raw_ip6.c revision 1.154.2.1 1 /* $NetBSD: raw_ip6.c,v 1.154.2.1 2017/04/21 16:54:06 bouyer Exp $ */
2 /* $KAME: raw_ip6.c,v 1.82 2001/07/23 18:57:56 jinmei Exp $ */
3
4 /*
5 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
6 * All rights reserved.
7 *
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
10 * are met:
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.
19 *
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
30 * SUCH DAMAGE.
31 */
32
33 /*
34 * Copyright (c) 1982, 1986, 1988, 1993
35 * The Regents of the University of California. All rights reserved.
36 *
37 * Redistribution and use in source and binary forms, with or without
38 * modification, are permitted provided that the following conditions
39 * are met:
40 * 1. Redistributions of source code must retain the above copyright
41 * notice, this list of conditions and the following disclaimer.
42 * 2. Redistributions in binary form must reproduce the above copyright
43 * notice, this list of conditions and the following disclaimer in the
44 * documentation and/or other materials provided with the distribution.
45 * 3. Neither the name of the University nor the names of its contributors
46 * may be used to endorse or promote products derived from this software
47 * without specific prior written permission.
48 *
49 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
50 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
51 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
52 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
53 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
54 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
55 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
56 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
57 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
58 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
59 * SUCH DAMAGE.
60 *
61 * @(#)raw_ip.c 8.2 (Berkeley) 1/4/94
62 */
63
64 #include <sys/cdefs.h>
65 __KERNEL_RCSID(0, "$NetBSD: raw_ip6.c,v 1.154.2.1 2017/04/21 16:54:06 bouyer Exp $");
66
67 #ifdef _KERNEL_OPT
68 #include "opt_ipsec.h"
69 #include "opt_net_mpsafe.h"
70 #endif
71
72 #include <sys/param.h>
73 #include <sys/sysctl.h>
74 #include <sys/mbuf.h>
75 #include <sys/socket.h>
76 #include <sys/protosw.h>
77 #include <sys/socketvar.h>
78 #include <sys/systm.h>
79 #include <sys/proc.h>
80 #include <sys/kauth.h>
81 #include <sys/kmem.h>
82
83 #include <net/if.h>
84 #include <net/if_types.h>
85 #include <net/net_stats.h>
86
87 #include <netinet/in.h>
88 #include <netinet/in_var.h>
89 #include <netinet/ip6.h>
90 #include <netinet6/ip6_var.h>
91 #include <netinet6/ip6_private.h>
92 #include <netinet6/ip6_mroute.h>
93 #include <netinet/icmp6.h>
94 #include <netinet6/icmp6_private.h>
95 #include <netinet6/in6_pcb.h>
96 #include <netinet6/ip6protosw.h>
97 #include <netinet6/scope6_var.h>
98 #include <netinet6/raw_ip6.h>
99
100 #ifdef IPSEC
101 #include <netipsec/ipsec.h>
102 #include <netipsec/ipsec_var.h>
103 #include <netipsec/ipsec_private.h>
104 #include <netipsec/ipsec6.h>
105 #endif
106
107 #include "faith.h"
108 #if defined(NFAITH) && 0 < NFAITH
109 #include <net/if_faith.h>
110 #endif
111
112 extern struct inpcbtable rawcbtable;
113 struct inpcbtable raw6cbtable;
114 #define ifatoia6(ifa) ((struct in6_ifaddr *)(ifa))
115
116 /*
117 * Raw interface to IP6 protocol.
118 */
119
120 static percpu_t *rip6stat_percpu;
121
122 #define RIP6_STATINC(x) _NET_STATINC(rip6stat_percpu, x)
123
124 static void sysctl_net_inet6_raw6_setup(struct sysctllog **);
125
126 /*
127 * Initialize raw connection block queue.
128 */
129 void
130 rip6_init(void)
131 {
132
133 sysctl_net_inet6_raw6_setup(NULL);
134 in6_pcbinit(&raw6cbtable, 1, 1);
135
136 rip6stat_percpu = percpu_alloc(sizeof(uint64_t) * RIP6_NSTATS);
137 }
138
139 /*
140 * Setup generic address and protocol structures
141 * for raw_input routine, then pass them along with
142 * mbuf chain.
