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. All advertising materials mentioning features or use of this software
14 * must display the following acknowledgement:
15 * This product includes software developed by the University of
16 * California, Berkeley and its contributors.
17 * 4. Neither the name of the University nor the names of its contributors
18 * may be used to endorse or promote products derived from this software
19 * without specific prior written permission.
21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
33 * @(#)raw_ip.c 8.7 (Berkeley) 5/15/95
34 * $FreeBSD: src/sys/netinet/raw_ip.c,v 1.64.2.16 2003/08/24 08:24:38 hsu Exp $
37 #include "opt_inet6.h"
38 #include "opt_ipsec.h"
40 #include <sys/param.h>
41 #include <sys/systm.h>
42 #include <sys/kernel.h>
44 #include <sys/malloc.h>
48 #include <sys/protosw.h>
49 #include <sys/socket.h>
50 #include <sys/socketvar.h>
51 #include <sys/sysctl.h>
53 #include <sys/thread2.h>
54 #include <sys/socketvar2.h>
55 #include <sys/msgport2.h>
57 #include <machine/stdarg.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;
161 ripsrc.sin_addr = ip->ip_src;
162 lwkt_gettoken(&raw_token);
163 LIST_FOREACH(inp, &ripcbinfo.pcblisthead, inp_list) {
164 if (inp->inp_flags & INP_PLACEMARKER)
167 if ((inp->inp_vflag & INP_IPV4) == 0)
170 if (inp->inp_ip_p && inp->inp_ip_p != proto)
172 if (inp->inp_laddr.s_addr != INADDR_ANY &&
173 inp->inp_laddr.s_addr != ip->ip_dst.s_addr)
175 if (inp->inp_faddr.s_addr != INADDR_ANY &&
176 inp->inp_faddr.s_addr != ip->ip_src.s_addr)
179 struct mbuf *n = m_copypacket(m, MB_DONTWAIT);
182 /* check AH/ESP integrity. */
183 if (n && ipsec4_in_reject_so(n, last->inp_socket)) {
185 ipsecstat.in_polvio++;
186 /* do not inject data to pcb */
190 /* check AH/ESP integrity. */
191 if (ipsec4_in_reject(n, last)) {
193 /* do not inject data to pcb */
195 #endif /*FAST_IPSEC*/
197 lwkt_gettoken(&last->inp_socket->so_rcv.ssb_token);
198 if (last->inp_flags & INP_CONTROLOPTS ||
199 last->inp_socket->so_options & SO_TIMESTAMP)
200 ip_savecontrol(last, &opts, ip, n);
201 if (ssb_appendaddr(&last->inp_socket->so_rcv,
202 (struct sockaddr *)&ripsrc, n,
204 /* should notify about lost packet */
209 sorwakeup(last->inp_socket);
211 lwkt_reltoken(&last->inp_socket->so_rcv.ssb_token);
218 /* check AH/ESP integrity. */
219 if (last && ipsec4_in_reject_so(m, last->inp_socket)) {
221 ipsecstat.in_polvio++;
222 ipstat.ips_delivered--;
223 /* do not inject data to pcb */
227 /* check AH/ESP integrity. */
228 if (last && ipsec4_in_reject(m, last)) {
230 ipstat.ips_delivered--;
231 /* do not inject data to pcb */
233 #endif /*FAST_IPSEC*/
234 /* Check the minimum TTL for socket. */
235 if (last && ip->ip_ttl < last->inp_ip_minttl) {
237 ipstat.ips_delivered--;
239 if (last->inp_flags & INP_CONTROLOPTS ||
240 last->inp_socket->so_options & SO_TIMESTAMP)
241 ip_savecontrol(last, &opts, ip, m);
242 lwkt_gettoken(&last->inp_socket->so_rcv.ssb_token);
243 if (ssb_appendaddr(&last->inp_socket->so_rcv,
244 (struct sockaddr *)&ripsrc, m, opts) == 0) {
249 sorwakeup(last->inp_socket);
251 lwkt_reltoken(&last->inp_socket->so_rcv.ssb_token);
254 ipstat.ips_noproto++;
255 ipstat.ips_delivered--;
257 lwkt_reltoken(&raw_token);
258 return(IPPROTO_DONE);
262 * Generate IP header and pass packet to ip_output.
