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"
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>
62 #include <net/if_types.h>
64 #include <net/route.h>
67 #include <netinet/in.h>
68 #include <netinet/in_systm.h>
69 #include <netinet/ip.h>
70 #include <netinet/in_pcb.h>
71 #include <netinet/in_var.h>
72 #include <netinet/ip_var.h>
74 #include <net/ip_mroute/ip_mroute.h>
75 #include <net/ipfw/ip_fw.h>
76 #include <net/dummynet/ip_dummynet.h>
79 #include <netproto/ipsec/ipsec.h>
83 #include <netinet6/ipsec.h>
86 struct inpcbinfo ripcbinfo
;
88 /* control hooks for ipfw and dummynet */
89 ip_fw_ctl_t
*ip_fw_ctl_ptr
;
90 ip_dn_ctl_t
*ip_dn_ctl_ptr
;
92 static struct lwkt_token raw_token
= LWKT_TOKEN_INITIALIZER(raw_token
);
96 * hooks for multicast routing. They all default to NULL,
97 * so leave them not initialized and rely on BSS being set to 0.
100 /* The socket used to communicate with the multicast routing daemon. */
101 struct socket
*ip_mrouter
;
103 /* The various mrouter and rsvp functions */
104 int (*ip_mrouter_set
)(struct socket
*, struct sockopt
*);
105 int (*ip_mrouter_get
)(struct socket
*, struct sockopt
*);
106 int (*ip_mrouter_done
)(void);
107 int (*ip_mforward
)(struct ip
*, struct ifnet
*, struct mbuf
*,
108 struct ip_moptions
*);
109 int (*mrt_ioctl
)(int, caddr_t
);
110 int (*legal_vif_num
)(int);
111 u_long (*ip_mcast_src
)(int);
113 int (*rsvp_input_p
)(struct mbuf
**, int *, int);
114 int (*ip_rsvp_vif
)(struct socket
*, struct sockopt
*);
115 void (*ip_rsvp_force_done
)(struct socket
*);
118 * Nominal space allocated to a raw ip socket.
124 * Raw interface to IP protocol.
128 * Initialize raw connection block queue.
133 in_pcbinfo_init(&ripcbinfo
);
135 * XXX We don't use the hash list for raw IP, but it's easier
136 * to allocate a one entry hash list than it is to check all
137 * over the place for hashbase == NULL.
139 ripcbinfo
.hashbase
= hashinit(1, M_PCB
, &ripcbinfo
.hashmask
);
140 ripcbinfo
.porthashbase
= hashinit(1, M_PCB
, &ripcbinfo
.porthashmask
);
141 ripcbinfo
.wildcardhashbase
= hashinit(1, M_PCB
,
142 &ripcbinfo
.wildcardhashmask
);
143 ripcbinfo
.ipi_size
= sizeof(struct inpcb
);
147 * Setup generic address and protocol structures
148 * for raw_input routine, then pass them along with
152 rip_input(struct mbuf
**mp
, int *offp
, int proto
)
154 struct sockaddr_in ripsrc
= { sizeof ripsrc
, AF_INET
};
155 struct mbuf
*m
= *mp
;
156 struct ip
*ip
= mtod(m
, struct ip
*);
158 struct inpcb
*last
= NULL
;
159 struct mbuf
*opts
= NULL
;
165 ripsrc
.sin_addr
= ip
->ip_src
;
166 lwkt_gettoken(&raw_token
);
167 LIST_FOREACH(inp
, &ripcbinfo
.pcblisthead
, inp_list
) {
168 if (inp
->inp_flags
& INP_PLACEMARKER
)
171 if ((inp
->inp_vflag
& INP_IPV4
) == 0)
174 if (inp
->inp_ip_p
&& inp
->inp_ip_p
!= proto
)
176 if (inp
->inp_laddr
.s_addr
!= INADDR_ANY
&&
177 inp
->inp_laddr
.s_addr
!= ip
->ip_dst
.s_addr
)
179 if (inp
->inp_faddr
.s_addr
!= INADDR_ANY
&&
180 inp
->inp_faddr
.s_addr
!= ip
->ip_src
.s_addr