143 */
144 int
145 rip6_input(struct mbuf **mp, int *offp, int proto)
146 {
147 struct mbuf *m = *mp;
148 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
149 struct inpcb_hdr *inph;
150 struct in6pcb *in6p;
151 struct in6pcb *last = NULL;
152 struct sockaddr_in6 rip6src;
153 struct mbuf *opts = NULL;
154
155 RIP6_STATINC(RIP6_STAT_IPACKETS);
156
157 #if defined(NFAITH) && 0 < NFAITH
158 if (faithprefix(&ip6->ip6_dst)) {
159 /* send icmp6 host unreach? */
160 m_freem(m);
161 return IPPROTO_DONE;
162 }
163 #endif
164
165 /* Be proactive about malicious use of IPv4 mapped address */
166 if (IN6_IS_ADDR_V4MAPPED(&ip6->ip6_src) ||
167 IN6_IS_ADDR_V4MAPPED(&ip6->ip6_dst)) {
168 /* XXX stat */
169 m_freem(m);
170 return IPPROTO_DONE;
171 }
172
173 sockaddr_in6_init(&rip6src, &ip6->ip6_src, 0, 0, 0);
174 if (sa6_recoverscope(&rip6src) != 0) {
175 /* XXX: should be impossible. */
176 m_freem(m);
177 return IPPROTO_DONE;
178 }
179
180 TAILQ_FOREACH(inph, &raw6cbtable.inpt_queue, inph_queue) {
181 in6p = (struct in6pcb *)inph;
182 if (in6p->in6p_af != AF_INET6)
183 continue;
184 if (in6p->in6p_ip6.ip6_nxt &&
185 in6p->in6p_ip6.ip6_nxt != proto)
186 continue;
187 if (!IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_laddr) &&
188 !IN6_ARE_ADDR_EQUAL(&in6p->in6p_laddr, &ip6->ip6_dst))
189 continue;
190 if (!IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_faddr) &&
191 !IN6_ARE_ADDR_EQUAL(&in6p->in6p_faddr, &ip6->ip6_src))
192 continue;
193 if (in6p->in6p_cksum != -1) {
194 RIP6_STATINC(RIP6_STAT_ISUM);
195 if (in6_cksum(m, proto, *offp,
196 m->m_pkthdr.len - *offp)) {
197 RIP6_STATINC(RIP6_STAT_BADSUM);
198 continue;
199 }
200 }
201 if (last) {
202 struct mbuf *n;
203
204 #ifdef IPSEC
205 /*
206 * Check AH/ESP integrity
207 */
208 if (ipsec_used && !ipsec6_in_reject(m, last))
209 #endif /* IPSEC */
210 if ((n = m_copy(m, 0, (int)M_COPYALL)) != NULL) {
211 if (last->in6p_flags & IN6P_CONTROLOPTS)
212 ip6_savecontrol(last, &opts, ip6, n);
213 /* strip intermediate headers */
214 m_adj(n, *offp);
215 if (sbappendaddr(&last->in6p_socket->so_rcv,
216 sin6tosa(&rip6src), n, opts) == 0) {
217 /* should notify about lost packet */
218 m_freem(n);
219 if (opts)
220 m_freem(opts);
221 RIP6_STATINC(RIP6_STAT_FULLSOCK);
222 } else
223 sorwakeup(last->in6p_socket);
224 opts = NULL;
225 }
226 }
227 last = in6p;
228 }
229 #ifdef IPSEC
230 if (ipsec_used && last && ipsec6_in_reject(m, last)) {
231 m_freem(m);
232 /*
233 * XXX ipsec6_in_reject update stat if there is an error
234 * so we just need to update stats by hand in the case of last is
235 * NULL
236 */
237 if (!last)
238 IPSEC6_STATINC(IPSEC_STAT_IN_POLVIO);
239 IP6_STATDEC(IP6_STAT_DELIVERED);
240 /* do not inject data into pcb */
241 } else
242 #endif /* IPSEC */
243 if (last) {
244 if (last->in6p_flags & IN6P_CONTROLOPTS)
245 ip6_savecontrol(last, &opts, ip6, m);
246 /* strip intermediate headers */
247 m_adj(m, *offp);
248 if (sbappendaddr(&last->in6p_socket->so_rcv,
249 sin6tosa(&rip6src), m, opts) == 0) {
250 m_freem(m);
251 if (opts)
252 m_freem(opts);
253 RIP6_STATINC(RIP6_STAT_FULLSOCK);
254 } else
255 sorwakeup(last->in6p_socket);
256 } else {
257 RIP6_STATINC(RIP6_STAT_NOSOCK);
258 if (m->m_flags & M_MCAST)
259 RIP6_STATINC(RIP6_STAT_NOSOCKMCAST);
260 if (proto == IPPROTO_NONE)
