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 * $KAME: in6_pcb.c,v 1.31 2001/05/21 05:45:10 jinmei Exp $
30 */
31
32 /*-
33 * Copyright (c) 1982, 1986, 1991, 1993
34 * The Regents of the University of California. All rights reserved.
35 *
36 * Redistribution and use in source and binary forms, with or without
37 * modification, are permitted provided that the following conditions
38 * are met:
39 * 1. Redistributions of source code must retain the above copyright
40 * notice, this list of conditions and the following disclaimer.
41 * 2. Redistributions in binary form must reproduce the above copyright
42 * notice, this list of conditions and the following disclaimer in the
43 * documentation and/or other materials provided with the distribution.
44 * 4. Neither the name of the University nor the names of its contributors
45 * may be used to endorse or promote products derived from this software
46 * without specific prior written permission.
47 *
48 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
49 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
50 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
51 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
52 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
53 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
54 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
55 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
56 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
57 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
58 * SUCH DAMAGE.
59 *
60 * @(#)in_pcb.c 8.2 (Berkeley) 1/4/94
61 */
62
63 #include <sys/cdefs.h>
64 __FBSDID("$FreeBSD$");
65
66 #include "opt_inet.h"
67 #include "opt_inet6.h"
68 #include "opt_ipsec.h"
69 #include "opt_mac.h"
70
71 #include <sys/param.h>
72 #include <sys/systm.h>
73 #include <sys/malloc.h>
74 #include <sys/mbuf.h>
75 #include <sys/domain.h>
76 #include <sys/protosw.h>
77 #include <sys/socket.h>
78 #include <sys/socketvar.h>
79 #include <sys/sockio.h>
80 #include <sys/errno.h>
81 #include <sys/time.h>
82 #include <sys/priv.h>
83 #include <sys/proc.h>
84 #include <sys/jail.h>
85
86 #include <vm/uma.h>
87
88 #include <net/if.h>
89 #include <net/if_types.h>
90 #include <net/route.h>
91
92 #include <netinet/in.h>
93 #include <netinet/in_var.h>
94 #include <netinet/in_systm.h>
95 #include <netinet/tcp_var.h>
96 #include <netinet/ip6.h>
97 #include <netinet/ip_var.h>
98 #include <netinet6/ip6_var.h>
99 #include <netinet6/nd6.h>
100 #include <netinet/in_pcb.h>
101 #include <netinet6/in6_pcb.h>
102 #include <netinet6/scope6_var.h>
103
104 #include <security/mac/mac_framework.h>
105
106 struct in6_addr zeroin6_addr;
107
108 int
109 in6_pcbbind(register struct inpcb *inp, struct sockaddr *nam,
110 struct ucred *cred)
111 {
112 struct socket *so = inp->inp_socket;
113 struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)NULL;
114 struct inpcbinfo *pcbinfo = inp->inp_pcbinfo;
115 u_short lport = 0;
116 int error, wild = 0, reuseport = (so->so_options & SO_REUSEPORT);
117
118 INP_INFO_WLOCK_ASSERT(pcbinfo);
119 INP_WLOCK_ASSERT(inp);
120
121 if (!in6_ifaddr) /* XXX broken! */
122 return (EADDRNOTAVAIL);
123 if (inp->inp_lport || !IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr))
124 return (EINVAL);
125 if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0)
126 wild = INPLOOKUP_WILDCARD;
127 if (nam == NULL) {
128 if ((error = prison_local_ip6(cred, &inp->in6p_laddr,
129 ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0))) != 0)
130 return (error);
131 } else {
132 sin6 = (struct sockaddr_in6 *)nam;
133 if (nam->sa_len != sizeof(*sin6))
134 return (EINVAL);
135 /*
136 * family check.
137 */
138 if (nam->sa_family != AF_INET6)
139 return (EAFNOSUPPORT);
140
141 if ((error = sa6_embedscope(sin6, ip6_use_defzone)) != 0)
142 return(error);
143
144 if ((error = prison_local_ip6(cred, &sin6->sin6_addr,
145 ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0))) != 0)
146 return (error);
147
148 lport = sin6->sin6_port;
149 if (IN6_IS_ADDR_MULTICAST(&sin6->sin6_addr)) {
150 /*
151 * Treat SO_REUSEADDR as SO_REUSEPORT for multicast;
152 * allow compepte duplication of binding if
153 * SO_REUSEPORT is set, or if SO_REUSEADDR is set
154 * and a multicast address is bound on both
155 * new and duplicated sockets.
