1 /* $FreeBSD: src/sys/netinet6/in6_pcb.c,v 1.10.2.9 2003/01/24 05:11:35 sam Exp $ */
2 /* $KAME: in6_pcb.c,v 1.31 2001/05/21 05:45:10 jinmei Exp $ */
5 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
11 * 1. Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
13 * 2. Redistributions in binary form must reproduce the above copyright
14 * notice, this list of conditions and the following disclaimer in the
15 * documentation and/or other materials provided with the distribution.
16 * 3. Neither the name of the project nor the names of its contributors
17 * may be used to endorse or promote products derived from this software
18 * without specific prior written permission.
20 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
21 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
24 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
35 * Copyright (c) 1982, 1986, 1991, 1993
36 * The Regents of the University of California. All rights reserved.
38 * Redistribution and use in source and binary forms, with or without
39 * modification, are permitted provided that the following conditions
41 * 1. Redistributions of source code must retain the above copyright
42 * notice, this list of conditions and the following disclaimer.
43 * 2. Redistributions in binary form must reproduce the above copyright
44 * notice, this list of conditions and the following disclaimer in the
45 * documentation and/or other materials provided with the distribution.
46 * 3. All advertising materials mentioning features or use of this software
47 * must display the following acknowledgement:
48 * This product includes software developed by the University of
49 * California, Berkeley and its contributors.
50 * 4. Neither the name of the University nor the names of its contributors
51 * may be used to endorse or promote products derived from this software
52 * without specific prior written permission.
54 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
55 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
56 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
57 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
58 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
59 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
60 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
61 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
62 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
63 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
66 * @(#)in_pcb.c 8.2 (Berkeley) 1/4/94
70 #include "opt_inet6.h"
71 #include "opt_ipsec.h"
73 #include <sys/param.h>
74 #include <sys/systm.h>
75 #include <sys/malloc.h>
77 #include <sys/domain.h>
78 #include <sys/protosw.h>
79 #include <sys/socket.h>
80 #include <sys/socketvar.h>
81 #include <sys/sockio.h>
82 #include <sys/errno.h>
88 #include <sys/thread2.h>
89 #include <sys/msgport2.h>
91 #include <vm/vm_zone.h>
94 #include <net/if_types.h>
95 #include <net/route.h>
97 #include <netinet/in.h>
98 #include <netinet/in_var.h>
99 #include <netinet/in_systm.h>
100 #include <netinet/ip6.h>
101 #include <netinet/ip_var.h>
102 #include <netinet6/ip6_var.h>
103 #include <netinet6/nd6.h>
104 #include <netinet/in_pcb.h>
105 #include <netinet6/in6_pcb.h>
108 #include <netinet6/ipsec.h>
110 #include <netinet6/ipsec6.h>
112 #include <netinet6/ah.h>
114 #include <netinet6/ah6.h>
116 #include <netproto/key/key.h>
120 #include <netproto/ipsec/ipsec.h>
121 #include <netproto/ipsec/ipsec6.h>
122 #include <netproto/ipsec/key.h>
124 #endif /* FAST_IPSEC */
126 struct in6_addr zeroin6_addr
;
129 in6_pcbbind(struct inpcb
*inp
, struct sockaddr
*nam
, struct thread
*td
)
131 struct socket
*so
= inp
->inp_socket
;
132 struct sockaddr_in6
*sin6
= NULL
;
133 struct sockaddr_in6 jsin6
;
134 struct inpcbinfo
*pcbinfo
= inp
->inp_pcbinfo
;
135 struct proc
*p
= td
->td_proc
;
136 struct ucred
*cred
= NULL
;
138 int wild
= 0, reuseport
= (so
->so_options
& SO_REUSEPORT
);
140 if (!in6_ifaddr
) /* XXX broken! */
141 return (EADDRNOTAVAIL
);
142 if (inp
->inp_lport
|| !IN6_IS_ADDR_UNSPECIFIED(&inp
->in6p_laddr
))
144 if ((so
->so_options
& (SO_REUSEADDR
|SO_REUSEPORT
)) == 0)
149 sin6
= (struct sockaddr_in6
*)nam
;
150 if (nam
->sa_len
!= sizeof(*sin6
))
155 if (nam
->sa_family
!= AF_INET6
)
156 return (EAFNOSUPPORT
);
158 if (!prison_replace_wildcards(td
, nam
))
161 /* KAME hack: embed scopeid */
162 if (in6_embedscope(&sin6
->sin6_addr
, sin6
, inp
, NULL
) != 0)
164 /* this must be cleared for ifa_ifwithaddr() */
165 sin6
->sin6_scope_id
= 0;
167 lport
= sin6
->sin6_port
;
168 if (IN6_IS_ADDR_MULTICAST(&sin6
->sin6_addr
)) {
170 * Treat SO_REUSEADDR as SO_REUSEPORT for multicast;
171 * allow compepte duplication of binding if
172 * SO_REUSEPORT is set, or if SO_REUSEADDR is set
173 * and a multicast address is bound on both
174 * new and duplicated sockets.
176 if (so
->so_options
& SO_REUSEADDR
)
177 reuseport
= SO_REUSEADDR
|SO_REUSEPORT
;
178 } else if (!IN6_IS_ADDR_UNSPECIFIED(&sin6
->sin6_addr
)) {
179 struct ifaddr
*ia
= NULL
;
181 sin6
->sin6_port
= 0; /* yech... */
182 if (!prison_replace_wildcards(td
, (struct sockaddr
*)sin6
)) {
183 sin6
->sin6_addr
= kin6addr_any
;
186 if ((ia
= ifa_ifwithaddr((struct sockaddr
*)sin6
)) == 0)
187 return (EADDRNOTAVAIL
);
190 * XXX: bind to an anycast address might accidentally
191 * cause sending a packet with anycast source address.
