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>
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_MP_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 void (*rsvp_input_p)(struct mbuf *m, ...);
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 *m, ...)
150 struct sockaddr_in ripsrc = { sizeof ripsrc, AF_INET };
151 struct ip *ip = mtod(m, struct ip *);
153 struct inpcb *last = NULL;
154 struct mbuf *opts = NULL;
159 off = __va_arg(ap, int);
160 proto = __va_arg(ap, int);
163 ripsrc.sin_addr = ip->ip_src;
164 lwkt_gettoken(&raw_token);
165 LIST_FOREACH(inp, &ripcbinfo.pcblisthead, inp_list) {
166 if (inp->inp_flags & INP_PLACEMARKER)
169 if ((inp->inp_vflag & INP_IPV4) == 0)
172 if (inp->inp_ip_p && inp->inp_ip_p != proto)
174 if (inp->inp_laddr.s_addr != INADDR_ANY &&
175 inp->inp_laddr.s_addr != ip->ip_dst.s_addr)
177 if (inp->inp_faddr.s_addr != INADDR_ANY &&
178 inp->inp_faddr.s_addr != ip->ip_src.s_addr)
181 struct mbuf *n = m_copypacket(m, MB_DONTWAIT);
184 /* check AH/ESP integrity. */
185 if (n && ipsec4_in_reject_so(n, last->inp_socket)) {
187 ipsecstat.in_polvio++;
188 /* do not inject data to pcb */
192 /* check AH/ESP integrity. */
193 if (ipsec4_in_reject(n, last)) {
195 /* do not inject data to pcb */
197 #endif /*FAST_IPSEC*/
199 lwkt_gettoken(&last->inp_socket->so_rcv.ssb_token);
200 if (last->inp_flags & INP_CONTROLOPTS ||
201 last->inp_socket->so_options & SO_TIMESTAMP)
202 ip_savecontrol(last, &opts, ip, n);
203 if (ssb_appendaddr(&last->inp_socket->so_rcv,
204 (struct sockaddr *)&ripsrc, n,
206 /* should notify about lost packet */
211 sorwakeup(last->inp_socket);
213 lwkt_reltoken(&last->inp_socket->so_rcv.ssb_token);
220 /* check AH/ESP integrity. */
221 if (last && ipsec4_in_reject_so(m, last->inp_socket)) {
223 ipsecstat.in_polvio++;
224 ipstat.ips_delivered--;
225 /* do not inject data to pcb */
229 /* check AH/ESP integrity. */
230 if (last && ipsec4_in_reject(m, last)) {
232 ipstat.ips_delivered--;
233 /* do not inject data to pcb */
235 #endif /*FAST_IPSEC*/
236 /* Check the minimum TTL for socket. */
237 if (last && ip->ip_ttl < last->inp_ip_minttl) {
239 ipstat.ips_delivered--;
241 if (last->inp_flags & INP_CONTROLOPTS ||
242 last->inp_socket->so_options & SO_TIMESTAMP)
243 ip_savecontrol(last, &opts, ip, m);
244 lwkt_gettoken(&last->inp_socket->so_rcv.ssb_token);
245 if (ssb_appendaddr(&last->inp_socket->so_rcv,
246 (struct sockaddr *)&ripsrc, m, opts) == 0) {
251 sorwakeup(last->inp_socket);
253 lwkt_reltoken(&last->inp_socket->so_rcv.ssb_token);
256 ipstat.ips_noproto++;
257 ipstat.ips_delivered--;
259 lwkt_reltoken(&raw_token);
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(struct socket *so, struct sockopt *sopt)
347 struct inpcb *inp = so->so_pcb;
350 if (sopt->sopt_level != IPPROTO_IP)
355 switch (sopt->sopt_dir) {
357 switch (sopt->sopt_name) {
359 optval = inp->inp_flags & INP_HDRINCL;
360 soopt_from_kbuf(sopt, &optval, sizeof optval);
363 case IP_FW_ADD: /* ADD actually returns the body... */
366 error = ip_fw_sockopt(sopt);
371 case IP_DUMMYNET_GET:
372 error = ip_dn_sockopt(sopt);
383 case MRT_API_SUPPORT:
385 case MRT_ADD_BW_UPCALL:
386 case MRT_DEL_BW_UPCALL:
387 error = ip_mrouter_get ? ip_mrouter_get(so, sopt) :
392 error = ip_ctloutput(so, sopt);
398 switch (sopt->sopt_name) {
400 error = soopt_to_kbuf(sopt, &optval, sizeof optval,
405 inp->inp_flags |= INP_HDRINCL;
407 inp->inp_flags &= ~INP_HDRINCL;
416 error = ip_fw_ctl_ptr(sopt);
421 case IP_DUMMYNET_CONFIGURE:
422 case IP_DUMMYNET_DEL:
423 case IP_DUMMYNET_FLUSH:
424 error = ip_dn_sockopt(sopt);
428 error = ip_rsvp_init(so);
432 error = ip_rsvp_done();
436 case IP_RSVP_VIF_OFF:
437 error = ip_rsvp_vif ?
438 ip_rsvp_vif(so, sopt) : EINVAL;
449 case MRT_API_SUPPORT:
451 case MRT_ADD_BW_UPCALL:
452 case MRT_DEL_BW_UPCALL:
453 error = ip_mrouter_set ? ip_mrouter_set(so, sopt) :
458 error = ip_ctloutput(so, sopt);
468 * This function exists solely to receive the PRC_IFDOWN messages which
469 * are sent by if_down(). It looks for an ifaddr whose ifa_addr is sa,
470 * and calls in_ifadown() to remove all routes corresponding to that address.
