1 /*-
2 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
3 * All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
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 project 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.
16 *
17 * THIS SOFTWARE IS PROVIDED BY THE PROJECT 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 PROJECT 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
27 * SUCH DAMAGE.
28 */
29
30 /*-
31 * Copyright (c) 1982, 1986, 1988, 1993
32 * The Regents of the University of California.
33 * All rights reserved.
34 *
35 * Redistribution and use in source and binary forms, with or without
36 * modification, are permitted provided that the following conditions
37 * are met:
38 * 1. Redistributions of source code must retain the above copyright
39 * notice, this list of conditions and the following disclaimer.
40 * 2. Redistributions in binary form must reproduce the above copyright
41 * notice, this list of conditions and the following disclaimer in the
42 * documentation and/or other materials provided with the distribution.
43 * 4. Neither the name of the University nor the names of its contributors
44 * may be used to endorse or promote products derived from this software
45 * without specific prior written permission.
46 *
47 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
48 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
49 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
50 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
51 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
52 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
53 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
54 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
55 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
56 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
57 * SUCH DAMAGE.
58 *
59 * @(#)raw_ip.c 8.2 (Berkeley) 1/4/94
60 */
61
62 #include <sys/cdefs.h>
63 __FBSDID("$FreeBSD$");
64
65 #include "opt_ipsec.h"
66 #include "opt_inet6.h"
67
68 #include <sys/param.h>
69 #include <sys/errno.h>
70 #include <sys/jail.h>
71 #include <sys/kernel.h>
72 #include <sys/lock.h>
73 #include <sys/malloc.h>
74 #include <sys/mbuf.h>
75 #include <sys/priv.h>
76 #include <sys/proc.h>
77 #include <sys/protosw.h>
78 #include <sys/signalvar.h>
79 #include <sys/socket.h>
80 #include <sys/socketvar.h>
81 #include <sys/sx.h>
82 #include <sys/syslog.h>
83
84 #include <net/if.h>
85 #include <net/if_var.h>
86 #include <net/if_types.h>
87 #include <net/route.h>
88 #include <net/vnet.h>
89
90 #include <netinet/in.h>
91 #include <netinet/in_var.h>
92 #include <netinet/in_systm.h>
93 #include <netinet/in_pcb.h>
94
95 #include <netinet/icmp6.h>
96 #include <netinet/ip6.h>
97 #include <netinet/ip_var.h>
98 #include <netinet6/ip6protosw.h>
99 #include <netinet6/ip6_mroute.h>
100 #include <netinet6/in6_pcb.h>
101 #include <netinet6/ip6_var.h>
102 #include <netinet6/nd6.h>
103 #include <netinet6/raw_ip6.h>
104 #include <netinet6/scope6_var.h>
105 #include <netinet6/send.h>
106
107 #include <netipsec/ipsec_support.h>
108
109 #include <machine/stdarg.h>
110
111 #define satosin6(sa) ((struct sockaddr_in6 *)(sa))
112 #define ifatoia6(ifa) ((struct in6_ifaddr *)(ifa))
113
114 /*
115 * Raw interface to IP6 protocol.
116 */
117
118 VNET_DECLARE(struct inpcbhead, ripcb);
119 VNET_DECLARE(struct inpcbinfo, ripcbinfo);
120 #define V_ripcb VNET(ripcb)
121 #define V_ripcbinfo VNET(ripcbinfo)
122
123 extern u_long rip_sendspace;
124 extern u_long rip_recvspace;
125
126 VNET_PCPUSTAT_DEFINE(struct rip6stat, rip6stat);
127 VNET_PCPUSTAT_SYSINIT(rip6stat);
128
129 #ifdef VIMAGE
130 VNET_PCPUSTAT_SYSUNINIT(rip6stat);
131 #endif /* VIMAGE */
132
133 /*
134 * Hooks for multicast routing. They all default to NULL, so leave them not
135 * initialized and rely on BSS being set to 0.
136 */
137
138 /*
139 * The socket used to communicate with the multicast routing daemon.
140 */
141 VNET_DEFINE(struct socket *, ip6_mrouter);
142
143 /*
144 * The various mrouter functions.