192 * We should allow to bind to a deprecated address, since
193 * the application dare to use it.
196 ((struct in6_ifaddr
*)ia
)->ia6_flags
&
197 (IN6_IFF_ANYCAST
|IN6_IFF_NOTREADY
|IN6_IFF_DETACHED
)) {
198 return (EADDRNOTAVAIL
);
205 if (ntohs(lport
) < IPV6PORT_RESERVED
&& cred
&&
206 priv_check_cred(cred
, PRIV_NETINET_RESERVEDPORT
, 0))
208 if (so
->so_cred
->cr_uid
!= 0 &&
209 !IN6_IS_ADDR_MULTICAST(&sin6
->sin6_addr
)) {
210 t
= in6_pcblookup_local(pcbinfo
,
211 &sin6
->sin6_addr
, lport
,
212 INPLOOKUP_WILDCARD
, cred
);
214 (!IN6_IS_ADDR_UNSPECIFIED(&sin6
->sin6_addr
) ||
215 !IN6_IS_ADDR_UNSPECIFIED(&t
->in6p_laddr
) ||
216 (t
->inp_socket
->so_options
&
217 SO_REUSEPORT
) == 0) &&
218 (so
->so_cred
->cr_uid
!=
219 t
->inp_socket
->so_cred
->cr_uid
))
221 if ((inp
->inp_flags
& IN6P_IPV6_V6ONLY
) == 0 &&
222 IN6_IS_ADDR_UNSPECIFIED(&sin6
->sin6_addr
)) {
223 struct sockaddr_in sin
;
225 in6_sin6_2_sin(&sin
, sin6
);
226 t
= in_pcblookup_local(pcbinfo
,
228 INPLOOKUP_WILDCARD
, cred
);
230 (so
->so_cred
->cr_uid
!=
231 t
->inp_socket
->so_cred
->cr_uid
) &&
232 (ntohl(t
->inp_laddr
.s_addr
) !=
235 INP_SOCKAF(t
->inp_socket
)))
239 if (cred
&& cred
->cr_prison
&&
240 !prison_replace_wildcards(td
, nam
))
241 return(EADDRNOTAVAIL
);
242 t
= in6_pcblookup_local(pcbinfo
, &sin6
->sin6_addr
,
244 if (t
&& (reuseport
& t
->inp_socket
->so_options
) == 0)
246 if ((inp
->inp_flags
& IN6P_IPV6_V6ONLY
) == 0 &&
247 IN6_IS_ADDR_UNSPECIFIED(&sin6
->sin6_addr
)) {
248 struct sockaddr_in sin
;
250 in6_sin6_2_sin(&sin
, sin6
);
251 t
= in_pcblookup_local(pcbinfo
,
255 (reuseport
& t
->inp_socket
->so_options
)
257 (ntohl(t
->inp_laddr
.s_addr
)
260 INP_SOCKAF(t
->inp_socket
)))
264 inp
->in6p_laddr
= sin6
->sin6_addr
;
269 jsin6
.sin6_addr
= inp
->in6p_laddr
;
270 jsin6
.sin6_family
= AF_INET6
;
271 if (!prison_replace_wildcards(td
, (struct sockaddr
*)&jsin6
)) {
272 inp
->in6p_laddr
= kin6addr_any
;
277 if ((e
= in6_pcbsetport(&inp
->in6p_laddr
, inp
, td
)) != 0)
281 inp
->inp_lport
= lport
;
282 if (in_pcbinsporthash(inp
) != 0) {
283 inp
->in6p_laddr
= kin6addr_any
;
292 * Transform old in6_pcbconnect() into an inner subroutine for new
293 * in6_pcbconnect(): Do some validity-checking on the remote
294 * address (in mbuf 'nam') and then determine local host address
295 * (i.e., which interface) to use to access that remote host.
297 * This preserves definition of in6_pcbconnect(), while supporting a
298 * slightly different version for T/TCP. (This is more than
299 * a bit of a kludge, but cleaning up the internal interfaces would
300 * have forced minor changes in every protocol).
304 in6_pcbladdr(struct inpcb
*inp
, struct sockaddr
*nam
,
305 struct in6_addr
**plocal_addr6
, struct thread
*td
)
307 struct sockaddr_in6
*sin6
= (struct sockaddr_in6
*)nam
;
308 struct ifnet
*ifp
= NULL
;
311 if (nam
->sa_len
!= sizeof (*sin6
))
313 if (sin6
->sin6_family
!= AF_INET6
)
314 return (EAFNOSUPPORT
);
315 if (sin6
->sin6_port
== 0)
316 return (EADDRNOTAVAIL
);
318 /* KAME hack: embed scopeid */
319 if (in6_embedscope(&sin6
->sin6_addr
, sin6
, inp
, &ifp
) != 0)
324 * If the destination address is UNSPECIFIED addr,
325 * use the loopback addr, e.g ::1.
327 if (IN6_IS_ADDR_UNSPECIFIED(&sin6
->sin6_addr
))
328 sin6
->sin6_addr
= kin6addr_loopback
;
332 * XXX: in6_selectsrc might replace the bound local address
333 * with the address specified by setsockopt(IPV6_PKTINFO).