156 */
157 if (so->so_options & SO_REUSEADDR)
158 reuseport = SO_REUSEADDR|SO_REUSEPORT;
159 } else if (!IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr)) {
160 struct ifaddr *ia = NULL;
161
162 sin6->sin6_port = 0; /* yech... */
163 if ((ia = ifa_ifwithaddr((struct sockaddr *)sin6)) == 0)
164 return (EADDRNOTAVAIL);
165
166 /*
167 * XXX: bind to an anycast address might accidentally
168 * cause sending a packet with anycast source address.
169 * We should allow to bind to a deprecated address, since
170 * the application dares to use it.
171 */
172 if (ia &&
173 ((struct in6_ifaddr *)ia)->ia6_flags &
174 (IN6_IFF_ANYCAST|IN6_IFF_NOTREADY|IN6_IFF_DETACHED)) {
175 return (EADDRNOTAVAIL);
176 }
177 }
178 if (lport) {
179 struct inpcb *t;
180
181 /* GROSS */
182 if (ntohs(lport) <= ipport_reservedhigh &&
183 ntohs(lport) >= ipport_reservedlow &&
184 priv_check_cred(cred, PRIV_NETINET_RESERVEDPORT,
185 0))
186 return (EACCES);
187 if (!IN6_IS_ADDR_MULTICAST(&sin6->sin6_addr) &&
188 priv_check_cred(inp->inp_cred,
189 PRIV_NETINET_REUSEPORT, 0) != 0) {
190 t = in6_pcblookup_local(pcbinfo,
191 &sin6->sin6_addr, lport,
192 INPLOOKUP_WILDCARD, cred);
193 if (t &&
194 ((t->inp_flags & INP_TIMEWAIT) == 0) &&
195 (so->so_type != SOCK_STREAM ||
196 IN6_IS_ADDR_UNSPECIFIED(&t->in6p_faddr)) &&
197 (!IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr) ||
198 !IN6_IS_ADDR_UNSPECIFIED(&t->in6p_laddr) ||
199 (t->inp_socket->so_options & SO_REUSEPORT)
200 == 0) && (inp->inp_cred->cr_uid !=
201 t->inp_cred->cr_uid))
202 return (EADDRINUSE);
203 if ((inp->inp_flags & IN6P_IPV6_V6ONLY) == 0 &&
204 IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr)) {
205 struct sockaddr_in sin;
206
207 in6_sin6_2_sin(&sin, sin6);
208 t = in_pcblookup_local(pcbinfo,
209 sin.sin_addr, lport,
210 INPLOOKUP_WILDCARD, cred);
211 if (t &&
212 ((t->inp_flags &
213 INP_TIMEWAIT) == 0) &&
214 (so->so_type != SOCK_STREAM ||
215 ntohl(t->inp_faddr.s_addr) ==
216 INADDR_ANY) &&
217 (inp->inp_cred->cr_uid !=
218 t->inp_cred->cr_uid))
219 return (EADDRINUSE);
220 }
221 }
222 t = in6_pcblookup_local(pcbinfo, &sin6->sin6_addr,
223 lport, wild, cred);
224 if (t && (reuseport & ((t->inp_flags & INP_TIMEWAIT) ?
