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/ipfw3/ip_fw.h>
73 #include <net/dummynet/ip_dummynet.h>
74 #include <net/dummynet3/ip_dummynet3.h>
77 #include <netproto/ipsec/ipsec.h>
81 #include <netinet6/ipsec.h>
84 struct inpcbinfo ripcbinfo
;
85 struct inpcbportinfo ripcbportinfo
;
87 /* control hooks for ipfw and dummynet */
88 ip_fw_ctl_t
*ip_fw_ctl_ptr
;
89 ip_fw_ctl_t
*ip_fw_ctl_x_ptr
;
90 ip_dn_ctl_t
*ip_dn_ctl_ptr
;
94 * hooks for multicast routing. They all default to NULL,
95 * so leave them not initialized and rely on BSS being set to 0.
98 /* The socket used to communicate with the multicast routing daemon. */
99 struct socket
*ip_mrouter
;
101 /* The various mrouter and rsvp functions */
102 int (*ip_mrouter_set
)(struct socket
*, struct sockopt
*);
103 int (*ip_mrouter_get
)(struct socket
*, struct sockopt
*);
104 int (*ip_mrouter_done
)(void);
105 int (*ip_mforward
)(struct ip
*, struct ifnet
*, struct mbuf
*,
106 struct ip_moptions
*);
107 int (*mrt_ioctl
)(u_long
, caddr_t
);
108 int (*legal_vif_num
)(int);
109 u_long (*ip_mcast_src
)(int);
111 int (*rsvp_input_p
)(struct mbuf
**, int *, int);
112 int (*ip_rsvp_vif
)(struct socket
*, struct sockopt
*);
113 void (*ip_rsvp_force_done
)(struct socket
*);
116 * Nominal space allocated to a raw ip socket.
122 * Raw interface to IP protocol.
126 * Initialize raw connection block queue.
131 in_pcbinfo_init(&ripcbinfo
, 0, FALSE
);
132 in_pcbportinfo_init(&ripcbportinfo
, 1, 0);
134 * XXX We don't use the hash list for raw IP, but it's easier
135 * to allocate a one entry hash list than it is to check all
136 * over the place for hashbase == NULL.
138 ripcbinfo
.hashbase
= hashinit(1, M_PCB
, &ripcbinfo
.hashmask
);
139 in_pcbportinfo_set(&ripcbinfo
, &ripcbportinfo
, 1);
140 ripcbinfo
.wildcardhashbase
= hashinit(1, M_PCB
,
141 &ripcbinfo
.wildcardhashmask
);
142 ripcbinfo
.ipi_size
= sizeof(struct inpcb
);
146 * Setup generic address and protocol structures
147 * for raw_input routine, then pass them along with
151 rip_input(struct mbuf
**mp
, int *offp
, int proto
)
153 struct sockaddr_in ripsrc
= { sizeof ripsrc
, AF_INET
};
154 struct mbuf
*m
= *mp
;
155 struct ip
*ip
= mtod(m
, struct ip
*);
157 struct inpcb
*last
= NULL
;
158 struct mbuf
*opts
= NULL
;
164 ripsrc
.sin_addr
= ip
->ip_src
;
165 LIST_FOREACH(inp
, &ripcbinfo
.pcblisthead
, inp_list
) {
166 if (inp
->inp_flags
& INP_PLACEMARKER
)
169 if (!INP_ISIPV4(inp
))
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
, M_NOWAIT
);
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 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
;
278 dst
= __va_arg(ap
, u_long
);
282 * If the user handed us a complete IP packet, use it.
283 * Otherwise, allocate an mbuf for a header and fill it in.