145 */
146 int (*ip6_mrouter_set)(struct socket *, struct sockopt *);
147 int (*ip6_mrouter_get)(struct socket *, struct sockopt *);
148 int (*ip6_mrouter_done)(void);
149 int (*ip6_mforward)(struct ip6_hdr *, struct ifnet *, struct mbuf *);
150 int (*mrt6_ioctl)(u_long, caddr_t);
151
152 /*
153 * Setup generic address and protocol structures for raw_input routine, then
154 * pass them along with mbuf chain.
155 */
156 int
rip6_input(struct mbuf ** mp,int * offp,int proto)157 rip6_input(struct mbuf **mp, int *offp, int proto)
158 {
159 struct ifnet *ifp;
160 struct mbuf *m = *mp;
161 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
162 struct inpcb *in6p;
163 struct inpcb *last = NULL;
164 struct mbuf *opts = NULL;
165 struct sockaddr_in6 fromsa;
166
167 RIP6STAT_INC(rip6s_ipackets);
168
169 init_sin6(&fromsa, m); /* general init */
170
171 ifp = m->m_pkthdr.rcvif;
172
173 INP_INFO_RLOCK(&V_ripcbinfo);
174 LIST_FOREACH(in6p, &V_ripcb, inp_list) {
175 /* XXX inp locking */
176 if ((in6p->inp_vflag & INP_IPV6) == 0)
177 continue;
178 if (in6p->inp_ip_p &&
179 in6p->inp_ip_p != proto)
180 continue;
181 if (!IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_laddr) &&
182 !IN6_ARE_ADDR_EQUAL(&in6p->in6p_laddr, &ip6->ip6_dst))
183 continue;
184 if (!IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_faddr) &&
185 !IN6_ARE_ADDR_EQUAL(&in6p->in6p_faddr, &ip6->ip6_src))
186 continue;
187 if (jailed_without_vnet(in6p->inp_cred)) {
188 /*
189 * Allow raw socket in jail to receive multicast;
190 * assume process had PRIV_NETINET_RAW at attach,
191 * and fall through into normal filter path if so.
192 */
193 if (!IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) &&
194 prison_check_ip6(in6p->inp_cred,
195 &ip6->ip6_dst) != 0)
196 continue;
197 }
198 INP_RLOCK(in6p);
199 if (in6p->in6p_cksum != -1) {
200 RIP6STAT_INC(rip6s_isum);
201 if (m->m_pkthdr.len - (*offp + in6p->in6p_cksum) < 2 ||
202 in6_cksum(m, proto, *offp,
203 m->m_pkthdr.len - *offp)) {
204 INP_RUNLOCK(in6p);
205 RIP6STAT_INC(rip6s_badsum);
206 continue;
207 }
208 }
209 /*
210 * If this raw socket has multicast state, and we
211 * have received a multicast, check if this socket
212 * should receive it, as multicast filtering is now
213 * the responsibility of the transport layer.
214 */
215 if (in6p->in6p_moptions &&
216 IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) {
217 /*
218 * If the incoming datagram is for MLD, allow it
219 * through unconditionally to the raw socket.
220 *
221 * Use the M_RTALERT_MLD flag to check for MLD
222 * traffic without having to inspect the mbuf chain
223 * more deeply, as all MLDv1/v2 host messages MUST
224 * contain the Router Alert option.
225 *
226 * In the case of MLDv1, we may not have explicitly
227 * joined the group, and may have set IFF_ALLMULTI
228 * on the interface. im6o_mc_filter() may discard
229 * control traffic we actually need to see.
230 *
231 * Userland multicast routing daemons should continue
232 * filter the control traffic appropriately.
233 */
234 int blocked;
235
236 blocked = MCAST_PASS;
237 if ((m->m_flags & M_RTALERT_MLD) == 0) {
238 struct sockaddr_in6 mcaddr;
239
240 bzero(&mcaddr, sizeof(struct sockaddr_in6));
241 mcaddr.sin6_len = sizeof(struct sockaddr_in6);
242 mcaddr.sin6_family = AF_INET6;
243 mcaddr.sin6_addr = ip6->ip6_dst;
244
245 blocked = im6o_mc_filter(in6p->in6p_moptions,
246 ifp,
247 (struct sockaddr *)&mcaddr,
248 (struct sockaddr *)&fromsa);
249 }
250 if (blocked != MCAST_PASS) {
251 IP6STAT_INC(ip6s_notmember);
252 INP_RUNLOCK(in6p);
253 continue;
254 }
255 }
256 if (last != NULL) {
257 struct mbuf *n = m_copy(m, 0, (int)M_COPYALL);
258
259 #if defined(IPSEC) || defined(IPSEC_SUPPORT)
260 /*
261 * Check AH/ESP integrity.