225 intotw(t)->tw_so_options :
226 t->inp_socket->so_options)) == 0)
227 return (EADDRINUSE);
228 if ((inp->inp_flags & IN6P_IPV6_V6ONLY) == 0 &&
229 IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr)) {
230 struct sockaddr_in sin;
231
232 in6_sin6_2_sin(&sin, sin6);
233 t = in_pcblookup_local(pcbinfo, sin.sin_addr,
234 lport, wild, cred);
235 if (t && t->inp_flags & INP_TIMEWAIT) {
236 if ((reuseport &
237 intotw(t)->tw_so_options) == 0 &&
238 (ntohl(t->inp_laddr.s_addr) !=
239 INADDR_ANY || ((inp->inp_vflag &
240 INP_IPV6PROTO) ==
241 (t->inp_vflag & INP_IPV6PROTO))))
242 return (EADDRINUSE);
243 }
244 else if (t &&
245 (reuseport & t->inp_socket->so_options)
246 == 0 && (ntohl(t->inp_laddr.s_addr) !=
247 INADDR_ANY || INP_SOCKAF(so) ==
248 INP_SOCKAF(t->inp_socket)))
249 return (EADDRINUSE);
250 }
251 }
252 inp->in6p_laddr = sin6->sin6_addr;
253 }
254 if (lport == 0) {
255 if ((error = in6_pcbsetport(&inp->in6p_laddr, inp, cred)) != 0) {
256 /* Undo an address bind that may have occurred. */
257 inp->in6p_laddr = in6addr_any;
258 return (error);
259 }
260 } else {
261 inp->inp_lport = lport;
262 if (in_pcbinshash(inp) != 0) {
263 inp->in6p_laddr = in6addr_any;
264 inp->inp_lport = 0;
265 return (EAGAIN);
266 }
267 }
268 return (0);
269 }
270
271 /*
272 * Transform old in6_pcbconnect() into an inner subroutine for new
273 * in6_pcbconnect(): Do some validity-checking on the remote
274 * address (in mbuf 'nam') and then determine local host address
275 * (i.e., which interface) to use to access that remote host.
276 *
277 * This preserves definition of in6_pcbconnect(), while supporting a
278 * slightly different version for T/TCP. (This is more than
279 * a bit of a kludge, but cleaning up the internal interfaces would
280 * have forced minor changes in every protocol).
281 */
282 int
283 in6_pcbladdr(register struct inpcb *inp, struct sockaddr *nam,
284 struct in6_addr *plocal_addr6)
285 {
286 register struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)nam;
287 int error = 0;
288 struct ifnet *ifp = NULL;
289 int scope_ambiguous = 0;
290 struct in6_addr in6a;
291
292 INP_INFO_WLOCK_ASSERT(inp->inp_pcbinfo);
293 INP_WLOCK_ASSERT(inp);
294
295 if (nam->sa_len != sizeof (*sin6))
296 return (EINVAL);
297 if (sin6->sin6_family != AF_INET6)
298 return (EAFNOSUPPORT);
299 if (sin6->sin6_port == 0)
300 return (EADDRNOTAVAIL);
301
302 if (sin6->sin6_scope_id == 0 && !ip6_use_defzone)
303 scope_ambiguous = 1;
304 if ((error = sa6_embedscope(sin6, ip6_use_defzone)) != 0)
305 return(error);
306
307 if (in6_ifaddr) {
308 /*
309 * If the destination address is UNSPECIFIED addr,
310 * use the loopback addr, e.g ::1.
311 */
312 if (IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr))
313 sin6->sin6_addr = in6addr_loopback;
314 }
315 if ((error = prison_remote_ip6(inp->inp_cred, &sin6->sin6_addr)) != 0)
316 return (error);
317
318 error = in6_selectsrc(sin6, inp->in6p_outputopts,
319 inp, NULL, inp->inp_cred, &ifp, &in6a);
320 if (error)
321 return (error);
322
323 if (ifp && scope_ambiguous &&
324 (error = in6_setscope(&sin6->sin6_addr, ifp, NULL)) != 0) {
325 return(error);
326 }
327
328 /*
329 * Do not update this earlier, in case we return with an error.
330 *
331 * XXX: this in6_selectsrc result might replace the bound local
332 * address with the address specified by setsockopt(IPV6_PKTINFO).
333 * Is it the intended behavior?
334 */
335 *plocal_addr6 = in6a;
336
337 /*
338 * Don't do pcblookup call here; return interface in
339 * plocal_addr6
340 * and exit to caller, that will do the lookup.
341 */
342
343 return (0);
344 }
345
346 /*
347 * Outer subroutine:
348 * Connect from a socket to a specified address.
349 * Both address and port must be specified in argument sin.
350 * If don't have a local address for this socket yet,
351 * then pick one.