261 m_freem(m);
262 else {
263 int s;
264 struct ifnet *rcvif = m_get_rcvif(m, &s);
265 u_int8_t *prvnxtp = ip6_get_prevhdr(m, *offp); /* XXX */
266 in6_ifstat_inc(rcvif, ifs6_in_protounknown);
267 m_put_rcvif(rcvif, &s);
268 icmp6_error(m, ICMP6_PARAM_PROB,
269 ICMP6_PARAMPROB_NEXTHEADER,
270 prvnxtp - mtod(m, u_int8_t *));
271 }
272 IP6_STATDEC(IP6_STAT_DELIVERED);
273 }
274 return IPPROTO_DONE;
275 }
276
277 void *
278 rip6_ctlinput(int cmd, const struct sockaddr *sa, void *d)
279 {
280 struct ip6_hdr *ip6;
281 struct ip6ctlparam *ip6cp = NULL;
282 const struct sockaddr_in6 *sa6_src = NULL;
283 void *cmdarg;
284 void (*notify)(struct in6pcb *, int) = in6_rtchange;
285 int nxt;
286
287 if (sa->sa_family != AF_INET6 ||
288 sa->sa_len != sizeof(struct sockaddr_in6))
289 return NULL;
290
291 if ((unsigned)cmd >= PRC_NCMDS)
292 return NULL;
293 if (PRC_IS_REDIRECT(cmd))
294 notify = in6_rtchange, d = NULL;
295 else if (cmd == PRC_HOSTDEAD)
296 d = NULL;
297 else if (cmd == PRC_MSGSIZE)
298 ; /* special code is present, see below */
299 else if (inet6ctlerrmap[cmd] == 0)
300 return NULL;
301
302 /* if the parameter is from icmp6, decode it. */
303 if (d != NULL) {
304 ip6cp = (struct ip6ctlparam *)d;
305 ip6 = ip6cp->ip6c_ip6;
306 cmdarg = ip6cp->ip6c_cmdarg;
307 sa6_src = ip6cp->ip6c_src;
308 nxt = ip6cp->ip6c_nxt;
309 } else {
310 ip6 = NULL;
311 cmdarg = NULL;
312 sa6_src = &sa6_any;
313 nxt = -1;
314 }
315
316 if (ip6 && cmd == PRC_MSGSIZE) {
317 const struct sockaddr_in6 *sa6 = (const struct sockaddr_in6 *)sa;
318 int valid = 0;
319 struct in6pcb *in6p;
320
321 /*
322 * Check to see if we have a valid raw IPv6 socket
323 * corresponding to the address in the ICMPv6 message
324 * payload, and the protocol (ip6_nxt) meets the socket.
325 * XXX chase extension headers, or pass final nxt value
326 * from icmp6_notify_error()
327 */
328 in6p = NULL;
329 in6p = in6_pcblookup_connect(&raw6cbtable, &sa6->sin6_addr, 0,
330 (const struct in6_addr *)&sa6_src->sin6_addr, 0, 0, 0);
331 #if 0
332 if (!in6p) {
333 /*
334 * As the use of sendto(2) is fairly popular,
335 * we may want to allow non-connected pcb too.
336 * But it could be too weak against attacks...
337 * We should at least check if the local
338 * address (= s) is really ours.
339 */
340 in6p = in6_pcblookup_bind(&raw6cbtable,
341 &sa6->sin6_addr, 0, 0);
342 }
343 #endif
344
345 if (in6p && in6p->in6p_ip6.ip6_nxt &&
346 in6p->in6p_ip6.ip6_nxt == nxt)
347 valid++;
348
349 /*
350 * Depending on the value of "valid" and routing table
351 * size (mtudisc_{hi,lo}wat), we will:
352 * - recalculate the new MTU and create the
353 * corresponding routing entry, or
354 * - ignore the MTU change notification.
355 */
356 icmp6_mtudisc_update((struct ip6ctlparam *)d, valid);
357
358 /*
359 * regardless of if we called icmp6_mtudisc_update(),
360 * we need to call in6_pcbnotify(), to notify path MTU
361 * change to the userland (RFC3542), because some
362 * unconnected sockets may share the same destination
363 * and want to know the path MTU.
364 */
365 }
366
367 (void) in6_pcbnotify(&raw6cbtable, sa, 0,
368 sin6tocsa(sa6_src), 0, cmd, cmdarg, notify);
369 return NULL;
370 }
371
372 /*
373 * Generate IPv6 header and pass packet to ip6_output.
374 * Tack on options user may have setup with control call.