334 * Is it the intended behavior?
336 *plocal_addr6
= in6_selectsrc(sin6
, inp
->in6p_outputopts
,
339 &inp
->in6p_laddr
, &error
, td
);
340 if (*plocal_addr6
== 0) {
342 error
= EADDRNOTAVAIL
;
346 * Don't do pcblookup call here; return interface in
348 * and exit to caller, that will do the lookup.
352 if (inp
->in6p_route
.ro_rt
)
353 ifp
= inp
->in6p_route
.ro_rt
->rt_ifp
;
360 * Connect from a socket to a specified address.
361 * Both address and port must be specified in argument sin.
362 * If don't have a local address for this socket yet,
366 in6_pcbconnect(struct inpcb
*inp
, struct sockaddr
*nam
, struct thread
*td
)
368 struct in6_addr
*addr6
;
369 struct sockaddr_in6
*sin6
= (struct sockaddr_in6
*)nam
;
373 * Call inner routine, to assign local interface address.
374 * in6_pcbladdr() may automatically fill in sin6_scope_id.
376 if ((error
= in6_pcbladdr(inp
, nam
, &addr6
, td
)) != 0)
379 if (in6_pcblookup_hash(inp
->inp_cpcbinfo
, &sin6
->sin6_addr
,
381 IN6_IS_ADDR_UNSPECIFIED(&inp
->in6p_laddr
)
382 ? addr6
: &inp
->in6p_laddr
,
383 inp
->inp_lport
, 0, NULL
) != NULL
) {
386 if (IN6_IS_ADDR_UNSPECIFIED(&inp
->in6p_laddr
)) {
387 if (inp
->inp_lport
== 0) {
388 error
= in6_pcbbind(inp
, NULL
, td
);
392 inp
->in6p_laddr
= *addr6
;
394 inp
->in6p_faddr
= sin6
->sin6_addr
;
395 inp
->inp_fport
= sin6
->sin6_port
;
396 /* update flowinfo - draft-itojun-ipv6-flowlabel-api-00 */
397 inp
->in6p_flowinfo
&= ~IPV6_FLOWLABEL_MASK
;
398 if (inp
->in6p_flags
& IN6P_AUTOFLOWLABEL
)
399 inp
->in6p_flowinfo
|=
400 (htonl(ip6_flow_seq
++) & IPV6_FLOWLABEL_MASK
);
402 in_pcbinsconnhash(inp
);
408 * Return an IPv6 address, which is the most appropriate for given
409 * destination and user specified options.
410 * If necessary, this function lookups the routing table and return
411 * an entry to the caller for later use.
414 in6_selectsrc(struct sockaddr_in6
*dstsock
, struct ip6_pktopts
*opts
,
415 struct ip6_moptions
*mopts
, struct route_in6
*ro
,
416 struct in6_addr
*laddr
, int *errorp
, struct thread
*td
)
418 struct sockaddr_in6 jsin6
;
419 struct ucred
*cred
= NULL
;
420 struct in6_addr
*dst
;
421 struct in6_ifaddr
*ia6
= 0;
422 struct in6_pktinfo
*pi
= NULL
;
425 if (td
&& td
->td_proc
&& td
->td_proc
->p_ucred
)
426 cred
= td
->td_proc
->p_ucred
;
427 if (cred
&& cred
->cr_prison
)
429 jsin6
.sin6_family
= AF_INET6
;
430 dst
= &dstsock
->sin6_addr
;
434 * If the source address is explicitly specified by the caller,
437 if (opts
&& (pi
= opts
->ip6po_pktinfo
) &&
438 !IN6_IS_ADDR_UNSPECIFIED(&pi
->ipi6_addr
)) {
439 jsin6
.sin6_addr
= pi
->ipi6_addr
;
440 if (jailed
&& !jailed_ip(cred
->cr_prison
,
441 (struct sockaddr
*)&jsin6
)) {
444 return (&pi
->ipi6_addr
);
449 * If the source address is not specified but the socket(if any)
450 * is already bound, use the bound address.
452 if (laddr
&& !IN6_IS_ADDR_UNSPECIFIED(laddr
)) {
453 jsin6
.sin6_addr
= *laddr
;
454 if (jailed
&& !jailed_ip(cred
->cr_prison
,
455 (struct sockaddr
*)&jsin6
)) {
463 * If the caller doesn't specify the source address but
464 * the outgoing interface, use an address associated with
467 if (pi
&& pi
->ipi6_ifindex
) {
468 /* XXX boundary check is assumed to be already done. */
469 ia6
= in6_ifawithscope(ifindex2ifnet
[pi
->ipi6_ifindex
],
473 jsin6
.sin6_addr
= (&ia6
->ia_addr
)->sin6_addr
;
474 if (!jailed_ip(cred
->cr_prison
,
475 (struct sockaddr
*)&jsin6
))
480 *errorp
= EADDRNOTAVAIL
;
483 return (&satosin6(&ia6
->ia_addr
)->sin6_addr
);
487 * If the destination address is a link-local unicast address or
488 * a multicast address, and if the outgoing interface is specified
489 * by the sin6_scope_id filed, use an address associated with the
491 * XXX: We're now trying to define more specific semantics of
492 * sin6_scope_id field, so this part will be rewritten in
495 if ((IN6_IS_ADDR_LINKLOCAL(dst
) || IN6_IS_ADDR_MULTICAST(dst
)) &&
496 dstsock
->sin6_scope_id
) {
498 * I'm not sure if boundary check for scope_id is done
501 if (dstsock
->sin6_scope_id
< 0 ||
502 if_index
< dstsock
->sin6_scope_id
) {
503 *errorp
= ENXIO
; /* XXX: better error? */
506 ia6
= in6_ifawithscope(ifindex2ifnet
[dstsock
->sin6_scope_id
],
510 jsin6
.sin6_addr
= (&ia6
->ia_addr
)->sin6_addr
;
511 if (!jailed_ip(cred
->cr_prison
,
512 (struct sockaddr
*)&jsin6
))
517 *errorp
= EADDRNOTAVAIL
;
520 return (&satosin6(&ia6
->ia_addr
)->sin6_addr
);
524 * If the destination address is a multicast address and
525 * the outgoing interface for the address is specified
526 * by the caller, use an address associated with the interface.