352 */
353 int
354 in6_pcbconnect(register struct inpcb *inp, struct sockaddr *nam,
355 struct ucred *cred)
356 {
357 register struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)nam;
358 struct in6_addr addr6;
359 int error;
360
361 INP_INFO_WLOCK_ASSERT(inp->inp_pcbinfo);
362 INP_WLOCK_ASSERT(inp);
363
364 /*
365 * Call inner routine, to assign local interface address.
366 * in6_pcbladdr() may automatically fill in sin6_scope_id.
367 */
368 if ((error = in6_pcbladdr(inp, nam, &addr6)) != 0)
369 return (error);
370
371 if (in6_pcblookup_hash(inp->inp_pcbinfo, &sin6->sin6_addr,
372 sin6->sin6_port,
373 IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr)
374 ? &addr6 : &inp->in6p_laddr,
375 inp->inp_lport, 0, NULL) != NULL) {
376 return (EADDRINUSE);
377 }
378 if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr)) {
379 if (inp->inp_lport == 0) {
380 error = in6_pcbbind(inp, (struct sockaddr *)0, cred);
381 if (error)
382 return (error);
383 }
384 inp->in6p_laddr = addr6;
385 }
386 inp->in6p_faddr = sin6->sin6_addr;
387 inp->inp_fport = sin6->sin6_port;
388 /* update flowinfo - draft-itojun-ipv6-flowlabel-api-00 */
389 inp->inp_flow &= ~IPV6_FLOWLABEL_MASK;
390 if (inp->inp_flags & IN6P_AUTOFLOWLABEL)
391 inp->inp_flow |=
392 (htonl(ip6_randomflowlabel()) & IPV6_FLOWLABEL_MASK);
393
394 in_pcbrehash(inp);
395
396 return (0);
397 }
398
399 void
400 in6_pcbdisconnect(struct inpcb *inp)
401 {
402
403 INP_INFO_WLOCK_ASSERT(inp->inp_pcbinfo);
404 INP_WLOCK_ASSERT(inp);
405
406 bzero((caddr_t)&inp->in6p_faddr, sizeof(inp->in6p_faddr));
407 inp->inp_fport = 0;
408 /* clear flowinfo - draft-itojun-ipv6-flowlabel-api-00 */
409 inp->inp_flow &= ~IPV6_FLOWLABEL_MASK;
410 in_pcbrehash(inp);
411 }
412
413 struct sockaddr *
414 in6_sockaddr(in_port_t port, struct in6_addr *addr_p)
415 {
416 struct sockaddr_in6 *sin6;
417
418 MALLOC(sin6, struct sockaddr_in6 *, sizeof *sin6, M_SONAME, M_WAITOK);
419 bzero(sin6, sizeof *sin6);
420 sin6->sin6_family = AF_INET6;
421 sin6->sin6_len = sizeof(*sin6);
422 sin6->sin6_port = port;
423 sin6->sin6_addr = *addr_p;
424 (void)sa6_recoverscope(sin6); /* XXX: should catch errors */
425
426 return (struct sockaddr *)sin6;
427 }
428
429 struct sockaddr *
430 in6_v4mapsin6_sockaddr(in_port_t port, struct in_addr *addr_p)
431 {
432 struct sockaddr_in sin;
433 struct sockaddr_in6 *sin6_p;
434
435 bzero(&sin, sizeof sin);
436 sin.sin_family = AF_INET;
437 sin.sin_len = sizeof(sin);
438 sin.sin_port = port;
439 sin.sin_addr = *addr_p;
440
441 MALLOC(sin6_p, struct sockaddr_in6 *, sizeof *sin6_p, M_SONAME,
442 M_WAITOK);
443 in6_sin_2_v4mapsin6(&sin, sin6_p);
444
445 return (struct sockaddr *)sin6_p;
446 }
447
448 int
449 in6_getsockaddr(struct socket *so, struct sockaddr **nam)
450 {
451 register struct inpcb *inp;
452 struct in6_addr addr;
453 in_port_t port;
454
455 inp = sotoinpcb(so);
456 KASSERT(inp != NULL, ("in6_getsockaddr: inp == NULL"));
457
458 INP_RLOCK(inp);
459 port = inp->inp_lport;
460 addr = inp->in6p_laddr;
461 INP_RUNLOCK(inp);
462
463 *nam = in6_sockaddr(port, &addr);