)
183 struct mbuf
*n
= m_copypacket(m
, MB_DONTWAIT
);
186 /* check AH/ESP integrity. */
187 if (n
&& ipsec4_in_reject_so(n
, last
->inp_socket
)) {
189 ipsecstat
.in_polvio
++;
190 /* do not inject data to pcb */
194 /* check AH/ESP integrity. */
195 if (ipsec4_in_reject(n
, last
)) {
197 /* do not inject data to pcb */
199 #endif /*FAST_IPSEC*/
201 lwkt_gettoken(&last
->inp_socket
->so_rcv
.ssb_token
);
202 if (last
->inp_flags
& INP_CONTROLOPTS
||
203 last
->inp_socket
->so_options
& SO_TIMESTAMP
)
204 ip_savecontrol(last
, &opts
, ip
, n
);
205 if (ssb_appendaddr(&last
->inp_socket
->so_rcv
,
206 (struct sockaddr
*)&ripsrc
, n
,
208 /* should notify about lost packet */
213 sorwakeup(last
->inp_socket
);
215 lwkt_reltoken(&last
->inp_socket
->so_rcv
.ssb_token
);
222 /* check AH/ESP integrity. */
223 if (last
&& ipsec4_in_reject_so(m
, last
->inp_socket
)) {
225 ipsecstat
.in_polvio
++;
226 ipstat
.ips_delivered
--;
227 /* do not inject data to pcb */
231 /* check AH/ESP integrity. */
232 if (last
&& ipsec4_in_reject(m
, last
)) {
234 ipstat
.ips_delivered
--;
235 /* do not inject data to pcb */
237 #endif /*FAST_IPSEC*/
238 /* Check the minimum TTL for socket. */
239 if (last
&& ip
->ip_ttl
< last
->inp_ip_minttl
) {
241 ipstat
.ips_delivered
--;
243 if (last
->inp_flags
& INP_CONTROLOPTS
||
244 last
->inp_socket
->so_options
& SO_TIMESTAMP
)
245 ip_savecontrol(last
, &opts
, ip
, m
);
246 lwkt_gettoken(&last
->inp_socket
->so_rcv
.ssb_token
);
247 if (ssb_appendaddr(&last
->inp_socket
->so_rcv
,
248 (struct sockaddr
*)&ripsrc
, m
, opts
) == 0) {
253 sorwakeup(last
->inp_socket
);
255 lwkt_reltoken(&last
->inp_socket
->so_rcv
.ssb_token
);
258 ipstat
.ips_noproto
++;
259 ipstat
.ips_delivered
--;
261 lwkt_reltoken(&raw_token
);
262 return(IPPROTO_DONE
);
266 * Generate IP header and pass packet to ip_output.
267 * Tack on options user may have setup with control call.
270 rip_output(struct mbuf
*m
, struct socket
*so
, ...)
273 struct inpcb
*inp
= so
->so_pcb
;
275 int flags
= (so
->so_options
& SO_DONTROUTE
) | IP_ALLOWBROADCAST
;
279 dst
= __va_arg(ap
, u_long
);
283 * If the user handed us a complete IP packet, use it.
284 * Otherwise, allocate an mbuf for a header and fill it in.
286 if ((inp
->inp_flags
& INP_HDRINCL
) == 0) {
287 if (m
->m_pkthdr
.len
+ sizeof(struct ip
) > IP_MAXPACKET
) {
291 M_PREPEND(m
, sizeof(struct ip
), MB_WAIT
);
294 ip
= mtod(m
, struct ip
*);
295 ip
->ip_tos
= inp
->inp_ip_tos
;
297 ip
->ip_p
= inp
->inp_ip_p
;
298 ip
->ip_len
= m
->m_pkthdr
.len
;
299 ip
->ip_src
= inp
->inp_laddr
;
300 ip
->ip_dst
.s_addr
= dst
;
301 ip
->ip_ttl
= inp
->inp_ip_ttl
;
305 if (m
->m_pkthdr
.len
> IP_MAXPACKET
) {
309 if (m
->m_len
< sizeof(struct ip
)) {
310 m
= m_pullup(m
, sizeof(struct ip
));
314 ip
= mtod(m
, struct ip
*);
315 hlen
= IP_VHL_HL(ip
->ip_vhl
) << 2;
317 /* Don't allow header length less than the minimum. */
318 if (hlen
< sizeof(struct ip
)) {
324 * Don't allow both user specified and setsockopt options.