285 if ((inp
->inp_flags
& INP_HDRINCL
) == 0) {
286 if (m
->m_pkthdr
.len
+ sizeof(struct ip
) > IP_MAXPACKET
) {
290 M_PREPEND(m
, sizeof(struct ip
), M_WAITOK
);
293 ip
= mtod(m
, struct ip
*);
294 ip
->ip_tos
= inp
->inp_ip_tos
;
296 ip
->ip_p
= inp
->inp_ip_p
;
297 ip
->ip_len
= m
->m_pkthdr
.len
;
298 ip
->ip_src
= inp
->inp_laddr
;
299 ip
->ip_dst
.s_addr
= dst
;
300 ip
->ip_ttl
= inp
->inp_ip_ttl
;
304 if (m
->m_pkthdr
.len
> IP_MAXPACKET
) {
308 if (m
->m_len
< sizeof(struct ip
)) {
309 m
= m_pullup(m
, sizeof(struct ip
));
313 ip
= mtod(m
, struct ip
*);
314 hlen
= IP_VHL_HL(ip
->ip_vhl
) << 2;
316 /* Don't allow header length less than the minimum. */
317 if (hlen
< sizeof(struct ip
)) {
323 * Don't allow both user specified and setsockopt options.
324 * Don't allow packet length sizes that will crash.
326 if ((hlen
!= sizeof(struct ip
) && inp
->inp_options
) ||
327 ip
->ip_len
> m
->m_pkthdr
.len
|| ip
->ip_len
< hlen
) {
332 ip
->ip_id
= ip_newid();
334 /* Prevent ip_output from overwriting header fields */
335 flags
|= IP_RAWOUTPUT
;
339 return ip_output(m
, inp
->inp_options
, &inp
->inp_route
, flags
,
340 inp
->inp_moptions
, inp
);
344 * Raw IP socket option processing.
347 rip_ctloutput(netmsg_t msg
)
349 struct socket
*so
= msg
->base
.nm_so
;
350 struct sockopt
*sopt
= msg
->ctloutput
.nm_sopt
;
351 struct inpcb
*inp
= so
->so_pcb
;
358 /* Get socket's owner cpuid hint */
359 if (sopt
->sopt_level
== SOL_SOCKET
&&
360 sopt
->sopt_dir
== SOPT_GET
&&
361 sopt
->sopt_name
== SO_CPUHINT
) {
363 soopt_from_kbuf(sopt
, &optval
, sizeof(optval
));
367 if (sopt
->sopt_level
!= IPPROTO_IP
) {
372 switch (sopt
->sopt_dir
) {
374 switch (sopt
->sopt_name
) {
376 optval
= inp
->inp_flags
& INP_HDRINCL
;
377 soopt_from_kbuf(sopt
, &optval
, sizeof optval
);
381 error
= ip_fw3_sockopt(sopt
);
385 case IP_FW_ADD
: /* ADD actually returns the body... */
388 error
= ip_fw_sockopt(sopt
);
393 case IP_DUMMYNET_GET
:
394 error
= ip_dn_sockopt(sopt
);
405 case MRT_API_SUPPORT
:
407 case MRT_ADD_BW_UPCALL
:
408 case MRT_DEL_BW_UPCALL
:
409 error
= ip_mrouter_get
? ip_mrouter_get(so
, sopt
) :
415 /* msg invalid now */
421 switch (sopt
->sopt_name
) {
423 error
= soopt_to_kbuf(sopt
, &optval
, sizeof optval
,
428 inp
->inp_flags
|= INP_HDRINCL
;
430 inp
->inp_flags
&= ~INP_HDRINCL
;
434 error
= ip_fw_ctl_x_ptr(sopt
);
444 error
= ip_fw_ctl_ptr(sopt
);
449 case IP_DUMMYNET_CONFIGURE
:
450 case IP_DUMMYNET_DEL
:
451 case IP_DUMMYNET_FLUSH
:
452 error
= ip_dn_sockopt(sopt
);
456 error
= ip_rsvp_init(so
);
460 error
= ip_rsvp_done();
464 case IP_RSVP_VIF_OFF
:
465 error
= ip_rsvp_vif
?
466 ip_rsvp_vif(so
, sopt
) : EINVAL
;
477 case MRT_API_SUPPORT
:
479 case MRT_ADD_BW_UPCALL
:
480 case MRT_DEL_BW_UPCALL
:
481 error
= ip_mrouter_set
? ip_mrouter_set(so
, sopt
) :
487 /* msg invalid now */
493 lwkt_replymsg(&msg
->lmsg
, error
);
497 * This function exists solely to receive the PRC_IFDOWN messages which
498 * are sent by if_down(). It looks for an ifaddr whose ifa_addr is sa,
499 * and calls in_ifadown() to remove all routes corresponding to that address.
