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 $
35 * $DragonFly: src/sys/netinet/raw_ip.c,v 1.33 2008/10/28 03:07:28 sephe Exp $
38 #include "opt_inet6.h"
39 #include "opt_ipsec.h"
41 #include <sys/param.h>
42 #include <sys/systm.h>
43 #include <sys/kernel.h>
45 #include <sys/malloc.h>
49 #include <sys/protosw.h>
50 #include <sys/socket.h>
51 #include <sys/socketvar.h>
52 #include <sys/sysctl.h>
54 #include <sys/thread2.h>
55 #include <sys/socketvar2.h>
56 #include <sys/msgport2.h>
58 #include <machine/stdarg.h>
61 #include <net/route.h>
64 #include <netinet/in.h>
65 #include <netinet/in_systm.h>
66 #include <netinet/ip.h>
67 #include <netinet/in_pcb.h>
68 #include <netinet/in_var.h>
69 #include <netinet/ip_var.h>
71 #include <net/ip_mroute/ip_mroute.h>
72 #include <net/ipfw/ip_fw.h>
73 #include <net/dummynet/ip_dummynet.h>
76 #include <netproto/ipsec/ipsec.h>
80 #include <netinet6/ipsec.h>
83 struct inpcbinfo ripcbinfo;
85 /* control hooks for ipfw and dummynet */
86 ip_fw_ctl_t *ip_fw_ctl_ptr;
87 ip_dn_ctl_t *ip_dn_ctl_ptr;
89 static struct lwkt_token raw_token = LWKT_TOKEN_INITIALIZER(raw_token);
93 * hooks for multicast routing. They all default to NULL,
94 * so leave them not initialized and rely on BSS being set to 0.
97 /* The socket used to communicate with the multicast routing daemon. */
98 struct socket *ip_mrouter;
100 /* The various mrouter and rsvp functions */
101 int (*ip_mrouter_set)(struct socket *, struct sockopt *);
102 int (*ip_mrouter_get)(struct socket *, struct sockopt *);
103 int (*ip_mrouter_done)(void);
104 int (*ip_mforward)(struct ip *, struct ifnet *, struct mbuf *,
105 struct ip_moptions *);
106 int (*mrt_ioctl)(int, caddr_t);
107 int (*legal_vif_num)(int);
108 u_long (*ip_mcast_src)(int);
110 int (*rsvp_input_p)(struct mbuf **, int *, int);
111 int (*ip_rsvp_vif)(struct socket *, struct sockopt *);
112 void (*ip_rsvp_force_done)(struct socket *);
115 * Nominal space allocated to a raw ip socket.
121 * Raw interface to IP protocol.
125 * Initialize raw connection block queue.
130 in_pcbinfo_init(&ripcbinfo);
132 * XXX We don't use the hash list for raw IP, but it's easier
133 * to allocate a one entry hash list than it is to check all
134 * over the place for hashbase == NULL.
136 ripcbinfo.hashbase = hashinit(1, M_PCB, &ripcbinfo.hashmask);
137 ripcbinfo.porthashbase = hashinit(1, M_PCB, &ripcbinfo.porthashmask);
138 ripcbinfo.wildcardhashbase = hashinit(1, M_PCB,
139 &ripcbinfo.wildcardhashmask);
140 ripcbinfo.ipi_size = sizeof(struct inpcb);
144 * Setup generic address and protocol structures
145 * for raw_input routine, then pass them along with
149 rip_input(struct mbuf **mp, int *offp, int proto)
151 struct sockaddr_in ripsrc = { sizeof ripsrc, AF_INET };
152 struct mbuf *m = *mp;
153 struct ip *ip = mtod(m, struct ip *);
155 struct inpcb *last = NULL;
156 struct mbuf *opts = NULL;
162 ripsrc.sin_addr = ip->ip_src;
163 lwkt_gettoken(&raw_token);
164 LIST_FOREACH(inp, &ripcbinfo.pcblisthead, inp_list) {
165 if (inp->inp_flags & INP_PLACEMARKER)
168 if ((inp->inp_vflag & INP_IPV4) == 0)
171 if (inp->inp_ip_p && inp->inp_ip_p != proto)
173 if (inp->inp_laddr.s_addr != INADDR_ANY &&
174 inp->inp_laddr.s_addr != ip->ip_dst.s_addr)
176 if (inp->inp_faddr.s_addr != INADDR_ANY &&
177 inp->inp_faddr.s_addr != ip->ip_src.s_addr)