471 * It also receives the PRC_IFUP messages from if_up() and reinstalls the
475 rip_ctlinput(int cmd, struct sockaddr *sa, void *vip)
477 struct in_ifaddr *ia;
478 struct in_ifaddr_container *iac;
485 TAILQ_FOREACH(iac, &in_ifaddrheads[mycpuid], ia_link) {
488 if (ia->ia_ifa.ifa_addr == sa &&
489 (ia->ia_flags & IFA_ROUTE)) {
491 * in_ifscrub kills the interface route.
493 in_ifscrub(ia->ia_ifp, ia);
495 * in_ifadown gets rid of all the rest of
496 * the routes. This is not quite the right
497 * thing to do, but at least if we are running
498 * a routing process they will come back.
500 in_ifadown(&ia->ia_ifa, 0);
508 TAILQ_FOREACH(iac, &in_ifaddrheads[mycpuid], ia_link) {
509 if (iac->ia->ia_ifa.ifa_addr == sa) {
514 if (ia == NULL || (ia->ia_flags & IFA_ROUTE))
517 ifp = ia->ia_ifa.ifa_ifp;
519 if ((ifp->if_flags & IFF_LOOPBACK) ||
520 (ifp->if_flags & IFF_POINTOPOINT))
523 err = rtinit(&ia->ia_ifa, RTM_ADD, flags);
525 ia->ia_flags |= IFA_ROUTE;
530 u_long rip_sendspace = RIPSNDQ;
531 u_long rip_recvspace = RIPRCVQ;
533 SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW,
534 &rip_sendspace, 0, "Maximum outgoing raw IP datagram size");
535 SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW,
536 &rip_recvspace, 0, "Maximum incoming raw IP datagram size");
539 rip_attach(struct socket *so, int proto, struct pru_attach_info *ai)
547 error = priv_check_cred(ai->p_ucred, PRIV_NETINET_RAW, NULL_CRED_OKAY);
551 error = soreserve(so, rip_sendspace, rip_recvspace, ai->sb_rlimit);
554 lwkt_gettoken(&raw_token);
555 error = in_pcballoc(so, &ripcbinfo);
557 inp = (struct inpcb *)so->so_pcb;
558 inp->inp_vflag |= INP_IPV4;
559 inp->inp_ip_p = proto;
560 inp->inp_ip_ttl = ip_defttl;
562 lwkt_reltoken(&raw_token);
567 rip_detach(struct socket *so)
574 if (so == ip_mrouter && ip_mrouter_done)
576 if (ip_rsvp_force_done)
577 ip_rsvp_force_done(so);
585 * NOTE: (so) is referenced from soabort*() and netmsg_pru_abort()
586 * will sofree() it when we return.
589 rip_abort(struct socket *so)
593 soisdisconnected(so);
594 if (so->so_state & SS_NOFDREF) /* XXX not sure why this test */
595 error = rip_detach(so);
603 rip_disconnect(struct socket *so)
607 if ((so->so_state & SS_ISCONNECTED) == 0)
610 error = rip_abort(so);
617 rip_bind(struct socket *so, struct sockaddr *nam, struct thread *td)
619 struct inpcb *inp = so->so_pcb;
620 struct sockaddr_in *addr = (struct sockaddr_in *)nam;
622 if (nam->sa_len != sizeof(*addr))
625 if (TAILQ_EMPTY(&ifnet) || ((addr->sin_family != AF_INET) &&
626 (addr->sin_family != AF_IMPLINK)) ||
627 (addr->sin_addr.s_addr != INADDR_ANY &&
628 ifa_ifwithaddr((struct sockaddr *)addr) == 0))
629 return EADDRNOTAVAIL;
630 inp->inp_laddr = addr->sin_addr;
635 rip_connect(struct socket *so, struct sockaddr *nam, struct thread *td)
637 struct inpcb *inp = so->so_pcb;
638 struct sockaddr_in *addr = (struct sockaddr_in *)nam;
640 if (nam->sa_len != sizeof(*addr))
642 if (TAILQ_EMPTY(&ifnet))
643 return EADDRNOTAVAIL;
644 if ((addr->sin_family != AF_INET) &&
645 (addr->sin_family != AF_IMPLINK))
647 inp->inp_faddr = addr->sin_addr;
653 rip_shutdown(struct socket *so)
660 rip_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam,
661 struct mbuf *control, struct thread *td)
663 struct inpcb *inp = so->so_pcb;
666 if (so->so_state & SS_ISCONNECTED) {
671 dst = inp->inp_faddr.s_addr;
677 dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr;
679 return rip_output(m, so, dst);
682 SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, &ripcbinfo, 0,
683 in_pcblist_global, "S,xinpcb", "List of active raw IP sockets");
685 struct pr_usrreqs rip_usrreqs = {
686 .pru_abort = rip_abort,
687 .pru_accept = pru_accept_notsupp,
688 .pru_attach = rip_attach,
689 .pru_bind = rip_bind,
690 .pru_connect = rip_connect,
691 .pru_connect2 = pru_connect2_notsupp,
692 .pru_control = in_control,
693 .pru_detach = rip_detach,
694 .pru_disconnect = rip_disconnect,
695 .pru_listen = pru_listen_notsupp,
696 .pru_peeraddr = in_setpeeraddr,
697 .pru_rcvd = pru_rcvd_notsupp,
698 .pru_rcvoob = pru_rcvoob_notsupp,
699 .pru_send = rip_send,
700 .pru_sense = pru_sense_null,
701 .pru_shutdown = rip_shutdown,
702 .pru_sockaddr = in_setsockaddr,
703 .pru_sosend = sosend,
704 .pru_soreceive = soreceive