527 * There is a sanity check here; if the destination has node-local
528 * scope, the outgoing interfacde should be a loopback address.
529 * Even if the outgoing interface is not specified, we also
530 * choose a loopback interface as the outgoing interface.
532 if (!jailed
&& IN6_IS_ADDR_MULTICAST(dst
)) {
533 struct ifnet
*ifp
= mopts
? mopts
->im6o_multicast_ifp
: NULL
;
535 if (ifp
== NULL
&& IN6_IS_ADDR_MC_NODELOCAL(dst
)) {
540 ia6
= in6_ifawithscope(ifp
, dst
);
542 *errorp
= EADDRNOTAVAIL
;
545 return (&ia6
->ia_addr
.sin6_addr
);
550 * If the next hop address for the packet is specified
551 * by caller, use an address associated with the route
555 struct sockaddr_in6
*sin6_next
;
558 if (opts
&& opts
->ip6po_nexthop
) {
559 sin6_next
= satosin6(opts
->ip6po_nexthop
);
560 rt
= nd6_lookup(&sin6_next
->sin6_addr
, 1, NULL
);
562 ia6
= in6_ifawithscope(rt
->rt_ifp
, dst
);
564 ia6
= ifatoia6(rt
->rt_ifa
);
567 jsin6
.sin6_addr
= (&ia6
->ia_addr
)->sin6_addr
;
568 if (!jailed_ip(cred
->cr_prison
,
569 (struct sockaddr
*)&jsin6
))
574 *errorp
= EADDRNOTAVAIL
;
577 return (&satosin6(&ia6
->ia_addr
)->sin6_addr
);
582 * If route is known or can be allocated now,
583 * our src addr is taken from the i/f, else punt.
587 !IN6_ARE_ADDR_EQUAL(&satosin6(&ro
->ro_dst
)->sin6_addr
, dst
)) {
591 if (ro
->ro_rt
== NULL
|| ro
->ro_rt
->rt_ifp
== NULL
) {
592 struct sockaddr_in6
*dst6
;
594 /* No route yet, so try to acquire one */
595 bzero(&ro
->ro_dst
, sizeof(struct sockaddr_in6
));
597 dst6
->sin6_family
= AF_INET6
;
598 dst6
->sin6_len
= sizeof(struct sockaddr_in6
);
599 dst6
->sin6_addr
= *dst
;
600 if (!jailed
&& IN6_IS_ADDR_MULTICAST(dst
)) {
602 rtpurelookup((struct sockaddr
*)&ro
->ro_dst
);
604 rtalloc((struct route
*)ro
);
609 * in_pcbconnect() checks out IFF_LOOPBACK to skip using
610 * the address. But we don't know why it does so.
611 * It is necessary to ensure the scope even for lo0
612 * so doesn't check out IFF_LOOPBACK.
616 ia6
= in6_ifawithscope(ro
->ro_rt
->rt_ifa
->ifa_ifp
, dst
);
618 jsin6
.sin6_addr
= (&ia6
->ia_addr
)->sin6_addr
;
619 if (!jailed_ip(cred
->cr_prison
,
620 (struct sockaddr
*)&jsin6
))
624 if (ia6
== 0) /* xxx scope error ?*/
625 ia6
= ifatoia6(ro
->ro_rt
->rt_ifa
);
628 jsin6
.sin6_addr
= (&ia6
->ia_addr
)->sin6_addr
;
629 if (!jailed_ip(cred
->cr_prison
,
630 (struct sockaddr
*)&jsin6
))
635 *errorp
= EHOSTUNREACH
; /* no route */
638 return (&satosin6(&ia6
->ia_addr
)->sin6_addr
);
641 *errorp
= EADDRNOTAVAIL
;
646 * Default hop limit selection. The precedence is as follows:
647 * 1. Hoplimit valued specified via ioctl.
648 * 2. (If the outgoing interface is detected) the current
649 * hop limit of the interface specified by router advertisement.
650 * 3. The system default hoplimit.