325 * Don't allow packet length sizes that will crash.
327 if ((hlen
!= sizeof(struct ip
) && inp
->inp_options
) ||
328 ip
->ip_len
> m
->m_pkthdr
.len
|| ip
->ip_len
< hlen
) {
333 ip
->ip_id
= ip_newid();
335 /* Prevent ip_output from overwriting header fields */
336 flags
|= IP_RAWOUTPUT
;
340 return ip_output(m
, inp
->inp_options
, &inp
->inp_route
, flags
,
341 inp
->inp_moptions
, inp
);
345 * Raw IP socket option processing.
348 rip_ctloutput(netmsg_t msg
)
350 struct socket
*so
= msg
->base
.nm_so
;
351 struct sockopt
*sopt
= msg
->ctloutput
.nm_sopt
;
352 struct inpcb
*inp
= so
->so_pcb
;
355 if (sopt
->sopt_level
!= IPPROTO_IP
) {
362 switch (sopt
->sopt_dir
) {
364 switch (sopt
->sopt_name
) {
366 optval
= inp
->inp_flags
& INP_HDRINCL
;
367 soopt_from_kbuf(sopt
, &optval
, sizeof optval
);
370 case IP_FW_ADD
: /* ADD actually returns the body... */
373 error
= ip_fw_sockopt(sopt
);
378 case IP_DUMMYNET_GET
:
379 error
= ip_dn_sockopt(sopt
);
390 case MRT_API_SUPPORT
:
392 case MRT_ADD_BW_UPCALL
:
393 case MRT_DEL_BW_UPCALL
:
394 error
= ip_mrouter_get
? ip_mrouter_get(so
, sopt
) :
400 /* msg invalid now */
406 switch (sopt
->sopt_name
) {
408 error
= soopt_to_kbuf(sopt
, &optval
, sizeof optval
,
413 inp
->inp_flags
|= INP_HDRINCL
;
415 inp
->inp_flags
&= ~INP_HDRINCL
;
424 error
= ip_fw_ctl_ptr(sopt
);
429 case IP_DUMMYNET_CONFIGURE
:
430 case IP_DUMMYNET_DEL
:
431 case IP_DUMMYNET_FLUSH
:
432 error
= ip_dn_sockopt(sopt
);
436 error
= ip_rsvp_init(so
);
440 error
= ip_rsvp_done();
444 case IP_RSVP_VIF_OFF
:
445 error
= ip_rsvp_vif
?
446 ip_rsvp_vif(so
, sopt
) : EINVAL
;
457 case MRT_API_SUPPORT
:
459 case MRT_ADD_BW_UPCALL
:
460 case MRT_DEL_BW_UPCALL
:
461 error
= ip_mrouter_set
? ip_mrouter_set(so
, sopt
) :
467 /* msg invalid now */
473 lwkt_replymsg(&msg
->lmsg
, error
);
477 * This function exists solely to receive the PRC_IFDOWN messages which
478 * are sent by if_down(). It looks for an ifaddr whose ifa_addr is sa,
479 * and calls in_ifadown() to remove all routes corresponding to that address.
480 * It also receives the PRC_IFUP messages from if_up() and reinstalls the
484 rip_ctlinput(netmsg_t msg
)
486 int cmd
= msg
->ctlinput
.nm_cmd
;
487 struct sockaddr
*sa
= msg
->ctlinput
.nm_arg
;
488 struct in_ifaddr
*ia
;
489 struct in_ifaddr_container
*iac
;
496 TAILQ_FOREACH(iac
, &in_ifaddrheads
[mycpuid
], ia_link
) {
499 if (ia
->ia_ifa
.ifa_addr
== sa
&&
500 (ia
->ia_flags
& IFA_ROUTE
)) {
502 * in_ifscrub kills the interface route.