180 struct mbuf *n = m_copypacket(m, MB_DONTWAIT);
183 /* check AH/ESP integrity. */
184 if (n && ipsec4_in_reject_so(n, last->inp_socket)) {
186 ipsecstat.in_polvio++;
187 /* do not inject data to pcb */
191 /* check AH/ESP integrity. */
192 if (ipsec4_in_reject(n, last)) {
194 /* do not inject data to pcb */
196 #endif /*FAST_IPSEC*/
198 lwkt_gettoken(&last->inp_socket->so_rcv.ssb_token);
199 if (last->inp_flags & INP_CONTROLOPTS ||
200 last->inp_socket->so_options & SO_TIMESTAMP)
201 ip_savecontrol(last, &opts, ip, n);
202 if (ssb_appendaddr(&last->inp_socket->so_rcv,
203 (struct sockaddr *)&ripsrc, n,
205 /* should notify about lost packet */
210 sorwakeup(last->inp_socket);
212 lwkt_reltoken(&last->inp_socket->so_rcv.ssb_token);
219 /* check AH/ESP integrity. */
220 if (last && ipsec4_in_reject_so(m, last->inp_socket)) {
222 ipsecstat.in_polvio++;
223 ipstat.ips_delivered--;
224 /* do not inject data to pcb */
228 /* check AH/ESP integrity. */
229 if (last && ipsec4_in_reject(m, last)) {
231 ipstat.ips_delivered--;
232 /* do not inject data to pcb */
234 #endif /*FAST_IPSEC*/
235 /* Check the minimum TTL for socket. */
236 if (last && ip->ip_ttl < last->inp_ip_minttl) {
238 ipstat.ips_delivered--;
240 if (last->inp_flags & INP_CONTROLOPTS ||
241 last->inp_socket->so_options & SO_TIMESTAMP)
242 ip_savecontrol(last, &opts, ip, m);
243 lwkt_gettoken(&last->inp_socket->so_rcv.ssb_token);
244 if (ssb_appendaddr(&last->inp_socket->so_rcv,
245 (struct sockaddr *)&ripsrc, m, opts) == 0) {
250 sorwakeup(last->inp_socket);
252 lwkt_reltoken(&last->inp_socket->so_rcv.ssb_token);
255 ipstat.ips_noproto++;
256 ipstat.ips_delivered--;
258 lwkt_reltoken(&raw_token);
259 return(IPPROTO_DONE);
263 * Generate IP header and pass packet to ip_output.
264 * Tack on options user may have setup with control call.
267 rip_output(struct mbuf *m, struct socket *so, ...)
270 struct inpcb *inp = so->so_pcb;
272 int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST;
276 dst = __va_arg(ap, u_long);
280 * If the user handed us a complete IP packet, use it.
281 * Otherwise, allocate an mbuf for a header and fill it in.
283 if ((inp->inp_flags & INP_HDRINCL) == 0) {
284 if (m->m_pkthdr.len + sizeof(struct ip) > IP_MAXPACKET) {
288 M_PREPEND(m, sizeof(struct ip), MB_WAIT);
291 ip = mtod(m, struct ip *);
292 ip->ip_tos = inp->inp_ip_tos;
294 ip->ip_p = inp->inp_ip_p;
295 ip->ip_len = m->m_pkthdr.len;
296 ip->ip_src = inp->inp_laddr;
297 ip->ip_dst.s_addr = dst;
298 ip->ip_ttl = inp->inp_ip_ttl;
302 if (m->m_pkthdr.len > IP_MAXPACKET) {
306 if (m->m_len < sizeof(struct ip)) {
307 m = m_pullup(m, sizeof(struct ip));
311 ip = mtod(m, struct ip *);
312 hlen = IP_VHL_HL(ip->ip_vhl) << 2;
314 /* Don't allow header length less than the minimum. */
315 if (hlen < sizeof(struct ip)) {
321 * Don't allow both user specified and setsockopt options.
322 * Don't allow packet length sizes that will crash.
324 if ((hlen != sizeof(struct ip) && inp->inp_options) ||
325 ip->ip_len > m->m_pkthdr.len || ip->ip_len < hlen) {
330 ip->ip_id = ip_newid();
332 /* Prevent ip_output from overwriting header fields */
333 flags |= IP_RAWOUTPUT;
337 return ip_output(m, inp->inp_options, &inp->inp_route, flags,
338 inp->inp_moptions, inp);
342 * Raw IP socket option processing.