262 */
263 if (IPSEC_ENABLED(ipv6)) {
264 if (n != NULL &&
265 IPSEC_CHECK_POLICY(ipv6, n, last) != 0) {
266 m_freem(n);
267 /* Do not inject data into pcb. */
268 n = NULL;
269 }
270 }
271 #endif /* IPSEC */
272 if (n) {
273 if (last->inp_flags & INP_CONTROLOPTS ||
274 last->inp_socket->so_options & SO_TIMESTAMP)
275 ip6_savecontrol(last, n, &opts);
276 /* strip intermediate headers */
277 m_adj(n, *offp);
278 if (sbappendaddr(&last->inp_socket->so_rcv,
279 (struct sockaddr *)&fromsa,
280 n, opts) == 0) {
281 m_freem(n);
282 if (opts)
283 m_freem(opts);
284 RIP6STAT_INC(rip6s_fullsock);
285 } else
286 sorwakeup(last->inp_socket);
287 opts = NULL;
288 }
289 INP_RUNLOCK(last);
290 }
291 last = in6p;
292 }
293 INP_INFO_RUNLOCK(&V_ripcbinfo);
294 #if defined(IPSEC) || defined(IPSEC_SUPPORT)
295 /*
296 * Check AH/ESP integrity.
297 */
298 if (IPSEC_ENABLED(ipv6) && last != NULL &&
299 IPSEC_CHECK_POLICY(ipv6, m, last) != 0) {
300 m_freem(m);
301 IP6STAT_DEC(ip6s_delivered);
302 /* Do not inject data into pcb. */
303 INP_RUNLOCK(last);
304 } else
305 #endif /* IPSEC */
306 if (last != NULL) {
307 if (last->inp_flags & INP_CONTROLOPTS ||
308 last->inp_socket->so_options & SO_TIMESTAMP)
309 ip6_savecontrol(last, m, &opts);
310 /* Strip intermediate headers. */
311 m_adj(m, *offp);
312 if (sbappendaddr(&last->inp_socket->so_rcv,
313 (struct sockaddr *)&fromsa, m, opts) == 0) {
314 m_freem(m);
315 if (opts)
316 m_freem(opts);
317 RIP6STAT_INC(rip6s_fullsock);
318 } else
319 sorwakeup(last->inp_socket);
320 INP_RUNLOCK(last);
321 } else {
322 RIP6STAT_INC(rip6s_nosock);
323 if (m->m_flags & M_MCAST)
324 RIP6STAT_INC(rip6s_nosockmcast);
325 if (proto == IPPROTO_NONE)
326 m_freem(m);
327 else
328 icmp6_error(m, ICMP6_PARAM_PROB,
329 ICMP6_PARAMPROB_NEXTHEADER,
330 ip6_get_prevhdr(m, *offp));
331 IP6STAT_DEC(ip6s_delivered);
332 }
333 return (IPPROTO_DONE);
334 }
335
336 void
rip6_ctlinput(int cmd,struct sockaddr * sa,void * d)337 rip6_ctlinput(int cmd, struct sockaddr *sa, void *d)
338 {
339 struct ip6_hdr *ip6;
340 struct mbuf *m;
341 int off = 0;
342 struct ip6ctlparam *ip6cp = NULL;
343 const struct sockaddr_in6 *sa6_src = NULL;
344 void *cmdarg;
345 struct inpcb *(*notify)(struct inpcb *, int) = in6_rtchange;
346
347 if (sa->sa_family != AF_INET6 ||
348 sa->sa_len != sizeof(struct sockaddr_in6))
349 return;
350
351 if ((unsigned)cmd >= PRC_NCMDS)
352 return;
353 if (PRC_IS_REDIRECT(cmd))
354 notify = in6_rtchange, d = NULL;
355 else if (cmd == PRC_HOSTDEAD)
356 d = NULL;
357 else if (inet6ctlerrmap[cmd] == 0)
358 return;
359
360 /*
361 * If the parameter is from icmp6, decode it.