500 * It also receives the PRC_IFUP messages from if_up() and reinstalls the
504 rip_ctlinput(netmsg_t msg
)
506 int cmd
= msg
->ctlinput
.nm_cmd
;
507 struct sockaddr
*sa
= msg
->ctlinput
.nm_arg
;
508 struct in_ifaddr
*ia
;
509 struct in_ifaddr_container
*iac
;
518 TAILQ_FOREACH(iac
, &in_ifaddrheads
[mycpuid
], ia_link
) {
521 if (ia
->ia_ifa
.ifa_addr
== sa
&&
522 (ia
->ia_flags
& IFA_ROUTE
)) {
524 * in_ifscrub kills the interface route.
526 in_ifscrub(ia
->ia_ifp
, ia
);
528 * in_ifadown gets rid of all the rest of
529 * the routes. This is not quite the right
530 * thing to do, but at least if we are running
531 * a routing process they will come back.
533 in_ifadown(&ia
->ia_ifa
, 0);
541 TAILQ_FOREACH(iac
, &in_ifaddrheads
[mycpuid
], ia_link
) {
542 if (iac
->ia
->ia_ifa
.ifa_addr
== sa
) {
547 if (ia
== NULL
|| (ia
->ia_flags
& IFA_ROUTE
))
550 ifp
= ia
->ia_ifa
.ifa_ifp
;
554 * Don't add prefix routes for CARP interfaces.
555 * Prefix routes creation is handled by CARP
556 * interfaces themselves.
558 if (ifp
->if_type
== IFT_CARP
)
562 if ((ifp
->if_flags
& IFF_LOOPBACK
) ||
563 (ifp
->if_flags
& IFF_POINTOPOINT
))
566 err
= rtinit(&ia
->ia_ifa
, RTM_ADD
, flags
);
568 ia
->ia_flags
|= IFA_ROUTE
;
572 lwkt_replymsg(&msg
->lmsg
, 0);
575 u_long rip_sendspace
= RIPSNDQ
;
576 u_long rip_recvspace
= RIPRCVQ
;
578 SYSCTL_INT(_net_inet_raw
, OID_AUTO
, maxdgram
, CTLFLAG_RW
,
579 &rip_sendspace
, 0, "Maximum outgoing raw IP datagram size");
580 SYSCTL_INT(_net_inet_raw
, OID_AUTO
, recvspace
, CTLFLAG_RW
,
581 &rip_recvspace
, 0, "Maximum incoming raw IP datagram size");
584 rip_attach(netmsg_t msg
)
586 struct socket
*so
= msg
->base
.nm_so
;
587 int proto
= msg
->attach
.nm_proto
;
588 struct pru_attach_info
*ai
= msg
->attach
.nm_ai
;
597 error
= priv_check_cred(ai
->p_ucred
, PRIV_NETINET_RAW
, NULL_CRED_OKAY
);
601 error
= soreserve(so
, rip_sendspace
, rip_recvspace
, ai
->sb_rlimit
);
605 error
= in_pcballoc(so
, &ripcbinfo
);
607 inp
= (struct inpcb
*)so
->so_pcb
;
608 inp
->inp_ip_p
= proto
;
609 inp
->inp_ip_ttl
= ip_defttl
;
612 lwkt_replymsg(&msg
->lmsg
, error
);
616 rip_detach(netmsg_t msg
)
618 struct socket
*so
= msg
->base
.nm_so
;
626 if (so
== ip_mrouter
&& ip_mrouter_done
)
628 if (ip_rsvp_force_done
)
629 ip_rsvp_force_done(so
);
633 lwkt_replymsg(&msg
->lmsg
, 0);
637 rip_abort(netmsg_t msg
)
640 * Raw socket does not support listen(2),
641 * so this should never be called.