464 return 0;
465 }
466
467 int
468 in6_getpeeraddr(struct socket *so, struct sockaddr **nam)
469 {
470 struct inpcb *inp;
471 struct in6_addr addr;
472 in_port_t port;
473
474 inp = sotoinpcb(so);
475 KASSERT(inp != NULL, ("in6_getpeeraddr: inp == NULL"));
476
477 INP_RLOCK(inp);
478 port = inp->inp_fport;
479 addr = inp->in6p_faddr;
480 INP_RUNLOCK(inp);
481
482 *nam = in6_sockaddr(port, &addr);
483 return 0;
484 }
485
486 int
487 in6_mapped_sockaddr(struct socket *so, struct sockaddr **nam)
488 {
489 struct inpcb *inp;
490 int error;
491
492 inp = sotoinpcb(so);
493 KASSERT(inp != NULL, ("in6_mapped_sockaddr: inp == NULL"));
494
495 if ((inp->inp_vflag & (INP_IPV4 | INP_IPV6)) == INP_IPV4) {
496 error = in_getsockaddr(so, nam);
497 if (error == 0)
498 in6_sin_2_v4mapsin6_in_sock(nam);
499 } else {
500 /* scope issues will be handled in in6_getsockaddr(). */
501 error = in6_getsockaddr(so, nam);
502 }
503
504 return error;
505 }
506
507 int
508 in6_mapped_peeraddr(struct socket *so, struct sockaddr **nam)
509 {
510 struct inpcb *inp;
511 int error;
512
513 inp = sotoinpcb(so);
514 KASSERT(inp != NULL, ("in6_mapped_peeraddr: inp == NULL"));
515
516 if ((inp->inp_vflag & (INP_IPV4 | INP_IPV6)) == INP_IPV4) {
517 error = in_getpeeraddr(so, nam);
518 if (error == 0)
519 in6_sin_2_v4mapsin6_in_sock(nam);
520 } else
521 /* scope issues will be handled in in6_getpeeraddr(). */
522 error = in6_getpeeraddr(so, nam);
523
524 return error;
525 }
526
527 /*
528 * Pass some notification to all connections of a protocol
529 * associated with address dst. The local address and/or port numbers
530 * may be specified to limit the search. The "usual action" will be
531 * taken, depending on the ctlinput cmd. The caller must filter any
532 * cmds that are uninteresting (e.g., no error in the map).
533 * Call the protocol specific routine (if any) to report
534 * any errors for each matching socket.
535 */
536 void
537 in6_pcbnotify(struct inpcbinfo *pcbinfo, struct sockaddr *dst,
538 u_int fport_arg, const struct sockaddr *src, u_int lport_arg,
539 int cmd, void *cmdarg,
540 struct inpcb *(*notify)(struct inpcb *, int))
541 {
542 struct inpcb *inp, *inp_temp;
543 struct sockaddr_in6 sa6_src, *sa6_dst;
544 u_short fport = fport_arg, lport = lport_arg;
545 u_int32_t flowinfo;
546 int errno;
547
548 if ((unsigned)cmd >= PRC_NCMDS || dst->sa_family != AF_INET6)
549 return;
550
551 sa6_dst = (struct sockaddr_in6 *)dst;
552 if (IN6_IS_ADDR_UNSPECIFIED(&sa6_dst->sin6_addr))
553 return;
554
555 /*
556 * note that src can be NULL when we get notify by local fragmentation.
557 */
558 sa6_src = (src == NULL) ? sa6_any : *(const struct sockaddr_in6 *)src;
559 flowinfo = sa6_src.sin6_flowinfo;
560
561 /*
562 * Redirects go to all references to the destination,
563 * and use in6_rtchange to invalidate the route cache.
564 * Dead host indications: also use in6_rtchange to invalidate
565 * the cache, and deliver the error to all the sockets.
566 * Otherwise, if we have knowledge of the local port and address,
567 * deliver only to that socket.