362 */
363 if (d != NULL) {
364 ip6cp = (struct ip6ctlparam *)d;
365 m = ip6cp->ip6c_m;
366 ip6 = ip6cp->ip6c_ip6;
367 off = ip6cp->ip6c_off;
368 cmdarg = ip6cp->ip6c_cmdarg;
369 sa6_src = ip6cp->ip6c_src;
370 } else {
371 m = NULL;
372 ip6 = NULL;
373 cmdarg = NULL;
374 sa6_src = &sa6_any;
375 }
376
377 (void) in6_pcbnotify(&V_ripcbinfo, sa, 0,
378 (const struct sockaddr *)sa6_src, 0, cmd, cmdarg, notify);
379 }
380
381 /*
382 * Generate IPv6 header and pass packet to ip6_output. Tack on options user
383 * may have setup with control call.
384 */
385 int
rip6_output(struct mbuf * m,struct socket * so,...)386 rip6_output(struct mbuf *m, struct socket *so, ...)
387 {
388 struct mbuf *control;
389 struct m_tag *mtag;
390 struct sockaddr_in6 *dstsock;
391 struct in6_addr *dst;
392 struct ip6_hdr *ip6;
393 struct inpcb *in6p;
394 u_int plen = m->m_pkthdr.len;
395 int error = 0;
396 struct ip6_pktopts opt, *optp;
397 struct ifnet *oifp = NULL;
398 int type = 0, code = 0; /* for ICMPv6 output statistics only */
399 int scope_ambiguous = 0;
400 int use_defzone = 0;
401 int hlim = 0;
402 struct in6_addr in6a;
403 va_list ap;
404
405 va_start(ap, so);
406 dstsock = va_arg(ap, struct sockaddr_in6 *);
407 control = va_arg(ap, struct mbuf *);
408 va_end(ap);
409
410 in6p = sotoinpcb(so);
411 INP_WLOCK(in6p);
412
413 dst = &dstsock->sin6_addr;
414 if (control != NULL) {
415 if ((error = ip6_setpktopts(control, &opt,
416 in6p->in6p_outputopts, so->so_cred,
417 so->so_proto->pr_protocol)) != 0) {
418 goto bad;
419 }
420 optp = &opt;
421 } else
422 optp = in6p->in6p_outputopts;
423
424 /*
425 * Check and convert scope zone ID into internal form.
426 *
427 * XXX: we may still need to determine the zone later.
428 */
429 if (!(so->so_state & SS_ISCONNECTED)) {
430 if (!optp || !optp->ip6po_pktinfo ||
431 !optp->ip6po_pktinfo->ipi6_ifindex)
432 use_defzone = V_ip6_use_defzone;
433 if (dstsock->sin6_scope_id == 0 && !use_defzone)
434 scope_ambiguous = 1;
435 if ((error = sa6_embedscope(dstsock, use_defzone)) != 0)
436 goto bad;
437 }
438
439 /*
440 * For an ICMPv6 packet, we should know its type and code to update
441 * statistics.
442 */
443 if (so->so_proto->pr_protocol == IPPROTO_ICMPV6) {
444 struct icmp6_hdr *icmp6;
445 if (m->m_len < sizeof(struct icmp6_hdr) &&
446 (m = m_pullup(m, sizeof(struct icmp6_hdr))) == NULL) {
447 error = ENOBUFS;
448 goto bad;
449 }
450 icmp6 = mtod(m, struct icmp6_hdr *);
451 type = icmp6->icmp6_type;
452 code = icmp6->icmp6_code;
453 }
454
455 M_PREPEND(m, sizeof(*ip6), M_NOWAIT);
456 if (m == NULL) {
457 error = ENOBUFS;
458 goto bad;
459 }
460 ip6 = mtod(m, struct ip6_hdr *);
461
462 /*
463 * Source address selection.