345 rip_ctloutput(netmsg_t msg)
347 struct socket *so = msg->base.nm_so;
348 struct sockopt *sopt = msg->ctloutput.nm_sopt;
349 struct inpcb *inp = so->so_pcb;
352 if (sopt->sopt_level != IPPROTO_IP) {
359 switch (sopt->sopt_dir) {
361 switch (sopt->sopt_name) {
363 optval = inp->inp_flags & INP_HDRINCL;
364 soopt_from_kbuf(sopt, &optval, sizeof optval);
367 case IP_FW_ADD: /* ADD actually returns the body... */
370 error = ip_fw_sockopt(sopt);
375 case IP_DUMMYNET_GET:
376 error = ip_dn_sockopt(sopt);
387 case MRT_API_SUPPORT:
389 case MRT_ADD_BW_UPCALL:
390 case MRT_DEL_BW_UPCALL:
391 error = ip_mrouter_get ? ip_mrouter_get(so, sopt) :
397 /* msg invalid now */
403 switch (sopt->sopt_name) {
405 error = soopt_to_kbuf(sopt, &optval, sizeof optval,
410 inp->inp_flags |= INP_HDRINCL;
412 inp->inp_flags &= ~INP_HDRINCL;
421 error = ip_fw_ctl_ptr(sopt);
426 case IP_DUMMYNET_CONFIGURE:
427 case IP_DUMMYNET_DEL:
428 case IP_DUMMYNET_FLUSH:
429 error = ip_dn_sockopt(sopt);
433 error = ip_rsvp_init(so);
437 error = ip_rsvp_done();
441 case IP_RSVP_VIF_OFF:
442 error = ip_rsvp_vif ?
443 ip_rsvp_vif(so, sopt) : EINVAL;
454 case MRT_API_SUPPORT:
456 case MRT_ADD_BW_UPCALL:
457 case MRT_DEL_BW_UPCALL:
458 error = ip_mrouter_set ? ip_mrouter_set(so, sopt) :
464 /* msg invalid now */
470 lwkt_replymsg(&msg->lmsg, error);
474 * This function exists solely to receive the PRC_IFDOWN messages which
475 * are sent by if_down(). It looks for an ifaddr whose ifa_addr is sa,
476 * and calls in_ifadown() to remove all routes corresponding to that address.
477 * It also receives the PRC_IFUP messages from if_up() and reinstalls the
481 rip_ctlinput(netmsg_t msg)
483 int cmd = msg->ctlinput.nm_cmd;
484 struct sockaddr *sa = msg->ctlinput.nm_arg;
485 struct in_ifaddr *ia;
486 struct in_ifaddr_container *iac;
493 TAILQ_FOREACH(iac, &in_ifaddrheads[mycpuid], ia_link) {
496 if (ia->ia_ifa.ifa_addr == sa &&
497 (ia->ia_flags & IFA_ROUTE)) {
499 * in_ifscrub kills the interface route.
501 in_ifscrub(ia->ia_ifp, ia);
503 * in_ifadown gets rid of all the rest of
504 * the routes. This is not quite the right
505 * thing to do, but at least if we are running
506 * a routing process they will come back.
508 in_ifadown(&ia->ia_ifa, 0);
516 TAILQ_FOREACH(iac, &in_ifaddrheads[mycpuid], ia_link) {
517 if (iac->ia->ia_ifa.ifa_addr == sa) {
522 if (ia == NULL || (ia->ia_flags & IFA_ROUTE))
525 ifp = ia->ia_ifa.ifa_ifp;
527 if ((ifp->if_flags & IFF_LOOPBACK) ||
528 (ifp->if_flags & IFF_POINTOPOINT))
531 err = rtinit(&ia->ia_ifa, RTM_ADD, flags);
533 ia->ia_flags |= IFA_ROUTE;
537 lwkt_replymsg(&msg->lmsg, 0);
540 u_long rip_sendspace = RIPSNDQ;
541 u_long rip_recvspace = RIPRCVQ;
543 SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW,
544 &rip_sendspace, 0, "Maximum outgoing raw IP datagram size");
545 SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW,
546 &rip_recvspace, 0, "Maximum incoming raw IP datagram size");
549 rip_attach(netmsg_t msg)
551 struct socket *so = msg->base.nm_so;
552 int proto = msg->attach.nm_proto;
553 struct pru_attach_info *ai = msg->attach.nm_ai;
560 error = priv_check_cred(ai->p_ucred, PRIV_NETINET_RAW, NULL_CRED_OKAY);
564 error = soreserve(so, rip_sendspace, rip_recvspace, ai->sb_rlimit);
568 lwkt_gettoken(&raw_token);
569 error = in_pcballoc(so, &ripcbinfo);
571 inp = (struct inpcb *)so->so_pcb;
572 inp->inp_vflag |= INP_IPV4;
573 inp->inp_ip_p = proto;
574 inp->inp_ip_ttl = ip_defttl;
576 lwkt_reltoken(&raw_token);
579 lwkt_replymsg(&msg->lmsg, error);
583 rip_detach(netmsg_t msg)
585 struct socket *so = msg->base.nm_so;
591 if (so == ip_mrouter && ip_mrouter_done)
593 if (ip_rsvp_force_done)
594 ip_rsvp_force_done(so);
598 lwkt_replymsg(&msg->lmsg, 0);
602 * NOTE: (so) is referenced from soabort*() and netmsg_pru_abort()
603 * will sofree() it when we return.