653 in6_selecthlim(struct in6pcb
*in6p
, struct ifnet
*ifp
)
655 if (in6p
&& in6p
->in6p_hops
>= 0)
656 return (in6p
->in6p_hops
);
658 return (ND_IFINFO(ifp
)->chlim
);
660 return (ip6_defhlim
);
665 in6_pcbdisconnect(struct inpcb
*inp
)
667 bzero((caddr_t
)&inp
->in6p_faddr
, sizeof(inp
->in6p_faddr
));
669 /* clear flowinfo - draft-itojun-ipv6-flowlabel-api-00 */
670 inp
->in6p_flowinfo
&= ~IPV6_FLOWLABEL_MASK
;
671 in_pcbremconnhash(inp
);
672 if (inp
->inp_socket
->so_state
& SS_NOFDREF
)
677 in6_pcbdetach(struct inpcb
*inp
)
679 struct socket
*so
= inp
->inp_socket
;
680 struct inpcbinfo
*ipi
= inp
->inp_pcbinfo
;
683 if (inp
->in6p_sp
!= NULL
)
684 ipsec6_delete_pcbpolicy(inp
);
686 inp
->inp_gencnt
= ++ipi
->ipi_gencnt
;
689 KKASSERT((so
->so_state
& SS_ASSERTINPROG
) == 0);
690 sofree(so
); /* remove pcb ref */
692 if (inp
->in6p_options
)
693 m_freem(inp
->in6p_options
);
694 ip6_freepcbopts(inp
->in6p_outputopts
);
695 ip6_freemoptions(inp
->in6p_moptions
);
696 if (inp
->in6p_route
.ro_rt
)
697 rtfree(inp
->in6p_route
.ro_rt
);
698 /* Check and free IPv4 related resources in case of mapped addr */
699 if (inp
->inp_options
)
700 m_free(inp
->inp_options
);
701 ip_freemoptions(inp
->inp_moptions
);
708 * The calling convention of in6_setsockaddr() and in6_setpeeraddr() was
709 * modified to match the pru_sockaddr() and pru_peeraddr() entry points
710 * in struct pr_usrreqs, so that protocols can just reference then directly
711 * without the need for a wrapper function. The socket must have a valid
712 * (i.e., non-nil) PCB, but it should be impossible to get an invalid one
713 * except through a kernel programming error, so it is acceptable to panic
714 * (or in this case trap) if the PCB is invalid. (Actually, we don't trap
715 * because there actually /is/ a programming error somewhere... XXX)
718 in6_setsockaddr_dispatch(netmsg_t msg
)
722 error
= in6_setsockaddr(msg
->sockaddr
.base
.nm_so
, msg
->sockaddr
.nm_nam
);
723 lwkt_replymsg(&msg
->sockaddr
.base
.lmsg
, error
);
727 in6_setsockaddr(struct socket
*so
, struct sockaddr
**nam
)
730 struct sockaddr_in6
*sin6
;
733 * Do the malloc first in case it blocks.
735 sin6
= kmalloc(sizeof *sin6
, M_SONAME
, M_WAITOK
| M_ZERO
);
736 sin6
->sin6_family
= AF_INET6
;
737 sin6
->sin6_len
= sizeof(*sin6
);
743 kfree(sin6
, M_SONAME
);
746 sin6
->sin6_port
= inp
->inp_lport
;
747 sin6
->sin6_addr
= inp
->in6p_laddr
;
749 if (IN6_IS_SCOPE_LINKLOCAL(&sin6
->sin6_addr
))
750 sin6
->sin6_scope_id
= ntohs(sin6
->sin6_addr
.s6_addr16
[1]);
752 sin6
->sin6_scope_id
= 0; /*XXX*/
753 if (IN6_IS_SCOPE_LINKLOCAL(&sin6
->sin6_addr
))
754 sin6
->sin6_addr
.s6_addr16
[1] = 0;
756 *nam
= (struct sockaddr
*)sin6
;
761 in6_setpeeraddr_dispatch(netmsg_t msg
)
765 error
= in6_setpeeraddr(msg
->peeraddr
.base
.nm_so
, msg
->peeraddr
.nm_nam
);
766 lwkt_replymsg(&msg
->peeraddr
.base
.lmsg
, error
);
770 in6_setpeeraddr(struct socket
*so
, struct sockaddr
**nam
)
773 struct sockaddr_in6
*sin6
;
776 * Do the malloc first in case it blocks.
778 sin6
= kmalloc(sizeof(*sin6
), M_SONAME
, M_WAITOK
| M_ZERO
);
779 sin6
->sin6_family
= AF_INET6
;
780 sin6
->sin6_len
= sizeof(struct sockaddr_in6
);
786 kfree(sin6
, M_SONAME
);
789 sin6
->sin6_port
= inp
->inp_fport
;
790 sin6
->sin6_addr
= inp
->in6p_faddr
;
792 if (IN6_IS_SCOPE_LINKLOCAL(&sin6
->sin6_addr
))
793 sin6
->sin6_scope_id
= ntohs(sin6
->sin6_addr
.s6_addr16
[1]);
795 sin6
->sin6_scope_id
= 0; /*XXX*/
796 if (IN6_IS_SCOPE_LINKLOCAL(&sin6
->sin6_addr
))
797 sin6
->sin6_addr
.s6_addr16
[1] = 0;
799 *nam
= (struct sockaddr
*)sin6
;
804 in6_mapped_sockaddr_dispatch(netmsg_t msg
)
808 error
= in6_mapped_sockaddr(msg
->sockaddr
.base
.nm_so
,
809 msg
->sockaddr
.nm_nam
);
810 lwkt_replymsg(&msg
->sockaddr
.base
.lmsg
, error
);
814 in6_mapped_sockaddr(struct socket
*so
, struct sockaddr
**nam
)
816 struct inpcb
*inp
= so
->so_pcb
;
821 if (inp
->inp_vflag
& INP_IPV4
) {
822 error
= in_setsockaddr(so
, nam
);
824 in6_sin_2_v4mapsin6_in_sock(nam
);
826 /* scope issues will be handled in in6_setsockaddr(). */
827 error
= in6_setsockaddr(so
, nam
);
833 in6_mapped_peeraddr(struct socket
*so
, struct sockaddr
**nam
)
835 struct inpcb
*inp
= so
->so_pcb
;
840 if (inp
->inp_vflag
& INP_IPV4
) {
841 error
= in_setpeeraddr(so
, nam
);
843 in6_sin_2_v4mapsin6_in_sock(nam
);
845 /* scope issues will be handled in in6_setpeeraddr(). */
846 error
= in6_setpeeraddr(so
, nam
);
852 in6_mapped_peeraddr_dispatch(netmsg_t msg
)
856 error
= in6_mapped_peeraddr(msg
->base
.nm_so
, msg
->peeraddr
.nm_nam
);
857 lwkt_replymsg(&msg
->base
.lmsg
, error
);
861 * Pass some notification to all connections of a protocol
862 * associated with address dst. The local address and/or port numbers
863 * may be specified to limit the search. The "usual action" will be
864 * taken, depending on the ctlinput cmd. The caller must filter any
865 * cmds that are uninteresting (e.g., no error in the map).