375 */
376 int
377 rip6_output(struct mbuf *m, struct socket * const so,
378 struct sockaddr_in6 * const dstsock, struct mbuf * const control)
379 {
380 struct in6_addr *dst;
381 struct ip6_hdr *ip6;
382 struct in6pcb *in6p;
383 u_int plen = m->m_pkthdr.len;
384 int error = 0;
385 struct ip6_pktopts opt, *optp = NULL;
386 struct ifnet *oifp = NULL;
387 int type, code; /* for ICMPv6 output statistics only */
388 int scope_ambiguous = 0;
389 int bound = curlwp_bind();
390 struct psref psref;
391
392 in6p = sotoin6pcb(so);
393
394 dst = &dstsock->sin6_addr;
395 if (control) {
396 if ((error = ip6_setpktopts(control, &opt,
397 in6p->in6p_outputopts,
398 kauth_cred_get(), so->so_proto->pr_protocol)) != 0) {
399 goto bad;
400 }
401 optp = &opt;
402 } else
403 optp = in6p->in6p_outputopts;
404
405 /*
406 * Check and convert scope zone ID into internal form.
407 * XXX: we may still need to determine the zone later.
408 */
409 if (!(so->so_state & SS_ISCONNECTED)) {
410 if (dstsock->sin6_scope_id == 0 && !ip6_use_defzone)
411 scope_ambiguous = 1;
412 if ((error = sa6_embedscope(dstsock, ip6_use_defzone)) != 0)
413 goto bad;
414 }
415
416 /*
417 * For an ICMPv6 packet, we should know its type and code
418 * to update statistics.
419 */
420 if (so->so_proto->pr_protocol == IPPROTO_ICMPV6) {
421 struct icmp6_hdr *icmp6;
422 if (m->m_len < sizeof(struct icmp6_hdr) &&
423 (m = m_pullup(m, sizeof(struct icmp6_hdr))) == NULL) {
424 error = ENOBUFS;
425 goto bad;
426 }
427 icmp6 = mtod(m, struct icmp6_hdr *);
428 type = icmp6->icmp6_type;
429 code = icmp6->icmp6_code;
430 } else {
431 type = 0;
432 code = 0;
433 }
434
435 M_PREPEND(m, sizeof(*ip6), M_DONTWAIT);
436 if (!m) {
437 error = ENOBUFS;
438 goto bad;
439 }
440 ip6 = mtod(m, struct ip6_hdr *);
441
442 /*
443 * Next header might not be ICMP6 but use its pseudo header anyway.
444 */
445 ip6->ip6_dst = *dst;
446
447 /*
448 * Source address selection.
449 */
450 error = in6_selectsrc(dstsock, optp, in6p->in6p_moptions,
451 &in6p->in6p_route, &in6p->in6p_laddr, &oifp, &psref, &ip6->ip6_src);
452 if (error != 0)
453 goto bad;
454
455 if (oifp && scope_ambiguous) {
456 /*
457 * Application should provide a proper zone ID or the use of
458 * default zone IDs should be enabled. Unfortunately, some
459 * applications do not behave as it should, so we need a
460 * workaround. Even if an appropriate ID is not determined
461 * (when it's required), if we can determine the outgoing
462 * interface. determine the zone ID based on the interface.
463 */
464 error = in6_setscope(&dstsock->sin6_addr, oifp, NULL);
465 if (error != 0)
466 goto bad;
467 }
468 ip6->ip6_dst = dstsock->sin6_addr;
469
470 /* fill in the rest of the IPv6 header fields */
471 ip6->ip6_flow = in6p->in6p_flowinfo & IPV6_FLOWINFO_MASK;
472 ip6->ip6_vfc &= ~IPV6_VERSION_MASK;
473 ip6->ip6_vfc |= IPV6_VERSION;
474 /* ip6_plen will be filled in ip6_output, so not fill it here. */
475 ip6->ip6_nxt = in6p->in6p_ip6.ip6_nxt;
476 ip6->ip6_hlim = in6_selecthlim(in6p, oifp);
477
478 if_put(oifp, &psref);
479 oifp = NULL;
480
481 if (so->so_proto->pr_protocol == IPPROTO_ICMPV6 ||
482 in6p->in6p_cksum != -1) {
483 int off;
484 u_int16_t sum;
485
486 /* compute checksum */
487 if (so->so_proto->pr_protocol == IPPROTO_ICMPV6)
488 off = offsetof(struct icmp6_hdr, icmp6_cksum);
489 else
490 off = in6p->in6p_cksum;
491 if (plen < off + 1) {
492 error = EINVAL;
493 goto bad;
494 }
495 off += sizeof(struct ip6_hdr);
496
497 sum = 0;
498 m = m_copyback_cow(m, off, sizeof(sum), (void *)&sum,
499 M_DONTWAIT);
500 if (m == NULL) {
501 error = ENOBUFS;
502 goto bad;
503 }
504 sum = in6_cksum(m, ip6->ip6_nxt, sizeof(*ip6), plen);
505 m = m_copyback_cow(m, off, sizeof(sum), (void *)&sum,
506 M_DONTWAIT);
507 if (m == NULL) {
508 error = ENOBUFS;
509 goto bad;
510 }
511 }
512
513 {
514 struct ifnet *ret_oifp = NULL;
515
516 error = ip6_output(m, optp, &in6p->in6p_route, 0,
517 in6p->in6p_moptions, in6p, &ret_oifp);
518 if (so->so_proto->pr_protocol == IPPROTO_ICMPV6) {
519 if (ret_oifp)
520 icmp6_ifoutstat_inc(ret_oifp, type, code);
521 ICMP6_STATINC(ICMP6_STAT_OUTHIST + type);
522 } else
523 RIP6_STATINC(RIP6_STAT_OPACKETS);
524 }
525
526 goto freectl;
527
528 bad:
529 if (m)
530 m_freem(m);
531
532 freectl:
533 if (control) {
534 ip6_clearpktopts(&opt, -1);
535 m_freem(control);
536 }
537 if_put(oifp, &psref);
538 curlwp_bindx(bound);
539 return error;
540 }
541
542 /*
543 * Raw IPv6 socket option processing.