568 */
569 if (PRC_IS_REDIRECT(cmd) || cmd == PRC_HOSTDEAD) {
570 fport = 0;
571 lport = 0;
572 bzero((caddr_t)&sa6_src.sin6_addr, sizeof(sa6_src.sin6_addr));
573
574 if (cmd != PRC_HOSTDEAD)
575 notify = in6_rtchange;
576 }
577 errno = inet6ctlerrmap[cmd];
578 INP_INFO_WLOCK(pcbinfo);
579 LIST_FOREACH_SAFE(inp, pcbinfo->ipi_listhead, inp_list, inp_temp) {
580 INP_WLOCK(inp);
581 if ((inp->inp_vflag & INP_IPV6) == 0) {
582 INP_WUNLOCK(inp);
583 continue;
584 }
585
586 /*
587 * If the error designates a new path MTU for a destination
588 * and the application (associated with this socket) wanted to
589 * know the value, notify. Note that we notify for all
590 * disconnected sockets if the corresponding application
591 * wanted. This is because some UDP applications keep sending
592 * sockets disconnected.
593 * XXX: should we avoid to notify the value to TCP sockets?
594 */
595 if (cmd == PRC_MSGSIZE && (inp->inp_flags & IN6P_MTU) != 0 &&
596 (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr) ||
597 IN6_ARE_ADDR_EQUAL(&inp->in6p_faddr, &sa6_dst->sin6_addr))) {
598 ip6_notify_pmtu(inp, (struct sockaddr_in6 *)dst,
599 (u_int32_t *)cmdarg);
600 }
601
602 /*
603 * Detect if we should notify the error. If no source and
604 * destination ports are specifed, but non-zero flowinfo and
605 * local address match, notify the error. This is the case
606 * when the error is delivered with an encrypted buffer
607 * by ESP. Otherwise, just compare addresses and ports
608 * as usual.
609 */
610 if (lport == 0 && fport == 0 && flowinfo &&
611 inp->inp_socket != NULL &&
612 flowinfo == (inp->inp_flow & IPV6_FLOWLABEL_MASK) &&
613 IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr, &sa6_src.sin6_addr))
614 goto do_notify;
615 else if (!IN6_ARE_ADDR_EQUAL(&inp->in6p_faddr,
616 &sa6_dst->sin6_addr) ||
617 inp->inp_socket == 0 ||
618 (lport && inp->inp_lport != lport) ||
619 (!IN6_IS_ADDR_UNSPECIFIED(&sa6_src.sin6_addr) &&
620 !IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr,
621 &sa6_src.sin6_addr)) ||
622 (fport && inp->inp_fport != fport)) {
623 INP_WUNLOCK(inp);
624 continue;
625 }
626
627 do_notify:
628 if (notify) {
629 if ((*notify)(inp, errno))
630 INP_WUNLOCK(inp);
631 } else
632 INP_WUNLOCK(inp);
633 }
634 INP_INFO_WUNLOCK(pcbinfo);
635 }
636
637 /*
638 * Lookup a PCB based on the local address and port.
639 */
640 struct inpcb *
641 in6_pcblookup_local(struct inpcbinfo *pcbinfo, struct in6_addr *laddr,
642 u_short lport, int wild_okay, struct ucred *cred)
643 {
644 register struct inpcb *inp;
645 int matchwild = 3, wildcard;
646
647 INP_INFO_WLOCK_ASSERT(pcbinfo);
648
649 if (!wild_okay) {
650 struct inpcbhead *head;
651 /*
652 * Look for an unconnected (wildcard foreign addr) PCB that
653 * matches the local address and port we're looking for.
654 */
655 head = &pcbinfo->ipi_hashbase[INP_PCBHASH(INADDR_ANY, lport,
656 0, pcbinfo->ipi_hashmask)];
657 LIST_FOREACH(inp, head, inp_hash) {
658 /* XXX inp locking */
659 if ((inp->inp_vflag & INP_IPV6) == 0)
660 continue;
661 if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr) &&
662 IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr, laddr) &&
663 inp->inp_lport == lport) {
664 /* Found. */
665 if (cred == NULL ||
666 inp->inp_cred->cr_prison == cred->cr_prison)
667 return (inp);
668 }
669 }
670 /*
671 * Not found.
