stdarg.h>
58 #include <net/if_types.h>
60 #include <net/route.h>
63 #include <netinet/in.h>
64 #include <netinet/in_systm.h>
65 #include <netinet/ip.h>
66 #include <netinet/in_pcb.h>
67 #include <netinet/in_var.h>
68 #include <netinet/ip_var.h>
70 #include <net/ip_mroute/ip_mroute.h>
71 #include <net/ipfw/ip_fw.h>
72 #include <net/dummynet/ip_dummynet.h>
75 #include <netproto/ipsec/ipsec.h>
79 #include <netinet6/ipsec.h>
82 struct inpcbinfo ripcbinfo;
84 /* control hooks for ipfw and dummynet */
85 ip_fw_ctl_t *ip_fw_ctl_ptr;
86 ip_dn_ctl_t *ip_dn_ctl_ptr;
88 static struct lwkt_token raw_token = LWKT_TOKEN_INITIALIZER(raw_token);
92 * hooks for multicast routing. They all default to NULL,
93 * so leave them not initialized and rely on BSS being set to 0.
96 /* The socket used to communicate with the multicast routing daemon. */
97 struct socket *ip_mrouter;
99 /* The various mrouter and rsvp functions */
100 int (*ip_mrouter_set)(struct socket *, struct sockopt *);
101 int (*ip_mrouter_get)(struct socket *, struct sockopt *);
102 int (*ip_mrouter_done)(void);
103 int (*ip_mforward)(struct ip *, struct ifnet *, struct mbuf *,
104 struct ip_moptions *);
105 int (*mrt_ioctl)(int, caddr_t);
106 int (*legal_vif_num)(int);
107 u_long (*ip_mcast_src)(int);
109 int (*rsvp_input_p)(struct mbuf **, int *, int);
110 int (*ip_rsvp_vif)(struct socket *, struct sockopt *);
111 void (*ip_rsvp_force_done)(struct socket *);
114 * Nominal space allocated to a raw ip socket.
120 * Raw interface to IP protocol.
124 * Initialize raw connection block queue.
129 in_pcbinfo_init(&ripcbinfo);
131 * XXX We don't use the hash list for raw IP, but it's easier
132 * to allocate a one entry hash list than it is to check all
133 * over the place for hashbase == NULL.
135 ripcbinfo.hashbase = hashinit(1, M_PCB, &ripcbinfo.hashmask);
136 ripcbinfo.porthashbase = hashinit(1, M_PCB, &ripcbinfo.porthashmask);
137 ripcbinfo.wildcardhashbase = hashinit(1, M_PCB,
138 &ripcbinfo.wildcardhashmask);
139 ripcbinfo.ipi_size = sizeof(struct inpcb);
143 * Setup generic address and protocol structures
144 * for raw_input routine, then pass them along with
148 rip_input(struct mbuf **mp, int *offp, int proto)
150 struct sockaddr_in ripsrc = { sizeof ripsrc, AF_INET };
151 struct mbuf *m = *mp;
152 struct ip *ip = mtod(m, struct ip *);
154 struct inpcb *last = NULL;
155 struct mbuf *opts = NULL;
159 ripsrc.sin_addr = ip->ip_src;
160 lwkt_gettoken(&raw_token);
161 LIST_FOREACH(inp, &ripcbinfo.pcblisthead, inp_list) {
162 if (inp->inp_flags & INP_PLACEMARKER)
165 if ((inp->inp_vflag & INP_IPV4) == 0)
168 if (inp->inp_ip_p && inp->inp_ip_p != proto)
170 if (inp->inp_laddr.s_addr != INADDR_ANY &&
171 inp->inp_laddr.s_addr != ip->ip_dst.s_addr)
173 if (inp->inp_faddr.s_addr != INADDR_ANY &&
174 inp->inp_faddr.s_addr != ip->ip_src.s_addr)
177 struct mbuf *n = m_copypacket(m, MB_DONTWAIT);
180 /* check AH/ESP integrity. */
181 if (n && ipsec4_in_reject_so(n, last->inp_socket)) {
183 ipsecstat.in_polvio++;
184 /* do not inject data to pcb */
188 /* check AH/ESP integrity. */