866 * Call the protocol specific routine (if any) to report
867 * any errors for each matching socket.
869 * Must be called under crit_enter().
872 in6_pcbnotify(struct inpcbhead
*head
, struct sockaddr
*dst
, in_port_t fport
,
873 const struct sockaddr
*src
, in_port_t lport
, int cmd
, int arg
,
874 void (*notify
) (struct inpcb
*, int))
876 struct inpcb
*inp
, *ninp
;
877 struct sockaddr_in6 sa6_src
, *sa6_dst
;
880 if ((unsigned)cmd
>= PRC_NCMDS
|| dst
->sa_family
!= AF_INET6
)
883 sa6_dst
= (struct sockaddr_in6
*)dst
;
884 if (IN6_IS_ADDR_UNSPECIFIED(&sa6_dst
->sin6_addr
))
888 * note that src can be NULL when we get notify by local fragmentation.
890 sa6_src
= (src
== NULL
) ? sa6_any
: *(const struct sockaddr_in6
*)src
;
891 flowinfo
= sa6_src
.sin6_flowinfo
;
894 * Redirects go to all references to the destination,
895 * and use in6_rtchange to invalidate the route cache.
896 * Dead host indications: also use in6_rtchange to invalidate
897 * the cache, and deliver the error to all the sockets.
898 * Otherwise, if we have knowledge of the local port and address,
899 * deliver only to that socket.
901 if (PRC_IS_REDIRECT(cmd
) || cmd
== PRC_HOSTDEAD
) {
904 bzero((caddr_t
)&sa6_src
.sin6_addr
, sizeof(sa6_src
.sin6_addr
));
906 if (cmd
!= PRC_HOSTDEAD
)
907 notify
= in6_rtchange
;
909 if (cmd
!= PRC_MSGSIZE
)
910 arg
= inet6ctlerrmap
[cmd
];
912 for (inp
= LIST_FIRST(head
); inp
!= NULL
; inp
= ninp
) {
913 ninp
= LIST_NEXT(inp
, inp_list
);
915 if (inp
->inp_flags
& INP_PLACEMARKER
)
918 if ((inp
->inp_vflag
& INP_IPV6
) == 0)
921 * If the error designates a new path MTU for a destination
922 * and the application (associated with this socket) wanted to
923 * know the value, notify. Note that we notify for all
924 * disconnected sockets if the corresponding application
925 * wanted. This is because some UDP applications keep sending
926 * sockets disconnected.
927 * XXX: should we avoid to notify the value to TCP sockets?
929 if (cmd
== PRC_MSGSIZE
&& (inp
->inp_flags
& IN6P_MTU
) != 0 &&
930 (IN6_IS_ADDR_UNSPECIFIED(&inp
->in6p_faddr
) ||
931 IN6_ARE_ADDR_EQUAL(&inp
->in6p_faddr
, &sa6_dst
->sin6_addr
))) {
932 ip6_notify_pmtu(inp
, (struct sockaddr_in6
*)dst
, &arg
);
936 * Detect if we should notify the error. If no source and
937 * destination ports are specifed, but non-zero flowinfo and
938 * local address match, notify the error. This is the case
939 * when the error is delivered with an encrypted buffer
940 * by ESP. Otherwise, just compare addresses and ports
943 if (lport
== 0 && fport
== 0 && flowinfo
&&
944 inp
->inp_socket
!= NULL
&&
945 flowinfo
== (inp
->in6p_flowinfo
& IPV6_FLOWLABEL_MASK
) &&
946 IN6_ARE_ADDR_EQUAL(&inp
->in6p_laddr
, &sa6_src
.sin6_addr
))
948 else if (!IN6_ARE_ADDR_EQUAL(&inp
->in6p_faddr
,
949 &sa6_dst
->sin6_addr
) ||
950 inp
->inp_socket
== 0 ||
951 (lport
&& inp
->inp_lport
!= lport
) ||
952 (!IN6_IS_ADDR_UNSPECIFIED(&sa6_src
.sin6_addr
) &&
953 !IN6_ARE_ADDR_EQUAL(&inp
->in6p_laddr
,
954 &sa6_src
.sin6_addr
)) ||
955 (fport
&& inp
->inp_fport
!= fport
))
966 * Lookup a PCB based on the local address and port.