263 * Tack on options user may have setup with control call.
266 rip_output(struct mbuf *m, struct socket *so, ...)
269 struct inpcb *inp = so->so_pcb;
271 int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST;
275 dst = __va_arg(ap, u_long);
279 * If the user handed us a complete IP packet, use it.
280 * Otherwise, allocate an mbuf for a header and fill it in.
282 if ((inp->inp_flags & INP_HDRINCL) == 0) {
283 if (m->m_pkthdr.len + sizeof(struct ip) > IP_MAXPACKET) {
287 M_PREPEND(m, sizeof(struct ip), MB_WAIT);
290 ip = mtod(m, struct ip *);
291 ip->ip_tos = inp->inp_ip_tos;
293 ip->ip_p = inp->inp_ip_p;
294 ip->ip_len = m->m_pkthdr.len;
295 ip->ip_src = inp->inp_laddr;
296 ip->ip_dst.s_addr = dst;
297 ip->ip_ttl = inp->inp_ip_ttl;
301 if (m->m_pkthdr.len > IP_MAXPACKET) {
305 if (m->m_len < sizeof(struct ip)) {
306 m = m_pullup(m, sizeof(struct ip));
310 ip = mtod(m, struct ip *);
311 hlen = IP_VHL_HL(ip->ip_vhl) << 2;
313 /* Don't allow header length less than the minimum. */
314 if (hlen < sizeof(struct ip)) {
320 * Don't allow both user specified and setsockopt options.
321 * Don't allow packet length sizes that will crash.
323 if ((hlen != sizeof(struct ip) && inp->inp_options) ||
324 ip->ip_len > m->m_pkthdr.len || ip->ip_len < hlen) {
329 ip->ip_id = ip_newid();
331 /* Prevent ip_output from overwriting header fields */
332 flags |= IP_RAWOUTPUT;
336 return ip_output(m, inp->inp_options, &inp->inp_route, flags,
337 inp->inp_moptions, inp);
341 * Raw IP socket option processing.
344 rip_ctloutput(netmsg_t msg)
346 struct socket *so = msg->base.nm_so;
347 struct sockopt *sopt = msg->ctloutput.nm_sopt;
348 struct inpcb *inp = so->so_pcb;
351 if (sopt->sopt_level != IPPROTO_IP) {
358 switch (sopt->sopt_dir) {
360 switch (sopt->sopt_name) {
362 optval = inp->inp_flags & INP_HDRINCL;
363 soopt_from_kbuf(sopt, &optval, sizeof optval);
366 case IP_FW_ADD: /* ADD actually returns the body... */
369 error = ip_fw_sockopt(sopt);
374 case IP_DUMMYNET_GET:
375 error = ip_dn_sockopt(sopt);
386 case MRT_API_SUPPORT:
388 case MRT_ADD_BW_UPCALL:
389 case MRT_DEL_BW_UPCALL:
390 error = ip_mrouter_get ? ip_mrouter_get(so, sopt) :
396 /* msg invalid now */
402 switch (sopt->sopt_name) {
404 error = soopt_to_kbuf(sopt, &optval, sizeof optval,
409 inp->inp_flags |= INP_HDRINCL;
411 inp->inp_flags &= ~INP_HDRINCL;
420 error = ip_fw_ctl_ptr(sopt);
425 case IP_DUMMYNET_CONFIGURE:
426 case IP_DUMMYNET_DEL:
427 case IP_DUMMYNET_FLUSH:
428 error = ip_dn_sockopt(sopt);
432 error = ip_rsvp_init(so);
436 error = ip_rsvp_done();
440 case IP_RSVP_VIF_OFF:
441 error = ip_rsvp_vif ?