464 */
465 error = in6_selectsrc_socket(dstsock, optp, in6p, so->so_cred,
466 scope_ambiguous, &in6a, &hlim);
467
468 if (error)
469 goto bad;
470 error = prison_check_ip6(in6p->inp_cred, &in6a);
471 if (error != 0)
472 goto bad;
473 ip6->ip6_src = in6a;
474
475 ip6->ip6_dst = dstsock->sin6_addr;
476
477 /*
478 * Fill in the rest of the IPv6 header fields.
479 */
480 ip6->ip6_flow = (ip6->ip6_flow & ~IPV6_FLOWINFO_MASK) |
481 (in6p->inp_flow & IPV6_FLOWINFO_MASK);
482 ip6->ip6_vfc = (ip6->ip6_vfc & ~IPV6_VERSION_MASK) |
483 (IPV6_VERSION & IPV6_VERSION_MASK);
484
485 /*
486 * ip6_plen will be filled in ip6_output, so not fill it here.
487 */
488 ip6->ip6_nxt = in6p->inp_ip_p;
489 ip6->ip6_hlim = hlim;
490
491 if (so->so_proto->pr_protocol == IPPROTO_ICMPV6 ||
492 in6p->in6p_cksum != -1) {
493 struct mbuf *n;
494 int off;
495 u_int16_t *p;
496
497 /* Compute checksum. */
498 if (so->so_proto->pr_protocol == IPPROTO_ICMPV6)
499 off = offsetof(struct icmp6_hdr, icmp6_cksum);
500 else
501 off = in6p->in6p_cksum;
502 if (plen < off + 2) {
503 error = EINVAL;
504 goto bad;
505 }
506 off += sizeof(struct ip6_hdr);
507
508 n = m;
509 while (n && n->m_len <= off) {
510 off -= n->m_len;
511 n = n->m_next;
512 }
513 if (!n)
514 goto bad;
515 p = (u_int16_t *)(mtod(n, caddr_t) + off);
516 *p = 0;
517 *p = in6_cksum(m, ip6->ip6_nxt, sizeof(*ip6), plen);
518 }
519
520 /*
521 * Send RA/RS messages to user land for protection, before sending
522 * them to rtadvd/rtsol.
523 */
524 if ((send_sendso_input_hook != NULL) &&
525 so->so_proto->pr_protocol == IPPROTO_ICMPV6) {
526 switch (type) {
527 case ND_ROUTER_ADVERT:
528 case ND_ROUTER_SOLICIT:
529 mtag = m_tag_get(PACKET_TAG_ND_OUTGOING,
530 sizeof(unsigned short), M_NOWAIT);
531 if (mtag == NULL)
532 goto bad;
533 m_tag_prepend(m, mtag);
534 }
535 }
536
537 error = ip6_output(m, optp, NULL, 0, in6p->in6p_moptions, &oifp, in6p);
538 if (so->so_proto->pr_protocol == IPPROTO_ICMPV6) {
539 if (oifp)
540 icmp6_ifoutstat_inc(oifp, type, code);
541 ICMP6STAT_INC(icp6s_outhist[type]);
542 } else
543 RIP6STAT_INC(rip6s_opackets);
544
545 goto freectl;
546
547 bad:
548 if (m)
549 m_freem(m);
550
551 freectl:
552 if (control != NULL) {
553 ip6_clearpktopts(&opt, -1);
554 m_freem(control);
555 }
556 INP_WUNLOCK(in6p);
557 return (error);
558 }
559
560 /*
561 * Raw IPv6 socket option processing.
562 */
563 int
rip6_ctloutput(struct socket * so,struct sockopt * sopt)564 rip6_ctloutput(struct socket *so, struct sockopt *sopt)
565 {
566 struct inpcb *inp;
567 int error;
568
569 if (sopt->sopt_level == IPPROTO_ICMPV6)
570 /*
571 * XXX: is it better to call icmp6_ctloutput() directly
572 * from protosw?