969 in6_pcblookup_local(struct inpcbinfo
*pcbinfo
, struct in6_addr
*laddr
,
970 u_int lport_arg
, int wild_okay
, struct ucred
*cred
)
973 int matchwild
= 3, wildcard
;
974 u_short lport
= lport_arg
;
975 struct inpcbporthead
*porthash
;
976 struct inpcbport
*phd
;
977 struct inpcb
*match
= NULL
;
980 * If the porthashbase is shared across several cpus we need
983 if (pcbinfo
->porttoken
)
984 lwkt_gettoken(pcbinfo
->porttoken
);
987 * Best fit PCB lookup.
989 * First see if this local port is in use by looking on the
992 porthash
= &pcbinfo
->porthashbase
[
993 INP_PCBPORTHASH(lport
, pcbinfo
->porthashmask
)];
994 LIST_FOREACH(phd
, porthash
, phd_hash
) {
995 if (phd
->phd_port
== lport
)
1001 * Port is in use by one or more PCBs. Look for best
1004 LIST_FOREACH(inp
, &phd
->phd_pcblist
, inp_portlist
) {
1006 if ((inp
->inp_vflag
& INP_IPV6
) == 0)
1008 if (!IN6_IS_ADDR_UNSPECIFIED(&inp
->in6p_faddr
))
1010 if (!IN6_IS_ADDR_UNSPECIFIED(&inp
->in6p_laddr
)) {
1011 if (IN6_IS_ADDR_UNSPECIFIED(laddr
))
1013 else if (!IN6_ARE_ADDR_EQUAL(
1014 &inp
->in6p_laddr
, laddr
))
1017 if (!IN6_IS_ADDR_UNSPECIFIED(laddr
))
1020 if (wildcard
&& !wild_okay
)
1022 if (wildcard
< matchwild
&&
1025 inp
->inp_socket
->so_cred
->cr_prison
)) {
1027 matchwild
= wildcard
;
1031 matchwild
= wildcard
;
1035 if (pcbinfo
->porttoken
)
1036 lwkt_reltoken(pcbinfo
->porttoken
);
1042 in6_pcbpurgeif0(struct in6pcb
*head
, struct ifnet
*ifp
)
1044 struct in6pcb
*in6p
;
1045 struct ip6_moptions
*im6o
;
1046 struct in6_multi_mship
*imm
, *nimm
;
1048 for (in6p
= head
; in6p
!= NULL
; in6p
= LIST_NEXT(in6p
, inp_list
)) {
1049 if (in6p
->in6p_flags
& INP_PLACEMARKER
)
1051 im6o
= in6p
->in6p_moptions
;
1052 if ((in6p
->inp_vflag
& INP_IPV6
) &&
1055 * Unselect the outgoing interface if it is being
1058 if (im6o
->im6o_multicast_ifp
== ifp
)
1059 im6o
->im6o_multicast_ifp
= NULL
;
1062 * Drop multicast group membership if we joined
1063 * through the interface being detached.
1064 * XXX controversial - is it really legal for kernel
1067 for (imm
= im6o
->im6o_memberships
.lh_first
;
1068 imm
!= NULL
; imm
= nimm
) {
1069 nimm
= imm
->i6mm_chain
.le_next
;
1070 if (imm
->i6mm_maddr
->in6m_ifp
== ifp
) {
1071 LIST_REMOVE(imm
, i6mm_chain
);
1072 in6_delmulti(imm
->i6mm_maddr
);
1073 kfree(imm
, M_IPMADDR
);
1081 * Check for alternatives when higher level complains
1082 * about service problems. For now, invalidate cached
1083 * routing information. If the route was created dynamically
1084 * (by a redirect), time to try a default gateway again.
1087 in6_losing(struct inpcb
*in6p
)
1090 struct rt_addrinfo info
;
1092 if ((rt
= in6p
->in6p_route
.ro_rt
) != NULL
) {
1093 bzero((caddr_t
)&info
, sizeof(info
));
1094 info
.rti_flags
= rt
->rt_flags
;
1095 info
.rti_info
[RTAX_DST
] = rt_key(rt
);
1096 info
.rti_info
[RTAX_GATEWAY
] = rt
->rt_gateway
;
1097 info
.rti_info
[RTAX_NETMASK
] = rt_mask(rt
);
1098 rt_missmsg(RTM_LOSING
, &info
, rt
->rt_flags
, 0);
1099 if (rt
->rt_flags
& RTF_DYNAMIC
)
1100 rtrequest1_global(RTM_DELETE
, &info
, NULL
, NULL
);
1101 in6p
->in6p_route
.ro_rt
= NULL
;
1104 * A new route can be allocated
1105 * the next time output is attempted.
1111 * After a routing change, flush old routing
1112 * and allocate a (hopefully) better one.
1115 in6_rtchange(struct inpcb
*inp
, int error
)
1117 if (inp
->in6p_route
.ro_rt
) {
1118 rtfree(inp
->in6p_route
.ro_rt
);
1119 inp
->in6p_route
.ro_rt
= 0;
1121 * A new route can be allocated the next time
1122 * output is attempted.
1128 * Lookup PCB in hash list.
1131 in6_pcblookup_hash(struct inpcbinfo
*pcbinfo
, struct in6_addr
*faddr
,
1132 u_int fport_arg
, struct in6_addr
*laddr
, u_int lport_arg
,
1133 int wildcard
, struct ifnet
*ifp
)
1135 struct inpcbhead
*head
;
1137 struct inpcb
*jinp
= NULL
;
1138 u_short fport
= fport_arg
, lport
= lport_arg
;
1141 if (faithprefix_p
!= NULL
)
1142 faith
= (*faithprefix_p
)(laddr
);
1147 * First look for an exact match.