643 panic("rip_abort is called");
647 rip_disconnect(netmsg_t msg
)
649 struct socket
*so
= msg
->base
.nm_so
;
654 if (so
->so_state
& SS_ISCONNECTED
) {
655 soisdisconnected(so
);
660 lwkt_replymsg(&msg
->lmsg
, error
);
664 rip_bind(netmsg_t msg
)
666 struct socket
*so
= msg
->base
.nm_so
;
667 struct sockaddr
*nam
= msg
->bind
.nm_nam
;
668 struct inpcb
*inp
= so
->so_pcb
;
669 struct sockaddr_in
*addr
= (struct sockaddr_in
*)nam
;
674 if (nam
->sa_len
== sizeof(*addr
)) {
675 if (ifnet_array_isempty() ||
676 ((addr
->sin_family
!= AF_INET
) &&
677 (addr
->sin_family
!= AF_IMPLINK
)) ||
678 (addr
->sin_addr
.s_addr
!= INADDR_ANY
&&
679 ifa_ifwithaddr((struct sockaddr
*)addr
) == 0)) {
680 error
= EADDRNOTAVAIL
;
682 inp
->inp_laddr
= addr
->sin_addr
;
688 lwkt_replymsg(&msg
->lmsg
, error
);
692 rip_connect(netmsg_t msg
)
694 struct socket
*so
= msg
->base
.nm_so
;
695 struct sockaddr
*nam
= msg
->connect
.nm_nam
;
696 struct inpcb
*inp
= so
->so_pcb
;
697 struct sockaddr_in
*addr
= (struct sockaddr_in
*)nam
;
702 if (nam
->sa_len
!= sizeof(*addr
)) {
704 } else if (ifnet_array_isempty()) {
705 error
= EADDRNOTAVAIL
;
707 if ((addr
->sin_family
!= AF_INET
) &&
708 (addr
->sin_family
!= AF_IMPLINK
)) {
709 error
= EAFNOSUPPORT
;
711 inp
->inp_faddr
= addr
->sin_addr
;
716 lwkt_replymsg(&msg
->lmsg
, error
);
720 rip_shutdown(netmsg_t msg
)
724 socantsendmore(msg
->base
.nm_so
);
725 lwkt_replymsg(&msg
->lmsg
, 0);
729 rip_send(netmsg_t msg
)
731 struct socket
*so
= msg
->base
.nm_so
;
732 struct mbuf
*m
= msg
->send
.nm_m
;
733 /*struct mbuf *control = msg->send.nm_control;*/
734 struct sockaddr
*nam
= msg
->send
.nm_addr
;
735 /*int flags = msg->send.nm_flags;*/
736 struct inpcb
*inp
= so
->so_pcb
;
742 if (so
->so_state
& SS_ISCONNECTED
) {
747 dst
= inp
->inp_faddr
.s_addr
;
748 error
= rip_output(m
, so
, dst
);
755 dst
= ((struct sockaddr_in
*)nam
)->sin_addr
.s_addr
;
756 error
= rip_output(m
, so
, dst
);
759 lwkt_replymsg(&msg
->lmsg
, error
);
762 SYSCTL_PROC(_net_inet_raw
, OID_AUTO
/*XXX*/, pcblist
, CTLFLAG_RD
, &ripcbinfo
, 1,
763 in_pcblist_range
, "S,xinpcb", "List of active raw IP sockets");
765 struct pr_usrreqs rip_usrreqs
= {
766 .pru_abort
= rip_abort
,
767 .pru_accept
= pr_generic_notsupp
,
768 .pru_attach
= rip_attach
,
769 .pru_bind
= rip_bind
,
770 .pru_connect
= rip_connect
,
771 .pru_connect2
= pr_generic_notsupp
,
772 .pru_control
= in_control_dispatch
,
773 .pru_detach
= rip_detach
,
774 .pru_disconnect
= rip_disconnect
,
775 .pru_listen
= pr_generic_notsupp
,
776 .pru_peeraddr
= in_setpeeraddr_dispatch
,
777 .pru_rcvd
= pr_generic_notsupp
,
778 .pru_rcvoob
= pr_generic_notsupp
,
779 .pru_send
= rip_send
,
780 .pru_sense
= pru_sense_null
,
781 .pru_shutdown
= rip_shutdown
,
782 .pru_sockaddr
= in_setsockaddr_dispatch
,
783 .pru_sosend
= sosend
,
784 .pru_soreceive
= soreceive