189 if (ipsec4_in_reject(n, last)) {
191 /* do not inject data to pcb */
193 #endif /*FAST_IPSEC*/
195 lwkt_gettoken(&last->inp_socket->so_rcv.ssb_token);
196 if (last->inp_flags & INP_CONTROLOPTS ||
197 last->inp_socket->so_options & SO_TIMESTAMP)
198 ip_savecontrol(last, &opts, ip, n);
199 if (ssb_appendaddr(&last->inp_socket->so_rcv,
200 (struct sockaddr *)&ripsrc, n,
202 /* should notify about lost packet */
207 sorwakeup(last->inp_socket);
209 lwkt_reltoken(&last->inp_socket->so_rcv.ssb_token);
216 /* check AH/ESP integrity. */
217 if (last && ipsec4_in_reject_so(m, last->inp_socket)) {
219 ipsecstat.in_polvio++;
220 ipstat.ips_delivered--;
221 /* do not inject data to pcb */
225 /* check AH/ESP integrity. */
226 if (last && ipsec4_in_reject(m, last)) {
228 ipstat.ips_delivered--;
229 /* do not inject data to pcb */
231 #endif /*FAST_IPSEC*/
232 /* Check the minimum TTL for socket. */
233 if (last && ip->ip_ttl < last->inp_ip_minttl) {
235 ipstat.ips_delivered--;
237 if (last->inp_flags & INP_CONTROLOPTS ||
238 last->inp_socket->so_options & SO_TIMESTAMP)
239 ip_savecontrol(last, &opts, ip, m);
240 lwkt_gettoken(&last->inp_socket->so_rcv.ssb_token);
241 if (ssb_appendaddr(&last->inp_socket->so_rcv,
242 (struct sockaddr *)&ripsrc, m, opts) == 0) {
247 sorwakeup(last->inp_socket);
249 lwkt_reltoken(&last->inp_socket->so_rcv.ssb_token);
252 ipstat.ips_noproto++;
253 ipstat.ips_delivered--;
255 lwkt_reltoken(&raw_token);
256 return(IPPROTO_DONE);
260 * Generate IP header and pass packet to ip_output.
261 * Tack on options user may have setup with control call.
264 rip_output(struct mbuf *m, struct socket *so, ...)
267 struct inpcb *inp = so->so_pcb;
269 int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST;
273 dst = __va_arg(ap, u_long);
277 * If the user handed us a complete IP packet, use it.
278 * Otherwise, allocate an mbuf for a header and fill it in.
280 if ((inp->inp_flags & INP_HDRINCL) == 0) {
281 if (m->m_pkthdr.len + sizeof(struct ip) > IP_MAXPACKET) {
285 M_PREPEND(m, sizeof(struct ip), MB_WAIT);
288 ip = mtod(m, struct ip *);
289 ip->ip_tos = inp->inp_ip_tos;
291 ip->ip_p = inp->inp_ip_p;
292 ip->ip_len = m->m_pkthdr.len;
293 ip->ip_src = inp->inp_laddr;
294 ip->ip_dst.s_addr = dst;
295 ip->ip_ttl = inp->inp_ip_ttl;
299 if (m->m_pkthdr.len > IP_MAXPACKET) {
303 if (m->m_len < sizeof(struct ip)) {
304 m = m_pullup(m, sizeof(struct ip));
308 ip = mtod(m, struct ip *);
309 hlen = IP_VHL_HL(ip->ip_vhl) << 2;
311 /* Don't allow header length less than the minimum. */
312 if (hlen < sizeof(struct ip)) {
318 * Don't allow both user specified and setsockopt options.
319 * Don't allow packet length sizes that will crash.
321 if ((hlen != sizeof(struct ip) && inp->inp_options) ||
322 ip->ip_len > m->m_pkthdr.len || ip->ip_len < hlen) {
327 ip->ip_id = ip_newid();
329 /* Prevent ip_output from overwriting header fields */
330 flags |= IP_RAWOUTPUT;
334 return ip_output(m, inp->inp_options, &inp->inp_route, flags,
335 inp->inp_moptions, inp);
339 * Raw IP socket option processing.