544 */
545 int
546 rip6_ctloutput(int op, struct socket *so, struct sockopt *sopt)
547 {
548 int error = 0;
549
550 if (sopt->sopt_level == SOL_SOCKET && sopt->sopt_name == SO_NOHEADER) {
551 int optval;
552
553 /* need to fiddle w/ opt(IPPROTO_IPV6, IPV6_CHECKSUM)? */
554 if (op == PRCO_GETOPT) {
555 optval = 1;
556 error = sockopt_set(sopt, &optval, sizeof(optval));
557 } else if (op == PRCO_SETOPT) {
558 error = sockopt_getint(sopt, &optval);
559 if (error)
560 goto out;
561 if (optval == 0)
562 error = EINVAL;
563 }
564
565 goto out;
566 } else if (sopt->sopt_level != IPPROTO_IPV6)
567 return ip6_ctloutput(op, so, sopt);
568
569 switch (sopt->sopt_name) {
570 case MRT6_INIT:
571 case MRT6_DONE:
572 case MRT6_ADD_MIF:
573 case MRT6_DEL_MIF:
574 case MRT6_ADD_MFC:
575 case MRT6_DEL_MFC:
576 case MRT6_PIM:
577 if (op == PRCO_SETOPT)
578 error = ip6_mrouter_set(so, sopt);
579 else if (op == PRCO_GETOPT)
580 error = ip6_mrouter_get(so, sopt);
581 else
582 error = EINVAL;
583 break;
584 case IPV6_CHECKSUM:
585 return ip6_raw_ctloutput(op, so, sopt);
586 default:
587 return ip6_ctloutput(op, so, sopt);
588 }
589 out:
590 return error;
591 }
592
593 extern u_long rip6_sendspace;
594 extern u_long rip6_recvspace;
595
596 int
597 rip6_attach(struct socket *so, int proto)
598 {
599 struct in6pcb *in6p;
600 int s, error;
601
602 KASSERT(sotoin6pcb(so) == NULL);
603 sosetlock(so);
604
605 error = kauth_authorize_network(curlwp->l_cred,
606 KAUTH_NETWORK_SOCKET, KAUTH_REQ_NETWORK_SOCKET_RAWSOCK,
607 KAUTH_ARG(AF_INET6),
608 KAUTH_ARG(SOCK_RAW),
609 KAUTH_ARG(so->so_proto->pr_protocol));
610 if (error) {
611 return error;
612 }
613 s = splsoftnet();
614 error = soreserve(so, rip6_sendspace, rip6_recvspace);
615 if (error) {
616 splx(s);
617 return error;
618 }
619 if ((error = in6_pcballoc(so, &raw6cbtable)) != 0) {
620 splx(s);
621 return error;
622 }
623 splx(s);
624 in6p = sotoin6pcb(so);
625 in6p->in6p_ip6.ip6_nxt = proto;
626 in6p->in6p_cksum = -1;
627
628 in6p->in6p_icmp6filt = kmem_alloc(sizeof(struct icmp6_filter), KM_SLEEP);
629 if (in6p->in6p_icmp6filt == NULL) {
630 in6_pcbdetach(in6p);
631 return ENOMEM;
632 }
633 ICMP6_FILTER_SETPASSALL(in6p->in6p_icmp6filt);
634 KASSERT(solocked(so));
635 return error;
636 }
637
638 static void
639 rip6_detach(struct socket *so)
640 {
641 struct in6pcb *in6p = sotoin6pcb(so);
642
643 KASSERT(solocked(so));
644 KASSERT(in6p != NULL);
645
646 if (so == ip6_mrouter) {
647 ip6_mrouter_done();
648 }
649 /* xxx: RSVP */
650 if (in6p->in6p_icmp6filt != NULL) {