573 */
574 return (icmp6_ctloutput(so, sopt));
575 else if (sopt->sopt_level != IPPROTO_IPV6) {
576 if (sopt->sopt_level == SOL_SOCKET &&
577 sopt->sopt_name == SO_SETFIB) {
578 inp = sotoinpcb(so);
579 INP_WLOCK(inp);
580 inp->inp_inc.inc_fibnum = so->so_fibnum;
581 INP_WUNLOCK(inp);
582 return (0);
583 }
584 return (EINVAL);
585 }
586
587 error = 0;
588
589 switch (sopt->sopt_dir) {
590 case SOPT_GET:
591 switch (sopt->sopt_name) {
592 case MRT6_INIT:
593 case MRT6_DONE:
594 case MRT6_ADD_MIF:
595 case MRT6_DEL_MIF:
596 case MRT6_ADD_MFC:
597 case MRT6_DEL_MFC:
598 case MRT6_PIM:
599 error = ip6_mrouter_get ? ip6_mrouter_get(so, sopt) :
600 EOPNOTSUPP;
601 break;
602 case IPV6_CHECKSUM:
603 error = ip6_raw_ctloutput(so, sopt);
604 break;
605 default:
606 error = ip6_ctloutput(so, sopt);
607 break;
608 }
609 break;
610
611 case SOPT_SET:
612 switch (sopt->sopt_name) {
613 case MRT6_INIT:
614 case MRT6_DONE:
615 case MRT6_ADD_MIF:
616 case MRT6_DEL_MIF:
617 case MRT6_ADD_MFC:
618 case MRT6_DEL_MFC:
619 case MRT6_PIM:
620 error = ip6_mrouter_set ? ip6_mrouter_set(so, sopt) :
621 EOPNOTSUPP;
622 break;
623 case IPV6_CHECKSUM:
624 error = ip6_raw_ctloutput(so, sopt);
625 break;
626 default:
627 error = ip6_ctloutput(so, sopt);
628 break;
629 }
630 break;
631 }
632
633 return (error);
634 }
635
636 static int
rip6_attach(struct socket * so,int proto,struct thread * td)637 rip6_attach(struct socket *so, int proto, struct thread *td)
638 {
639 struct inpcb *inp;
640 struct icmp6_filter *filter;
641 int error;
642
643 inp = sotoinpcb(so);
644 KASSERT(inp == NULL, ("rip6_attach: inp != NULL"));
645
646 error = priv_check(td, PRIV_NETINET_RAW);
647 if (error)
648 return (error);
649 error = soreserve(so, rip_sendspace, rip_recvspace);
650 if (error)
651 return (error);
652 filter = malloc(sizeof(struct icmp6_filter), M_PCB, M_NOWAIT);
653 if (filter == NULL)
654 return (ENOMEM);
655 INP_INFO_WLOCK(&V_ripcbinfo);
656 error = in_pcballoc(so, &V_ripcbinfo);
657 if (error) {
658 INP_INFO_WUNLOCK(&V_ripcbinfo);
659 free(filter, M_PCB);
660 return (error);
661 }
662 inp = (struct inpcb *)so->so_pcb;
663 INP_INFO_WUNLOCK(&V_ripcbinfo);
664 inp->inp_vflag |= INP_IPV6;
665 inp->inp_ip_p = (long)proto;
666 inp->in6p_hops = -1; /* use kernel default */
667 inp->in6p_cksum = -1;
668 inp->in6p_icmp6filt = filter;
669 ICMP6_FILTER_SETPASSALL(inp->in6p_icmp6filt);
670 INP_WUNLOCK(inp);
671 return (0);
672 }
673
674 static void
rip6_detach(struct socket * so)675 rip6_detach(struct socket *so)
676 {
677 struct inpcb *inp;
678
679 inp = sotoinpcb(so);
680 KASSERT(inp != NULL, ("rip6_detach: inp == NULL"));
681
682 if (so == V_ip6_mrouter && ip6_mrouter_done)
683 ip6_mrouter_done();
684 /* xxx: RSVP */
685 INP_INFO_WLOCK(&V_ripcbinfo);
686 INP_WLOCK(inp);
687 free(inp->in6p_icmp6filt, M_PCB);
688 in_pcbdetach(inp);
689 in_pcbfree(inp);
690 INP_INFO_WUNLOCK(&V_ripcbinfo);
691 }
692
693 /* XXXRW: This can't ever be called. */
694 static void
rip6_abort(struct socket * so)695 rip6_abort(struct socket *so)