442 ip_rsvp_vif(so, sopt) : EINVAL;
453 case MRT_API_SUPPORT:
455 case MRT_ADD_BW_UPCALL:
456 case MRT_DEL_BW_UPCALL:
457 error = ip_mrouter_set ? ip_mrouter_set(so, sopt) :
463 /* msg invalid now */
469 lwkt_replymsg(&msg->lmsg, error);
473 * This function exists solely to receive the PRC_IFDOWN messages which
474 * are sent by if_down(). It looks for an ifaddr whose ifa_addr is sa,
475 * and calls in_ifadown() to remove all routes corresponding to that address.
476 * It also receives the PRC_IFUP messages from if_up() and reinstalls the
480 rip_ctlinput(netmsg_t msg)
482 int cmd = msg->ctlinput.nm_cmd;
483 struct sockaddr *sa = msg->ctlinput.nm_arg;
484 struct in_ifaddr *ia;
485 struct in_ifaddr_container *iac;
492 TAILQ_FOREACH(iac, &in_ifaddrheads[mycpuid], ia_link) {
495 if (ia->ia_ifa.ifa_addr == sa &&
496 (ia->ia_flags & IFA_ROUTE)) {
498 * in_ifscrub kills the interface route.
500 in_ifscrub(ia->ia_ifp, ia);
502 * in_ifadown gets rid of all the rest of
503 * the routes. This is not quite the right
504 * thing to do, but at least if we are running
505 * a routing process they will come back.
507 in_ifadown(&ia->ia_ifa, 0);
515 TAILQ_FOREACH(iac, &in_ifaddrheads[mycpuid], ia_link) {
516 if (iac->ia->ia_ifa.ifa_addr == sa) {
521 if (ia == NULL || (ia->ia_flags & IFA_ROUTE))
524 ifp = ia->ia_ifa.ifa_ifp;
526 if ((ifp->if_flags & IFF_LOOPBACK) ||
527 (ifp->if_flags & IFF_POINTOPOINT))
530 err = rtinit(&ia->ia_ifa, RTM_ADD, flags);
532 ia->ia_flags |= IFA_ROUTE;
536 lwkt_replymsg(&msg->lmsg, 0);
539 u_long rip_sendspace = RIPSNDQ;
540 u_long rip_recvspace = RIPRCVQ;
542 SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW,
543 &rip_sendspace, 0, "Maximum outgoing raw IP datagram size");
544 SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW,
545 &rip_recvspace, 0, "Maximum incoming raw IP datagram size");
548 rip_attach(netmsg_t msg)
550 struct socket *so = msg->base.nm_so;
551 int proto = msg->attach.nm_proto;
552 struct pru_attach_info *ai = msg->attach.nm_ai;
559 error = priv_check_cred(ai->p_ucred, PRIV_NETINET_RAW, NULL_CRED_OKAY);
563 error = soreserve(so, rip_sendspace, rip_recvspace, ai->sb_rlimit);
567 lwkt_gettoken(&raw_token);
568 error = in_pcballoc(so, &ripcbinfo);
570 inp = (struct inpcb *)so->so_pcb;
571 inp->inp_vflag |= INP_IPV4;
572 inp->inp_ip_p = proto;
573 inp->inp_ip_ttl = ip_defttl;
575 lwkt_reltoken(&raw_token);
578 lwkt_replymsg(&msg->lmsg, error);
582 rip_detach(netmsg_t msg)
584 struct socket *so = msg->base.nm_so;
590 if (so == ip_mrouter && ip_mrouter_done)
592 if (ip_rsvp_force_done)
593 ip_rsvp_force_done(so);
597 lwkt_replymsg(&msg->lmsg, 0);
601 * NOTE: (so) is referenced from soabort*() and netmsg_pru_abort()
602 * will sofree() it when we return.