651 kmem_free(in6p->in6p_icmp6filt, sizeof(struct icmp6_filter));
652 in6p->in6p_icmp6filt = NULL;
653 }
654 in6_pcbdetach(in6p);
655 }
656
657 static int
658 rip6_accept(struct socket *so, struct sockaddr *nam)
659 {
660 KASSERT(solocked(so));
661
662 return EOPNOTSUPP;
663 }
664
665 static int
666 rip6_bind(struct socket *so, struct sockaddr *nam, struct lwp *l)
667 {
668 struct in6pcb *in6p = sotoin6pcb(so);
669 struct sockaddr_in6 *addr = (struct sockaddr_in6 *)nam;
670 struct ifaddr *ifa = NULL;
671 int error = 0;
672 int s;
673
674 KASSERT(solocked(so));
675 KASSERT(in6p != NULL);
676 KASSERT(nam != NULL);
677
678 if (addr->sin6_len != sizeof(*addr))
679 return EINVAL;
680 if (IFNET_READER_EMPTY() || addr->sin6_family != AF_INET6)
681 return EADDRNOTAVAIL;
682
683 if ((error = sa6_embedscope(addr, ip6_use_defzone)) != 0)
684 return error;
685
686 /*
687 * we don't support mapped address here, it would confuse
688 * users so reject it
689 */
690 if (IN6_IS_ADDR_V4MAPPED(&addr->sin6_addr))
691 return EADDRNOTAVAIL;
692 s = pserialize_read_enter();
693 if (!IN6_IS_ADDR_UNSPECIFIED(&addr->sin6_addr) &&
694 (ifa = ifa_ifwithaddr(sin6tosa(addr))) == NULL) {
695 error = EADDRNOTAVAIL;
696 goto out;
697 }
698 if (ifa && (ifatoia6(ifa))->ia6_flags &
699 (IN6_IFF_ANYCAST | IN6_IFF_DUPLICATED)) {
700 error = EADDRNOTAVAIL;
701 goto out;
702 }
703
704 in6p->in6p_laddr = addr->sin6_addr;
705 error = 0;
706 out:
707 pserialize_read_exit(s);
708 return error;
709 }
710
711 static int
712 rip6_listen(struct socket *so, struct lwp *l)
713 {
714 KASSERT(solocked(so));
715
716 return EOPNOTSUPP;
717 }
718
719 static int
720 rip6_connect(struct socket *so, struct sockaddr *nam, struct lwp *l)
721 {
722 struct in6pcb *in6p = sotoin6pcb(so);
723 struct sockaddr_in6 *addr = (struct sockaddr_in6 *)nam;
724 struct in6_addr in6a;
725 struct ifnet *ifp = NULL;
726 int scope_ambiguous = 0;
727 int error = 0;
728 struct psref psref;
729 int bound;
730
731 KASSERT(solocked(so));
732 KASSERT(in6p != NULL);
733 KASSERT(nam != NULL);
734
735 if (IFNET_READER_EMPTY())
736 return EADDRNOTAVAIL;
737 if (addr->sin6_family != AF_INET6)
738 return EAFNOSUPPORT;
739
740 /*
741 * Application should provide a proper zone ID or the use of
742 * default zone IDs should be enabled. Unfortunately, some
743 * applications do not behave as it should, so we need a
744 * workaround. Even if an appropriate ID is not determined,
745 * we'll see if we can determine the outgoing interface. If we
746 * can, determine the zone ID based on the interface below.