696 {
697 struct inpcb *inp;
698
699 inp = sotoinpcb(so);
700 KASSERT(inp != NULL, ("rip6_abort: inp == NULL"));
701
702 soisdisconnected(so);
703 }
704
705 static void
rip6_close(struct socket * so)706 rip6_close(struct socket *so)
707 {
708 struct inpcb *inp;
709
710 inp = sotoinpcb(so);
711 KASSERT(inp != NULL, ("rip6_close: inp == NULL"));
712
713 soisdisconnected(so);
714 }
715
716 static int
rip6_disconnect(struct socket * so)717 rip6_disconnect(struct socket *so)
718 {
719 struct inpcb *inp;
720
721 inp = sotoinpcb(so);
722 KASSERT(inp != NULL, ("rip6_disconnect: inp == NULL"));
723
724 if ((so->so_state & SS_ISCONNECTED) == 0)
725 return (ENOTCONN);
726 inp->in6p_faddr = in6addr_any;
727 rip6_abort(so);
728 return (0);
729 }
730
731 static int
rip6_bind(struct socket * so,struct sockaddr * nam,struct thread * td)732 rip6_bind(struct socket *so, struct sockaddr *nam, struct thread *td)
733 {
734 struct inpcb *inp;
735 struct sockaddr_in6 *addr = (struct sockaddr_in6 *)nam;
736 struct ifaddr *ifa = NULL;
737 int error = 0;
738
739 inp = sotoinpcb(so);
740 KASSERT(inp != NULL, ("rip6_bind: inp == NULL"));
741
742 if (nam->sa_len != sizeof(*addr))
743 return (EINVAL);
744 if ((error = prison_check_ip6(td->td_ucred, &addr->sin6_addr)) != 0)
745 return (error);
746 if (TAILQ_EMPTY(&V_ifnet) || addr->sin6_family != AF_INET6)
747 return (EADDRNOTAVAIL);
748 if ((error = sa6_embedscope(addr, V_ip6_use_defzone)) != 0)
749 return (error);
750
751 if (!IN6_IS_ADDR_UNSPECIFIED(&addr->sin6_addr) &&
752 (ifa = ifa_ifwithaddr((struct sockaddr *)addr)) == NULL)
753 return (EADDRNOTAVAIL);
754 if (ifa != NULL &&
755 ((struct in6_ifaddr *)ifa)->ia6_flags &
756 (IN6_IFF_ANYCAST|IN6_IFF_NOTREADY|
757 IN6_IFF_DETACHED|IN6_IFF_DEPRECATED)) {
758 ifa_free(ifa);
759 return (EADDRNOTAVAIL);
760 }
761 if (ifa != NULL)
762 ifa_free(ifa);
763 INP_INFO_WLOCK(&V_ripcbinfo);
764 INP_WLOCK(inp);
765 inp->in6p_laddr = addr->sin6_addr;
766 INP_WUNLOCK(inp);
767 INP_INFO_WUNLOCK(&V_ripcbinfo);
768 return (0);
769 }
770
771 static int
rip6_connect(struct socket * so,struct sockaddr * nam,struct thread * td)772 rip6_connect(struct socket *so, struct sockaddr *nam, struct thread *td)
773 {
774 struct inpcb *inp;
775 struct sockaddr_in6 *addr = (struct sockaddr_in6 *)nam;
776 struct in6_addr in6a;
777 int error = 0, scope_ambiguous = 0;
778
779 inp = sotoinpcb(so);
780 KASSERT(inp != NULL, ("rip6_connect: inp == NULL"));
781
782 if (nam->sa_len != sizeof(*addr))
783 return (EINVAL);
784 if (TAILQ_EMPTY(&V_ifnet))
785 return (EADDRNOTAVAIL);
786 if (addr->sin6_family != AF_INET6)
787 return (EAFNOSUPPORT);
788
789 /*
790 * Application should provide a proper zone ID or the use of default
791 * zone IDs should be enabled. Unfortunately, some applications do
792 * not behave as it should, so we need a workaround. Even if an
793 * appropriate ID is not determined, we'll see if we can determine
794 * the outgoing interface. If we can, determine the zone ID based on
795 * the interface below.