672 */
673 return (NULL);
674 } else {
675 struct inpcbporthead *porthash;
676 struct inpcbport *phd;
677 struct inpcb *match = NULL;
678 /*
679 * Best fit PCB lookup.
680 *
681 * First see if this local port is in use by looking on the
682 * port hash list.
683 */
684 porthash = &pcbinfo->ipi_porthashbase[INP_PCBPORTHASH(lport,
685 pcbinfo->ipi_porthashmask)];
686 LIST_FOREACH(phd, porthash, phd_hash) {
687 if (phd->phd_port == lport)
688 break;
689 }
690 if (phd != NULL) {
691 /*
692 * Port is in use by one or more PCBs. Look for best
693 * fit.
694 */
695 LIST_FOREACH(inp, &phd->phd_pcblist, inp_portlist) {
696 wildcard = 0;
697 if (cred != NULL &&
698 inp->inp_cred->cr_prison != cred->cr_prison)
699 continue;
700 /* XXX inp locking */
701 if ((inp->inp_vflag & INP_IPV6) == 0)
702 continue;
703 if (!IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr))
704 wildcard++;
705 if (!IN6_IS_ADDR_UNSPECIFIED(
706 &inp->in6p_laddr)) {
707 if (IN6_IS_ADDR_UNSPECIFIED(laddr))
708 wildcard++;
709 else if (!IN6_ARE_ADDR_EQUAL(
710 &inp->in6p_laddr, laddr))
711 continue;
712 } else {
713 if (!IN6_IS_ADDR_UNSPECIFIED(laddr))
714 wildcard++;
715 }
716 if (wildcard < matchwild) {
717 match = inp;
718 matchwild = wildcard;
719 if (matchwild == 0)
720 break;
721 }
722 }
723 }
724 return (match);
725 }
726 }
727
728 void
729 in6_pcbpurgeif0(struct inpcbinfo *pcbinfo, struct ifnet *ifp)
730 {
731 struct inpcb *in6p;
732 struct ip6_moptions *im6o;
733 struct in6_multi_mship *imm, *nimm;
734
735 INP_INFO_RLOCK(pcbinfo);
736 LIST_FOREACH(in6p, pcbinfo->ipi_listhead, inp_list) {
737 INP_WLOCK(in6p);
738 im6o = in6p->in6p_moptions;
739 if ((in6p->inp_vflag & INP_IPV6) &&
740 im6o) {
741 /*
742 * Unselect the outgoing interface if it is being
743 * detached.
744 */
745 if (im6o->im6o_multicast_ifp == ifp)
746 im6o->im6o_multicast_ifp = NULL;
747
748 /*
749 * Drop multicast group membership if we joined
750 * through the interface being detached.
751 * XXX controversial - is it really legal for kernel
752 * to force this?
753 */
754 for (imm = im6o->im6o_memberships.lh_first;
755 imm != NULL; imm = nimm) {
756 nimm = imm->i6mm_chain.le_next;
757 if (imm->i6mm_maddr->in6m_ifp == ifp) {
758 LIST_REMOVE(imm, i6mm_chain);
759 in6_delmulti(imm->i6mm_maddr);
760 free(imm, M_IP6MADDR);
761 }
762 }
763 }
764 INP_WUNLOCK(in6p);
765 }
766 INP_INFO_RUNLOCK(pcbinfo);
767 }
768
769 /*
770 * Check for alternatives when higher level complains
771 * about service problems. For now, invalidate cached
772 * routing information. If the route was created dynamically
773 * (by a redirect), time to try a default gateway again.
774 */
775 void
776 in6_losing(struct inpcb *in6p)
777 {
778
779 /*
780 * We don't store route pointers in the routing table anymore
781 */
782 return;
783 }
784
785 /*
786 * After a routing change, flush old routing
787 * and allocate a (hopefully) better one.
788 */
789 struct inpcb *
790 in6_rtchange(struct inpcb *inp, int errno)
791 {
792 /*
793 * We don't store route pointers in the routing table anymore
794 */
795 return inp;
796 }
797
798 /*
799 * Lookup PCB in hash list.