747 */
748 if (addr->sin6_scope_id == 0 && !ip6_use_defzone)
749 scope_ambiguous = 1;
750 if ((error = sa6_embedscope(addr, ip6_use_defzone)) != 0)
751 return error;
752
753 bound = curlwp_bind();
754 /* Source address selection. XXX: need pcblookup? */
755 error = in6_selectsrc(addr, in6p->in6p_outputopts,
756 in6p->in6p_moptions, &in6p->in6p_route,
757 &in6p->in6p_laddr, &ifp, &psref, &in6a);
758 if (error != 0)
759 goto out;
760 /* XXX: see above */
761 if (ifp && scope_ambiguous &&
762 (error = in6_setscope(&addr->sin6_addr, ifp, NULL)) != 0) {
763 goto out;
764 }
765 in6p->in6p_laddr = in6a;
766 in6p->in6p_faddr = addr->sin6_addr;
767 soisconnected(so);
768 out:
769 if_put(ifp, &psref);
770 curlwp_bindx(bound);
771 return error;
772 }
773
774 static int
775 rip6_connect2(struct socket *so, struct socket *so2)
776 {
777 KASSERT(solocked(so));
778
779 return EOPNOTSUPP;
780 }
781
782 static int
783 rip6_disconnect(struct socket *so)
784 {
785 struct in6pcb *in6p = sotoin6pcb(so);
786
787 KASSERT(solocked(so));
788 KASSERT(in6p != NULL);
789
790 if ((so->so_state & SS_ISCONNECTED) == 0)
791 return ENOTCONN;
792
793 in6p->in6p_faddr = in6addr_any;
794 so->so_state &= ~SS_ISCONNECTED; /* XXX */
795 return 0;
796 }
797
798 static int
799 rip6_shutdown(struct socket *so)
800 {
801 KASSERT(solocked(so));
802
803 /*
804 * Mark the connection as being incapable of futther input.
805 */
806 socantsendmore(so);
807 return 0;
808 }
809
810 static int
811 rip6_abort(struct socket *so)
812 {
813 KASSERT(solocked(so));
814
815 soisdisconnected(so);
816 rip6_detach(so);
817 return 0;
818 }
819
820 static int
821 rip6_ioctl(struct socket *so, u_long cmd, void *nam, struct ifnet *ifp)
822 {
823 return in6_control(so, cmd, nam, ifp);
824 }
825
826 static int
827 rip6_stat(struct socket *so, struct stat *ub)
828 {
829 KASSERT(solocked(so));
830
831 /* stat: don't bother with a blocksize */
832 return 0;
833 }
834
835 static int
836 rip6_peeraddr(struct socket *so, struct sockaddr *nam)
837 {
838 KASSERT(solocked(so));
839 KASSERT(sotoin6pcb(so) != NULL);
840 KASSERT(nam != NULL);
841
842 in6_setpeeraddr(sotoin6pcb(so), (struct sockaddr_in6 *)nam);
843 return 0;
844 }
845
846 static int
847 rip6_sockaddr(struct socket *so, struct sockaddr *nam)
848 {
849 KASSERT(solocked(so));
850 KASSERT(sotoin6pcb(so) != NULL);
851 KASSERT(nam != NULL);
852
853 in6_setsockaddr(sotoin6pcb(so), (struct sockaddr_in6 *)nam);
854 return 0;
855 }
856
857 static int
858 rip6_rcvd(struct socket *so, int flags, struct lwp *l)
859 {
860 KASSERT(solocked(so));
861
862 return EOPNOTSUPP;
863 }
864
865 static int
866 rip6_recvoob(struct socket *so, struct mbuf *m, int flags)
867 {
868 KASSERT(solocked(so));
869
870 return EOPNOTSUPP;
871 }
872
873 static int
874 rip6_send(struct socket *so, struct mbuf *m, struct sockaddr *nam,
875 struct mbuf *control, struct lwp *l)
876 {
877 struct in6pcb *in6p = sotoin6pcb(so);
878 struct sockaddr_in6 tmp;
879 struct sockaddr_in6 *dst;
880 int error = 0;
881
882 KASSERT(solocked(so));
883 KASSERT(in6p != NULL);
884 KASSERT(m != NULL);
885
886 /*
887 * Ship a packet out. The appropriate raw output
888 * routine handles any messaging necessary.