342 rip_ctloutput(netmsg_t msg)
344 struct socket *so = msg->base.nm_so;
345 struct sockopt *sopt = msg->ctloutput.nm_sopt;
346 struct inpcb *inp = so->so_pcb;
349 if (sopt->sopt_level != IPPROTO_IP) {
356 switch (sopt->sopt_dir) {
358 switch (sopt->sopt_name) {
360 optval = inp->inp_flags & INP_HDRINCL;
361 soopt_from_kbuf(sopt, &optval, sizeof optval);
364 case IP_FW_ADD: /* ADD actually returns the body... */
367 error = ip_fw_sockopt(sopt);
372 case IP_DUMMYNET_GET:
373 error = ip_dn_sockopt(sopt);
384 case MRT_API_SUPPORT:
386 case MRT_ADD_BW_UPCALL:
387 case MRT_DEL_BW_UPCALL:
388 error = ip_mrouter_get ? ip_mrouter_get(so, sopt) :
394 /* msg invalid now */
400 switch (sopt->sopt_name) {
402 error = soopt_to_kbuf(sopt, &optval, sizeof optval,
407 inp->inp_flags |= INP_HDRINCL;
409 inp->inp_flags &= ~INP_HDRINCL;
418 error = ip_fw_ctl_ptr(sopt);
423 case IP_DUMMYNET_CONFIGURE:
424 case IP_DUMMYNET_DEL:
425 case IP_DUMMYNET_FLUSH:
426 error = ip_dn_sockopt(sopt);
430 error = ip_rsvp_init(so);
434 error = ip_rsvp_done();
438 case IP_RSVP_VIF_OFF:
439 error = ip_rsvp_vif ?
440 ip_rsvp_vif(so, sopt) : EINVAL;
451 case MRT_API_SUPPORT:
453 case MRT_ADD_BW_UPCALL:
454 case MRT_DEL_BW_UPCALL:
455 error = ip_mrouter_set ? ip_mrouter_set(so, sopt) :
461 /* msg invalid now */
467 lwkt_replymsg(&msg->lmsg, error);
471 * This function exists solely to receive the PRC_IFDOWN messages which
472 * are sent by if_down(). It looks for an ifaddr whose ifa_addr is sa,
473 * and calls in_ifadown() to remove all routes corresponding to that address.
474 * It also receives the PRC_IFUP messages from if_up() and reinstalls the
478 rip_ctlinput(netmsg_t msg)
480 int cmd = msg->ctlinput.nm_cmd;
481 struct sockaddr *sa = msg->ctlinput.nm_arg;
482 struct in_ifaddr *ia;
483 struct in_ifaddr_container *iac;
490 TAILQ_FOREACH(iac, &in_ifaddrheads[mycpuid], ia_link) {
493 if (ia->ia_ifa.ifa_addr == sa &&
494 (ia->ia_flags & IFA_ROUTE)) {
496 * in_ifscrub kills the interface route.
498 in_ifscrub(ia->ia_ifp, ia);
500 * in_ifadown gets rid of all the rest of
501 * the routes. This is not quite the right
502 * thing to do, but at least if we are running
503 * a routing process they will come back.
505 in_ifadown(&ia->ia_ifa, 0);
513 TAILQ_FOREACH(iac, &in_ifaddrheads[mycpuid], ia_link) {
514 if (iac->ia->ia_ifa.ifa_addr == sa) {
519 if (ia == NULL || (ia->ia_flags & IFA_ROUTE))
522 ifp = ia->ia_ifa.ifa_ifp;
526 * Don't add prefix routes for CARP interfaces.
527 * Prefix routes creation is handled by CARP
528 * interfaces themselves.
530 if (ifp->if_type == IFT_CARP)
534 if ((ifp->if_flags & IFF_LOOPBACK) ||
535 (ifp->if_flags & IFF_POINTOPOINT))
538 err = rtinit(&ia->ia_ifa, RTM_ADD, flags);
540 ia->ia_flags |= IFA_ROUTE;
544 lwkt_replymsg(&msg->lmsg, 0);
547 u_long rip_sendspace = RIPSNDQ;
548 u_long rip_recvspace = RIPRCVQ;
550 SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW,
551 &rip_sendspace, 0, "Maximum outgoing raw IP datagram size");
552 SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW,
553 &rip_recvspace, 0, "Maximum incoming raw IP datagram size");
556 rip_attach(netmsg_t msg)
558 struct socket *so = msg->base.nm_so;
559 int proto = msg->attach.nm_proto;
560 struct pru_attach_info *ai = msg->attach.nm_ai;
567 error = priv_check_cred(ai->p_ucred, PRIV_NETINET_RAW, NULL_CRED_OKAY);
571 error = soreserve(so, rip_sendspace, rip_recvspace, ai->sb_rlimit);
575 lwkt_gettoken(&raw_token);
576 error = in_pcballoc(so, &ripcbinfo);
578 inp = (struct inpcb *)so->so_pcb;
579 inp->inp_vflag |= INP_IPV4;
580 inp->inp_ip_p = proto;
581 inp->inp_ip_ttl = ip_defttl;
583 lwkt_reltoken(&raw_token);
586 lwkt_replymsg(&msg->lmsg, error);
590 rip_detach(netmsg_t msg)
592 struct socket *so = msg->base.nm_so;
598 if (so == ip_mrouter && ip_mrouter_done)
600 if (ip_rsvp_force_done)
601 ip_rsvp_force_done(so);
605 lwkt_replymsg(&msg->lmsg, 0);
609 * NOTE: (so) is referenced from soabort*() and netmsg_pru_abort()
610 * will sofree() it when we return.