1149 head
= &pcbinfo
->hashbase
[INP_PCBCONNHASH(faddr
->s6_addr32
[3] /* XXX */,
1151 laddr
->s6_addr32
[3], /* XXX JH */
1153 pcbinfo
->hashmask
)];
1154 LIST_FOREACH(inp
, head
, inp_hash
) {
1155 if ((inp
->inp_vflag
& INP_IPV6
) == 0)
1157 if (IN6_ARE_ADDR_EQUAL(&inp
->in6p_faddr
, faddr
) &&
1158 IN6_ARE_ADDR_EQUAL(&inp
->in6p_laddr
, laddr
) &&
1159 inp
->inp_fport
== fport
&&
1160 inp
->inp_lport
== lport
) {
1164 if (inp
->inp_socket
== NULL
||
1165 inp
->inp_socket
->so_cred
->cr_prison
== NULL
) {
1176 struct inpcontainerhead
*chead
;
1177 struct inpcontainer
*ic
;
1178 struct inpcb
*local_wild
= NULL
;
1179 struct inpcb
*jinp_wild
= NULL
;
1180 struct sockaddr_in6 jsin6
;
1184 * Order of socket selection:
1185 * 1. non-jailed, non-wild.
1186 * 2. non-jailed, wild.
1187 * 3. jailed, non-wild.
1190 jsin6
.sin6_family
= AF_INET6
;
1191 chead
= &pcbinfo
->wildcardhashbase
[INP_PCBWILDCARDHASH(lport
,
1192 pcbinfo
->wildcardhashmask
)];
1193 LIST_FOREACH(ic
, chead
, ic_list
) {
1196 if (!(inp
->inp_vflag
& INP_IPV6
))
1198 if (inp
->inp_socket
!= NULL
)
1199 cred
= inp
->inp_socket
->so_cred
;
1203 if (cred
!= NULL
&& jailed(cred
)) {
1207 jsin6
.sin6_addr
= *laddr
;
1208 if (!jailed_ip(cred
->cr_prison
,
1209 (struct sockaddr
*)&jsin6
))
1213 if (IN6_IS_ADDR_UNSPECIFIED(&inp
->in6p_faddr
) &&
1214 inp
->inp_lport
== lport
) {
1215 if (faith
&& (inp
->inp_flags
& INP_FAITH
) == 0)
1217 if (IN6_ARE_ADDR_EQUAL(&inp
->in6p_laddr
,
1219 if (cred
!= NULL
&& jailed(cred
))
1223 } else if (IN6_IS_ADDR_UNSPECIFIED(&inp
->in6p_laddr
)) {
1224 if (cred
!= NULL
&& jailed(cred
))
1231 if (local_wild
!= NULL
)
1232 return (local_wild
);
1245 init_sin6(struct sockaddr_in6
*sin6
, struct mbuf
*m
)
1249 ip
= mtod(m
, struct ip6_hdr
*);
1250 bzero(sin6
, sizeof(*sin6
));
1251 sin6
->sin6_len
= sizeof(*sin6
);
1252 sin6
->sin6_family
= AF_INET6
;
1253 sin6
->sin6_addr
= ip
->ip6_src
;
1254 if (IN6_IS_SCOPE_LINKLOCAL(&sin6
->sin6_addr
))
1255 sin6
->sin6_addr
.s6_addr16
[1] = 0;
1256 sin6
->sin6_scope_id
=
1257 (m
->m_pkthdr
.rcvif
&& IN6_IS_SCOPE_LINKLOCAL(&sin6
->sin6_addr
))
1258 ? m
->m_pkthdr
.rcvif
->if_index
: 0;
1264 in6_savefaddr(struct socket
*so
, const struct sockaddr
*faddr
)
1266 struct sockaddr_in6
*sin6
;
1268 KASSERT(faddr
->sa_family
== AF_INET6
,
1269 ("not AF_INET6 faddr %d\n", faddr
->sa_family
));
1271 sin6
= kmalloc(sizeof(*sin6
), M_SONAME
, M_WAITOK
| M_ZERO
);
1272 sin6
->sin6_family
= AF_INET6
;
1273 sin6
->sin6_len
= sizeof(*sin6
);
1275 sin6
->sin6_port
= ((const struct sockaddr_in6
*)faddr
)->sin6_port
;
1276 sin6
->sin6_addr
= ((const struct sockaddr_in6
*)faddr
)->sin6_addr
;
1278 if (IN6_IS_SCOPE_LINKLOCAL(&sin6
->sin6_addr
))
1279 sin6
->sin6_scope_id
= ntohs(sin6
->sin6_addr
.s6_addr16
[1]);
1281 sin6
->sin6_scope_id
= 0; /*XXX*/
1282 if (IN6_IS_SCOPE_LINKLOCAL(&sin6
->sin6_addr
))
1283 sin6
->sin6_addr
.s6_addr16
[1] = 0;
1285 so
->so_faddr
= (struct sockaddr
*)sin6
;
1289 in6_mapped_savefaddr(struct socket
*so
, const struct sockaddr
*faddr
)
1291 if (faddr
->sa_family
== AF_INET
) {
1292 in_savefaddr(so
, faddr
);
1293 in6_sin_2_v4mapsin6_in_sock(&so
->so_faddr
);
1295 in6_savefaddr(so
, faddr
);