504 in_ifscrub(ia
->ia_ifp
, ia
);
506 * in_ifadown gets rid of all the rest of
507 * the routes. This is not quite the right
508 * thing to do, but at least if we are running
509 * a routing process they will come back.
511 in_ifadown(&ia
->ia_ifa
, 0);
519 TAILQ_FOREACH(iac
, &in_ifaddrheads
[mycpuid
], ia_link
) {
520 if (iac
->ia
->ia_ifa
.ifa_addr
== sa
) {
525 if (ia
== NULL
|| (ia
->ia_flags
& IFA_ROUTE
))
528 ifp
= ia
->ia_ifa
.ifa_ifp
;
532 * Don't add prefix routes for CARP interfaces.
533 * Prefix routes creation is handled by CARP
534 * interfaces themselves.
536 if (ifp
->if_type
== IFT_CARP
)
540 if ((ifp
->if_flags
& IFF_LOOPBACK
) ||
541 (ifp
->if_flags
& IFF_POINTOPOINT
))
544 err
= rtinit(&ia
->ia_ifa
, RTM_ADD
, flags
);
546 ia
->ia_flags
|= IFA_ROUTE
;
550 lwkt_replymsg(&msg
->lmsg
, 0);
553 u_long rip_sendspace
= RIPSNDQ
;
554 u_long rip_recvspace
= RIPRCVQ
;
556 SYSCTL_INT(_net_inet_raw
, OID_AUTO
, maxdgram
, CTLFLAG_RW
,
557 &rip_sendspace
, 0, "Maximum outgoing raw IP datagram size");
558 SYSCTL_INT(_net_inet_raw
, OID_AUTO
, recvspace
, CTLFLAG_RW
,
559 &rip_recvspace
, 0, "Maximum incoming raw IP datagram size");
562 rip_attach(netmsg_t msg
)
564 struct socket
*so
= msg
->base
.nm_so
;
565 int proto
= msg
->attach
.nm_proto
;
566 struct pru_attach_info
*ai
= msg
->attach
.nm_ai
;
573 error
= priv_check_cred(ai
->p_ucred
, PRIV_NETINET_RAW
, NULL_CRED_OKAY
);
577 error
= soreserve(so
, rip_sendspace
, rip_recvspace
, ai
->sb_rlimit
);
581 lwkt_gettoken(&raw_token
);
582 error
= in_pcballoc(so
, &ripcbinfo
);
584 inp
= (struct inpcb
*)so
->so_pcb
;
585 inp
->inp_vflag
|= INP_IPV4
;
586 inp
->inp_ip_p
= proto
;
587 inp
->inp_ip_ttl
= ip_defttl
;
589 lwkt_reltoken(&raw_token
);
592 lwkt_replymsg(&msg
->lmsg
, error
);
596 rip_detach(netmsg_t msg
)
598 struct socket
*so
= msg
->base
.nm_so
;
604 if (so
== ip_mrouter
&& ip_mrouter_done
)
606 if (ip_rsvp_force_done
)
607 ip_rsvp_force_done(so
);
611 lwkt_replymsg(&msg
->lmsg
, 0);
615 * NOTE: (so) is referenced from soabort*() and netmsg_pru_abort()
616 * will sofree() it when we return.