606 rip_abort(netmsg_t msg)
608 struct socket *so = msg->base.nm_so;
611 soisdisconnected(so);
612 if (so->so_state & SS_NOFDREF) { /* XXX not sure why this test */
614 /* msg invalid now */
618 lwkt_replymsg(&msg->lmsg, error);
622 rip_disconnect(netmsg_t msg)
624 struct socket *so = msg->base.nm_so;
627 if (so->so_state & SS_ISCONNECTED) {
630 /* msg invalid now */
635 lwkt_replymsg(&msg->lmsg, error);
639 rip_bind(netmsg_t msg)
641 struct socket *so = msg->base.nm_so;
642 struct sockaddr *nam = msg->bind.nm_nam;
643 struct inpcb *inp = so->so_pcb;
644 struct sockaddr_in *addr = (struct sockaddr_in *)nam;
647 if (nam->sa_len == sizeof(*addr)) {
648 if (TAILQ_EMPTY(&ifnet) ||
649 ((addr->sin_family != AF_INET) &&
650 (addr->sin_family != AF_IMPLINK)) ||
651 (addr->sin_addr.s_addr != INADDR_ANY &&
652 ifa_ifwithaddr((struct sockaddr *)addr) == 0)) {
653 error = EADDRNOTAVAIL;
655 inp->inp_laddr = addr->sin_addr;
661 lwkt_replymsg(&msg->lmsg, error);
665 rip_connect(netmsg_t msg)
667 struct socket *so = msg->base.nm_so;
668 struct sockaddr *nam = msg->connect.nm_nam;
669 struct inpcb *inp = so->so_pcb;
670 struct sockaddr_in *addr = (struct sockaddr_in *)nam;
673 if (nam->sa_len != sizeof(*addr)) {
675 } else if (TAILQ_EMPTY(&ifnet)) {
676 error = EADDRNOTAVAIL;
678 if ((addr->sin_family != AF_INET) &&
679 (addr->sin_family != AF_IMPLINK)) {
680 error = EAFNOSUPPORT;
682 inp->inp_faddr = addr->sin_addr;
687 lwkt_replymsg(&msg->lmsg, error);
691 rip_shutdown(netmsg_t msg)
693 socantsendmore(msg->base.nm_so);
694 lwkt_replymsg(&msg->lmsg, 0);
698 rip_send(netmsg_t msg)
700 struct socket *so = msg->base.nm_so;
701 struct mbuf *m = msg->send.nm_m;
702 /*struct mbuf *control = msg->send.nm_control;*/
703 struct sockaddr *nam = msg->send.nm_addr;
704 /*int flags = msg->send.nm_flags;*/
705 struct inpcb *inp = so->so_pcb;
709 if (so->so_state & SS_ISCONNECTED) {
714 dst = inp->inp_faddr.s_addr;
715 error = rip_output(m, so, dst);
722 dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr;
723 error = rip_output(m, so, dst);
726 lwkt_replymsg(&msg->lmsg, error);
729 SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, &ripcbinfo, 0,
730 in_pcblist_global, "S,xinpcb", "List of active raw IP sockets");
732 struct pr_usrreqs rip_usrreqs = {
733 .pru_abort = rip_abort,
734 .pru_accept = pr_generic_notsupp,
735 .pru_attach = rip_attach,
736 .pru_bind = rip_bind,
737 .pru_connect = rip_connect,
738 .pru_connect2 = pr_generic_notsupp,
739 .pru_control = in_control_dispatch,
740 .pru_detach = rip_detach,
741 .pru_disconnect = rip_disconnect,
742 .pru_listen = pr_generic_notsupp,
743 .pru_peeraddr = in_setpeeraddr_dispatch,
744 .pru_rcvd = pr_generic_notsupp,
745 .pru_rcvoob = pr_generic_notsupp,
746 .pru_send = rip_send,
747 .pru_sense = pru_sense_null,
748 .pru_shutdown = rip_shutdown,
749 .pru_sockaddr = in_setsockaddr_dispatch,
750 .pru_sosend = sosend,
751 .pru_soreceive = soreceive