800 */
801 struct inpcb *
802 in6_pcblookup_hash(struct inpcbinfo *pcbinfo, struct in6_addr *faddr,
803 u_int fport_arg, struct in6_addr *laddr, u_int lport_arg, int wildcard,
804 struct ifnet *ifp)
805 {
806 struct inpcbhead *head;
807 struct inpcb *inp, *tmpinp;
808 u_short fport = fport_arg, lport = lport_arg;
809 int faith;
810
811 INP_INFO_LOCK_ASSERT(pcbinfo);
812
813 if (faithprefix_p != NULL)
814 faith = (*faithprefix_p)(laddr);
815 else
816 faith = 0;
817
818 /*
819 * First look for an exact match.
820 */
821 tmpinp = NULL;
822 head = &pcbinfo->ipi_hashbase[
823 INP_PCBHASH(faddr->s6_addr32[3] /* XXX */, lport, fport,
824 pcbinfo->ipi_hashmask)];
825 LIST_FOREACH(inp, head, inp_hash) {
826 /* XXX inp locking */
827 if ((inp->inp_vflag & INP_IPV6) == 0)
828 continue;
829 if (IN6_ARE_ADDR_EQUAL(&inp->in6p_faddr, faddr) &&
830 IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr, laddr) &&
831 inp->inp_fport == fport &&
832 inp->inp_lport == lport) {
833 /*
834 * XXX We should be able to directly return
835 * the inp here, without any checks.
836 * Well unless both bound with SO_REUSEPORT?
837 */
838 if (jailed(inp->inp_cred))
839 return (inp);
840 if (tmpinp == NULL)
841 tmpinp = inp;
842 }
843 }
844 if (tmpinp != NULL)
845 return (tmpinp);
846
847 /*
848 * Then look for a wildcard match, if requested.
849 */
850 if (wildcard == INPLOOKUP_WILDCARD) {
851 struct inpcb *local_wild = NULL, *local_exact = NULL;
852 struct inpcb *jail_wild = NULL;
853 int injail;
854
855 /*
856 * Order of socket selection - we always prefer jails.
857 * 1. jailed, non-wild.
858 * 2. jailed, wild.
859 * 3. non-jailed, non-wild.
860 * 4. non-jailed, wild.
861 */
862 head = &pcbinfo->ipi_hashbase[INP_PCBHASH(INADDR_ANY, lport,
863 0, pcbinfo->ipi_hashmask)];
864 LIST_FOREACH(inp, head, inp_hash) {
865 /* XXX inp locking */
866 if ((inp->inp_vflag & INP_IPV6) == 0)
867 continue;
868
869 if (!IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr) ||
870 inp->inp_lport != lport) {
871 continue;
872 }
873
874 /* XXX inp locking */
875 if (faith && (inp->inp_flags & INP_FAITH) == 0)
876 continue;
877
878 injail = jailed(inp->inp_cred);
879 if (injail) {
880 if (prison_check_ip6(inp->inp_cred,
881 laddr) != 0)
882 continue;
883 } else {
884 if (local_exact != NULL)
885 continue;
886 }
887
888 if (IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr, laddr)) {
889 if (injail)
890 return (inp);
891 else
892 local_exact = inp;
893 } else if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr)) {
894 if (injail)
895 jail_wild = inp;
896 else
897 local_wild = inp;
898 }
899 } /* LIST_FOREACH */
900
901 if (jail_wild != NULL)
902 return (jail_wild);
903 if (local_exact != NULL)
904 return (local_exact);
905 if (local_wild != NULL)
906 return (local_wild);
907 } /* if (wildcard == INPLOOKUP_WILDCARD) */
908
909 /*
910 * Not found.
911 */
912 return (NULL);
913 }
914
915 void
916 init_sin6(struct sockaddr_in6 *sin6, struct mbuf *m)
917 {
918 struct ip6_hdr *ip;
919
920 ip = mtod(m, struct ip6_hdr *);
921 bzero(sin6, sizeof(*sin6));
922 sin6->sin6_len = sizeof(*sin6);
923 sin6->sin6_family = AF_INET6;
924 sin6->sin6_addr = ip->ip6_src;
925
926 (void)sa6_recoverscope(sin6); /* XXX: should catch errors... */
927
928 return;
929 }
Cache object: 53186dfa76e894ae4da46faaf0346886
|