613 rip_abort(netmsg_t msg)
615 struct socket *so = msg->base.nm_so;
618 soisdisconnected(so);
619 if (so->so_state & SS_NOFDREF) { /* XXX not sure why this test */
621 /* msg invalid now */
625 lwkt_replymsg(&msg->lmsg, error);
629 rip_disconnect(netmsg_t msg)
631 struct socket *so = msg->base.nm_so;
634 if (so->so_state & SS_ISCONNECTED) {
637 /* msg invalid now */
642 lwkt_replymsg(&msg->lmsg, error);
646 rip_bind(netmsg_t msg)
648 struct socket *so = msg->base.nm_so;
649 struct sockaddr *nam = msg->bind.nm_nam;
650 struct inpcb *inp = so->so_pcb;
651 struct sockaddr_in *addr = (struct sockaddr_in *)nam;
654 if (nam->sa_len == sizeof(*addr)) {
655 if (TAILQ_EMPTY(&ifnet) ||
656 ((addr->sin_family != AF_INET) &&
657 (addr->sin_family != AF_IMPLINK)) ||
658 (addr->sin_addr.s_addr != INADDR_ANY &&
659 ifa_ifwithaddr((struct sockaddr *)addr) == 0)) {
660 error = EADDRNOTAVAIL;
662 inp->inp_laddr = addr->sin_addr;
668 lwkt_replymsg(&msg->lmsg, error);
672 rip_connect(netmsg_t msg)
674 struct socket *so = msg->base.nm_so;
675 struct sockaddr *nam = msg->connect.nm_nam;
676 struct inpcb *inp = so->so_pcb;
677 struct sockaddr_in *addr = (struct sockaddr_in *)nam;
680 if (nam->sa_len != sizeof(*addr)) {
682 } else if (TAILQ_EMPTY(&ifnet)) {
683 error = EADDRNOTAVAIL;
685 if ((addr->sin_family != AF_INET) &&
686 (addr->sin_family != AF_IMPLINK)) {
687 error = EAFNOSUPPORT;
689 inp->inp_faddr = addr->sin_addr;
694 lwkt_replymsg(&msg->lmsg, error);
698 rip_shutdown(netmsg_t msg)
700 socantsendmore(msg->base.nm_so);
701 lwkt_replymsg(&msg->lmsg, 0);
705 rip_send(netmsg_t msg)
707 struct socket *so = msg->base.nm_so;
708 struct mbuf *m = msg->send.nm_m;
709 /*struct mbuf *control = msg->send.nm_control;*/
710 struct sockaddr *nam = msg->send.nm_addr;
711 /*int flags = msg->send.nm_flags;*/
712 struct inpcb *inp = so->so_pcb;
716 if (so->so_state & SS_ISCONNECTED) {
721 dst = inp->inp_faddr.s_addr;
722 error = rip_output(m, so, dst);
729 dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr;
730 error = rip_output(m, so, dst);
733 lwkt_replymsg(&msg->lmsg, error);
736 SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, &ripcbinfo, 0,
737 in_pcblist_global, "S,xinpcb", "List of active raw IP sockets");
739 struct pr_usrreqs rip_usrreqs = {
740 .pru_abort = rip_abort,
741 .pru_accept = pr_generic_notsupp,
742 .pru_attach = rip_attach,
743 .pru_bind = rip_bind,
744 .pru_connect = rip_connect,
745 .pru_connect2 = pr_generic_notsupp,
746 .pru_control = in_control_dispatch,
747 .pru_detach = rip_detach,
748 .pru_disconnect = rip_disconnect,
749 .pru_listen = pr_generic_notsupp,
750 .pru_peeraddr = in_setpeeraddr_dispatch,
751 .pru_rcvd = pr_generic_notsupp,
752 .pru_rcvoob = pr_generic_notsupp,
753 .pru_send = rip_send,
754 .pru_sense = pru_sense_null,
755 .pru_shutdown = rip_shutdown,
756 .pru_sockaddr = in_setsockaddr_dispatch,
757 .pru_sosend = sosend,
758 .pru_soreceive = soreceive