605 rip_abort(netmsg_t msg)
607 struct socket *so = msg->base.nm_so;
610 soisdisconnected(so);
611 if (so->so_state & SS_NOFDREF) { /* XXX not sure why this test */
613 /* msg invalid now */
617 lwkt_replymsg(&msg->lmsg, error);
621 rip_disconnect(netmsg_t msg)
623 struct socket *so = msg->base.nm_so;
626 if (so->so_state & SS_ISCONNECTED) {
629 /* msg invalid now */
634 lwkt_replymsg(&msg->lmsg, error);
638 rip_bind(netmsg_t msg)
640 struct socket *so = msg->base.nm_so;
641 struct sockaddr *nam = msg->bind.nm_nam;
642 struct inpcb *inp = so->so_pcb;
643 struct sockaddr_in *addr = (struct sockaddr_in *)nam;
646 if (nam->sa_len == sizeof(*addr)) {
647 if (TAILQ_EMPTY(&ifnet) ||
648 ((addr->sin_family != AF_INET) &&
649 (addr->sin_family != AF_IMPLINK)) ||
650 (addr->sin_addr.s_addr != INADDR_ANY &&
651 ifa_ifwithaddr((struct sockaddr *)addr) == 0)) {
652 error = EADDRNOTAVAIL;
654 inp->inp_laddr = addr->sin_addr;
660 lwkt_replymsg(&msg->lmsg, error);
664 rip_connect(netmsg_t msg)
666 struct socket *so = msg->base.nm_so;
667 struct sockaddr *nam = msg->connect.nm_nam;
668 struct inpcb *inp = so->so_pcb;
669 struct sockaddr_in *addr = (struct sockaddr_in *)nam;
672 if (nam->sa_len != sizeof(*addr)) {
674 } else if (TAILQ_EMPTY(&ifnet)) {
675 error = EADDRNOTAVAIL;
677 if ((addr->sin_family != AF_INET) &&
678 (addr->sin_family != AF_IMPLINK)) {
679 error = EAFNOSUPPORT;
681 inp->inp_faddr = addr->sin_addr;
686 lwkt_replymsg(&msg->lmsg, error);
690 rip_shutdown(netmsg_t msg)
692 socantsendmore(msg->base.nm_so);
693 lwkt_replymsg(&msg->lmsg, 0);
697 rip_send(netmsg_t msg)
699 struct socket *so = msg->base.nm_so;
700 struct mbuf *m = msg->send.nm_m;
701 /*struct mbuf *control = msg->send.nm_control;*/
702 struct sockaddr *nam = msg->send.nm_addr;
703 /*int flags = msg->send.nm_flags;*/
704 struct inpcb *inp = so->so_pcb;
708 if (so->so_state & SS_ISCONNECTED) {
713 dst = inp->inp_faddr.s_addr;
714 error = rip_output(m, so, dst);
721 dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr;
722 error = rip_output(m, so, dst);
725 lwkt_replymsg(&msg->lmsg, error);
728 SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, &ripcbinfo, 0,
729 in_pcblist_global, "S,xinpcb", "List of active raw IP sockets");
731 struct pr_usrreqs rip_usrreqs = {
732 .pru_abort = rip_abort,
733 .pru_accept = pr_generic_notsupp,
734 .pru_attach = rip_attach,
735 .pru_bind = rip_bind,
736 .pru_connect = rip_connect,
737 .pru_connect2 = pr_generic_notsupp,
738 .pru_control = in_control_dispatch,
739 .pru_detach = rip_detach,
740 .pru_disconnect = rip_disconnect,
741 .pru_listen = pr_generic_notsupp,
742 .pru_peeraddr = in_setpeeraddr_dispatch,
743 .pru_rcvd = pr_generic_notsupp,
744 .pru_rcvoob = pr_generic_notsupp,
745 .pru_send = rip_send,
746 .pru_sense = pru_sense_null,
747 .pru_shutdown = rip_shutdown,
748 .pru_sockaddr = in_setsockaddr_dispatch,
749 .pru_sosend = sosend,
750 .pru_soreceive = soreceive