889 */
890
891 /* always copy sockaddr to avoid overwrites */
892 if (so->so_state & SS_ISCONNECTED) {
893 if (nam) {
894 error = EISCONN;
895 goto release;
896 }
897 /* XXX */
898 sockaddr_in6_init(&tmp, &in6p->in6p_faddr, 0, 0, 0);
899 dst = &tmp;
900 } else {
901 if (nam == NULL) {
902 error = ENOTCONN;
903 goto release;
904 }
905 tmp = *(struct sockaddr_in6 *)nam;
906 dst = &tmp;
907
908 if (dst->sin6_family != AF_INET6) {
909 error = EAFNOSUPPORT;
910 goto release;
911 }
912 }
913 error = rip6_output(m, so, dst, control);
914 m = NULL;
915
916 release:
917 if (m)
918 m_freem(m);
919
920 return error;
921 }
922
923 static int
924 rip6_sendoob(struct socket *so, struct mbuf *m, struct mbuf *control)
925 {
926 KASSERT(solocked(so));
927
928 if (m)
929 m_freem(m);
930
931 return EOPNOTSUPP;
932 }
933
934 static int
935 rip6_purgeif(struct socket *so, struct ifnet *ifp)
936 {
937
938 mutex_enter(softnet_lock);
939 in6_pcbpurgeif0(&raw6cbtable, ifp);
940 #ifdef NET_MPSAFE
941 mutex_exit(softnet_lock);
942 #endif
943 in6_purgeif(ifp);
944 #ifdef NET_MPSAFE
945 mutex_enter(softnet_lock);
946 #endif
947 in6_pcbpurgeif(&raw6cbtable, ifp);
948 mutex_exit(softnet_lock);
949
950 return 0;
951 }
952
953 static int
954 sysctl_net_inet6_raw6_stats(SYSCTLFN_ARGS)
955 {
956
957 return (NETSTAT_SYSCTL(rip6stat_percpu, RIP6_NSTATS));
958 }
959
960 static void
961 sysctl_net_inet6_raw6_setup(struct sysctllog **clog)
962 {
963
964 sysctl_createv(clog, 0, NULL, NULL,
965 CTLFLAG_PERMANENT,
966 CTLTYPE_NODE, "inet6", NULL,
967 NULL, 0, NULL, 0,
968 CTL_NET, PF_INET6, CTL_EOL);
969 sysctl_createv(clog, 0, NULL, NULL,
970 CTLFLAG_PERMANENT,
971 CTLTYPE_NODE, "raw6",
972 SYSCTL_DESCR("Raw IPv6 settings"),
973 NULL, 0, NULL, 0,
974 CTL_NET, PF_INET6, IPPROTO_RAW, CTL_EOL);
975
976 sysctl_createv(clog, 0, NULL, NULL,
977 CTLFLAG_PERMANENT,
978 CTLTYPE_STRUCT, "pcblist",
979 SYSCTL_DESCR("Raw IPv6 control block list"),
980 sysctl_inpcblist, 0, &raw6cbtable, 0,
981 CTL_NET, PF_INET6, IPPROTO_RAW,
982 CTL_CREATE, CTL_EOL);
983 sysctl_createv(clog, 0, NULL, NULL,
984 CTLFLAG_PERMANENT,
985 CTLTYPE_STRUCT, "stats",
986 SYSCTL_DESCR("Raw IPv6 statistics"),
987 sysctl_net_inet6_raw6_stats, 0, NULL, 0,
988 CTL_NET, PF_INET6, IPPROTO_RAW, RAW6CTL_STATS,
989 CTL_EOL);
990 }
991
992 PR_WRAP_USRREQS(rip6)
993 #define rip6_attach rip6_attach_wrapper
994 #define rip6_detach rip6_detach_wrapper
995 #define rip6_accept rip6_accept_wrapper
996 #define rip6_bind rip6_bind_wrapper
997 #define rip6_listen rip6_listen_wrapper
998 #define rip6_connect rip6_connect_wrapper
999 #define rip6_connect2 rip6_connect2_wrapper
1000 #define rip6_disconnect rip6_disconnect_wrapper
1001 #define rip6_shutdown rip6_shutdown_wrapper
1002 #define rip6_abort rip6_abort_wrapper
1003 #define rip6_ioctl rip6_ioctl_wrapper
1004 #define rip6_stat rip6_stat_wrapper
1005 #define rip6_peeraddr rip6_peeraddr_wrapper
1006 #define rip6_sockaddr rip6_sockaddr_wrapper
1007 #define rip6_rcvd rip6_rcvd_wrapper
1008 #define rip6_recvoob rip6_recvoob_wrapper
1009 #define rip6_send rip6_send_wrapper
1010 #define rip6_sendoob rip6_sendoob_wrapper
1011 #define rip6_purgeif rip6_purgeif_wrapper
1012
1013 const struct pr_usrreqs rip6_usrreqs = {
1014 .pr_attach = rip6_attach,
1015 .pr_detach = rip6_detach,
1016 .pr_accept = rip6_accept,
1017 .pr_bind = rip6_bind,
1018 .pr_listen = rip6_listen,
1019 .pr_connect = rip6_connect,
1020 .pr_connect2 = rip6_connect2,
1021 .pr_disconnect = rip6_disconnect,
1022 .pr_shutdown = rip6_shutdown,
1023 .pr_abort = rip6_abort,
1024 .pr_ioctl = rip6_ioctl,
1025 .pr_stat = rip6_stat,
1026 .pr_peeraddr = rip6_peeraddr,
1027 .pr_sockaddr = rip6_sockaddr,
1028 .pr_rcvd = rip6_rcvd,
1029 .pr_recvoob = rip6_recvoob,
1030 .pr_send = rip6_send,
1031 .pr_sendoob = rip6_sendoob,
1032 .pr_purgeif = rip6_purgeif,
1033 };
1034