796 */
797 if (addr->sin6_scope_id == 0 && !V_ip6_use_defzone)
798 scope_ambiguous = 1;
799 if ((error = sa6_embedscope(addr, V_ip6_use_defzone)) != 0)
800 return (error);
801
802 INP_INFO_WLOCK(&V_ripcbinfo);
803 INP_WLOCK(inp);
804 /* Source address selection. XXX: need pcblookup? */
805 error = in6_selectsrc_socket(addr, inp->in6p_outputopts,
806 inp, so->so_cred, scope_ambiguous, &in6a, NULL);
807 if (error) {
808 INP_WUNLOCK(inp);
809 INP_INFO_WUNLOCK(&V_ripcbinfo);
810 return (error);
811 }
812
813 inp->in6p_faddr = addr->sin6_addr;
814 inp->in6p_laddr = in6a;
815 soisconnected(so);
816 INP_WUNLOCK(inp);
817 INP_INFO_WUNLOCK(&V_ripcbinfo);
818 return (0);
819 }
820
821 static int
rip6_shutdown(struct socket * so)822 rip6_shutdown(struct socket *so)
823 {
824 struct inpcb *inp;
825
826 inp = sotoinpcb(so);
827 KASSERT(inp != NULL, ("rip6_shutdown: inp == NULL"));
828
829 INP_WLOCK(inp);
830 socantsendmore(so);
831 INP_WUNLOCK(inp);
832 return (0);
833 }
834
835 static int
rip6_send(struct socket * so,int flags,struct mbuf * m,struct sockaddr * nam,struct mbuf * control,struct thread * td)836 rip6_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam,
837 struct mbuf *control, struct thread *td)
838 {
839 struct inpcb *inp;
840 struct sockaddr_in6 tmp;
841 struct sockaddr_in6 *dst;
842 int ret;
843
844 inp = sotoinpcb(so);
845 KASSERT(inp != NULL, ("rip6_send: inp == NULL"));
846
847 /* Always copy sockaddr to avoid overwrites. */
848 /* Unlocked read. */
849 if (so->so_state & SS_ISCONNECTED) {
850 if (nam) {
851 m_freem(m);
852 return (EISCONN);
853 }
854 /* XXX */
855 bzero(&tmp, sizeof(tmp));
856 tmp.sin6_family = AF_INET6;
857 tmp.sin6_len = sizeof(struct sockaddr_in6);
858 INP_RLOCK(inp);
859 bcopy(&inp->in6p_faddr, &tmp.sin6_addr,
860 sizeof(struct in6_addr));
861 INP_RUNLOCK(inp);
862 dst = &tmp;
863 } else {
864 if (nam == NULL) {
865 m_freem(m);
866 return (ENOTCONN);
867 }
868 if (nam->sa_len != sizeof(struct sockaddr_in6)) {
869 m_freem(m);
870 return (EINVAL);
871 }
872 tmp = *(struct sockaddr_in6 *)nam;
873 dst = &tmp;
874
875 if (dst->sin6_family == AF_UNSPEC) {
876 /*
877 * XXX: we allow this case for backward
878 * compatibility to buggy applications that
879 * rely on old (and wrong) kernel behavior.
880 */
881 log(LOG_INFO, "rip6 SEND: address family is "
882 "unspec. Assume AF_INET6\n");
883 dst->sin6_family = AF_INET6;
884 } else if (dst->sin6_family != AF_INET6) {
885 m_freem(m);
886 return(EAFNOSUPPORT);
887 }
888 }
889 ret = rip6_output(m, so, dst, control);
890 return (ret);
891 }
892
893 struct pr_usrreqs rip6_usrreqs = {
894 .pru_abort = rip6_abort,
895 .pru_attach = rip6_attach,
896 .pru_bind = rip6_bind,
897 .pru_connect = rip6_connect,
898 .pru_control = in6_control,
899 .pru_detach = rip6_detach,
900 .pru_disconnect = rip6_disconnect,
901 .pru_peeraddr = in6_getpeeraddr,
902 .pru_send = rip6_send,
903 .pru_shutdown = rip6_shutdown,
904 .pru_sockaddr = in6_getsockaddr,
905 .pru_close = rip6_close,
906 };
907