619 rip_abort(netmsg_t msg
)
621 struct socket
*so
= msg
->base
.nm_so
;
624 soisdisconnected(so
);
625 if (so
->so_state
& SS_NOFDREF
) { /* XXX not sure why this test */
627 /* msg invalid now */
631 lwkt_replymsg(&msg
->lmsg
, error
);
635 rip_disconnect(netmsg_t msg
)
637 struct socket
*so
= msg
->base
.nm_so
;
640 if (so
->so_state
& SS_ISCONNECTED
) {
643 /* msg invalid now */
648 lwkt_replymsg(&msg
->lmsg
, error
);
652 rip_bind(netmsg_t msg
)
654 struct socket
*so
= msg
->base
.nm_so
;
655 struct sockaddr
*nam
= msg
->bind
.nm_nam
;
656 struct inpcb
*inp
= so
->so_pcb
;
657 struct sockaddr_in
*addr
= (struct sockaddr_in
*)nam
;
660 if (nam
->sa_len
== sizeof(*addr
)) {
661 if (TAILQ_EMPTY(&ifnet
) ||
662 ((addr
->sin_family
!= AF_INET
) &&
663 (addr
->sin_family
!= AF_IMPLINK
)) ||
664 (addr
->sin_addr
.s_addr
!= INADDR_ANY
&&
665 ifa_ifwithaddr((struct sockaddr
*)addr
) == 0)) {
666 error
= EADDRNOTAVAIL
;
668 inp
->inp_laddr
= addr
->sin_addr
;
674 lwkt_replymsg(&msg
->lmsg
, error
);
678 rip_connect(netmsg_t msg
)
680 struct socket
*so
= msg
->base
.nm_so
;
681 struct sockaddr
*nam
= msg
->connect
.nm_nam
;
682 struct inpcb
*inp
= so
->so_pcb
;
683 struct sockaddr_in
*addr
= (struct sockaddr_in
*)nam
;
686 if (nam
->sa_len
!= sizeof(*addr
)) {
688 } else if (TAILQ_EMPTY(&ifnet
)) {
689 error
= EADDRNOTAVAIL
;
691 if ((addr
->sin_family
!= AF_INET
) &&
692 (addr
->sin_family
!= AF_IMPLINK
)) {
693 error
= EAFNOSUPPORT
;
695 inp
->inp_faddr
= addr
->sin_addr
;
700 lwkt_replymsg(&msg
->lmsg
, error
);
704 rip_shutdown(netmsg_t msg
)
706 socantsendmore(msg
->base
.nm_so
);
707 lwkt_replymsg(&msg
->lmsg
, 0);
711 rip_send(netmsg_t msg
)
713 struct socket
*so
= msg
->base
.nm_so
;
714 struct mbuf
*m
= msg
->send
.nm_m
;
715 /*struct mbuf *control = msg->send.nm_control;*/
716 struct sockaddr
*nam
= msg
->send
.nm_addr
;
717 /*int flags = msg->send.nm_flags;*/
718 struct inpcb
*inp
= so
->so_pcb
;
722 if (so
->so_state
& SS_ISCONNECTED
) {
727 dst
= inp
->inp_faddr
.s_addr
;
728 error
= rip_output(m
, so
, dst
);
735 dst
= ((struct sockaddr_in
*)nam
)->sin_addr
.s_addr
;
736 error
= rip_output(m
, so
, dst
);
739 lwkt_replymsg(&msg
->lmsg
, error
);
742 SYSCTL_PROC(_net_inet_raw
, OID_AUTO
/*XXX*/, pcblist
, CTLFLAG_RD
, &ripcbinfo
, 0,
743 in_pcblist_global
, "S,xinpcb", "List of active raw IP sockets");
745 struct pr_usrreqs rip_usrreqs
= {
746 .pru_abort
= rip_abort
,
747 .pru_accept
= pr_generic_notsupp
,
748 .pru_attach
= rip_attach
,
749 .pru_bind
= rip_bind
,
750 .pru_connect
= rip_connect
,
751 .pru_connect2
= pr_generic_notsupp
,
752 .pru_control
= in_control_dispatch
,
753 .pru_detach
= rip_detach
,
754 .pru_disconnect
= rip_disconnect
,
755 .pru_listen
= pr_generic_notsupp
,
756 .pru_peeraddr
= in_setpeeraddr_dispatch
,
757 .pru_rcvd
= pr_generic_notsupp
,
758 .pru_rcvoob
= pr_generic_notsupp
,
759 .pru_send
= rip_send
,
760 .pru_sense
= pru_sense_null
,
761 .pru_shutdown
= rip_shutdown
,
762 .pru_sockaddr
= in_setsockaddr_dispatch
,
763 .pru_sosend
= sosend
,
764 .pru_soreceive
= soreceive