2 * Copyright (c) 2004 Jeffrey M. Hsu. All rights reserved.
3 * Copyright (c) 2004 The DragonFly Project. All rights reserved.
5 * This code is derived from software contributed to The DragonFly 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 DragonFly Project nor the names of its
17 * contributors may be used to endorse or promote products derived
18 * from this software without specific, prior written permission.
20 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
21 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
22 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
23 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
24 * COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
25 * INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING,
26 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
27 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
28 * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
29 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
30 * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
35 * Copyright (c) 2004 Jeffrey M. Hsu. All rights reserved.
37 * License terms: all terms for the DragonFly license above plus the following:
39 * 4. All advertising materials mentioning features or use of this software
40 * must display the following acknowledgement:
42 * This product includes software developed by Jeffrey M. Hsu
43 * for the DragonFly Project.
45 * This requirement may be waived with permission from Jeffrey Hsu.
46 * This requirement will sunset and may be removed on July 8 2005,
47 * after which the standard DragonFly license (as shown above) will
52 * Copyright (c) 1982, 1986, 1991, 1993, 1995
53 * The Regents of the University of California. All rights reserved.
55 * Redistribution and use in source and binary forms, with or without
56 * modification, are permitted provided that the following conditions
58 * 1. Redistributions of source code must retain the above copyright
59 * notice, this list of conditions and the following disclaimer.
60 * 2. Redistributions in binary form must reproduce the above copyright
61 * notice, this list of conditions and the following disclaimer in the
62 * documentation and/or other materials provided with the distribution.
63 * 3. All advertising materials mentioning features or use of this software
64 * must display the following acknowledgement:
65 * This product includes software developed by the University of
66 * California, Berkeley and its contributors.
67 * 4. Neither the name of the University nor the names of its contributors
68 * may be used to endorse or promote products derived from this software
69 * without specific prior written permission.
71 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
72 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
73 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
74 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
75 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
76 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
77 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
78 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
79 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
80 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
83 * @(#)in_pcb.c 8.4 (Berkeley) 5/24/95
84 * $FreeBSD: src/sys/netinet/in_pcb.c,v 1.59.2.27 2004/01/02 04:06:42 ambrisko Exp $
85 * $DragonFly: src/sys/netinet/in_pcb.c,v 1.24 2004/07/08 22:07:35 hsu Exp $
88 #include "opt_ipsec.h"
89 #include "opt_inet6.h"
91 #include <sys/param.h>
92 #include <sys/systm.h>
93 #include <sys/malloc.h>
95 #include <sys/domain.h>
96 #include <sys/protosw.h>
97 #include <sys/socket.h>
98 #include <sys/socketvar.h>
100 #include <sys/jail.h>
101 #include <sys/kernel.h>
102 #include <sys/sysctl.h>
104 #include <machine/limits.h>
106 #include <vm/vm_zone.h>
109 #include <net/if_types.h>
110 #include <net/route.h>
112 #include <netinet/in.h>
113 #include <netinet/in_pcb.h>
114 #include <netinet/in_var.h>
115 #include <netinet/ip_var.h>
117 #include <netinet/ip6.h>
118 #include <netinet6/ip6_var.h>
122 #include <netinet6/ipsec.h>
123 #include <netproto/key/key.h>
127 #if defined(IPSEC) || defined(IPSEC_ESP)
128 #error "Bad idea: don't compile with both IPSEC and FAST_IPSEC!"
131 #include <netipsec/ipsec.h>
132 #include <netipsec/key.h>
134 #endif /* FAST_IPSEC */
136 struct in_addr zeroin_addr;
139 * These configure the range of local port addresses assigned to
140 * "unspecified" outgoing connections/packets/whatever.
142 int ipport_lowfirstauto = IPPORT_RESERVED - 1; /* 1023 */
143 int ipport_lowlastauto = IPPORT_RESERVEDSTART; /* 600 */
145 int ipport_firstauto = IPPORT_RESERVED; /* 1024 */
146 int ipport_lastauto = IPPORT_USERRESERVED; /* 5000 */
148 int ipport_hifirstauto = IPPORT_HIFIRSTAUTO; /* 49152 */
149 int ipport_hilastauto = IPPORT_HILASTAUTO; /* 65535 */
151 /* Allocate ephermal source ports in random order. */
152 int ipport_randomized = 1;
155 RANGECHK(int var, int min, int max)
164 sysctl_net_ipport_check(SYSCTL_HANDLER_ARGS)
168 error = sysctl_handle_int(oidp, oidp->oid_arg1, oidp->oid_arg2, req);
170 RANGECHK(ipport_lowfirstauto, 1, IPPORT_RESERVED - 1);
171 RANGECHK(ipport_lowlastauto, 1, IPPORT_RESERVED - 1);
173 RANGECHK(ipport_firstauto, IPPORT_RESERVED, USHRT_MAX);
174 RANGECHK(ipport_lastauto, IPPORT_RESERVED, USHRT_MAX);
176 RANGECHK(ipport_hifirstauto, IPPORT_RESERVED, USHRT_MAX);
177 RANGECHK(ipport_hilastauto, IPPORT_RESERVED, USHRT_MAX);
182 SYSCTL_NODE(_net_inet_ip, IPPROTO_IP, portrange, CTLFLAG_RW, 0, "IP Ports");
184 SYSCTL_PROC(_net_inet_ip_portrange, OID_AUTO, lowfirst, CTLTYPE_INT|CTLFLAG_RW,
185 &ipport_lowfirstauto, 0, &sysctl_net_ipport_check, "I", "");
186 SYSCTL_PROC(_net_inet_ip_portrange, OID_AUTO, lowlast, CTLTYPE_INT|CTLFLAG_RW,
187 &ipport_lowlastauto, 0, &sysctl_net_ipport_check, "I", "");
188 SYSCTL_PROC(_net_inet_ip_portrange, OID_AUTO, first, CTLTYPE_INT|CTLFLAG_RW,
189 &ipport_firstauto, 0, &sysctl_net_ipport_check, "I", "");
190 SYSCTL_PROC(_net_inet_ip_portrange, OID_AUTO, last, CTLTYPE_INT|CTLFLAG_RW,
191 &ipport_lastauto, 0, &sysctl_net_ipport_check, "I", "");
192 SYSCTL_PROC(_net_inet_ip_portrange, OID_AUTO, hifirst, CTLTYPE_INT|CTLFLAG_RW,
193 &ipport_hifirstauto, 0, &sysctl_net_ipport_check, "I", "");
194 SYSCTL_PROC(_net_inet_ip_portrange, OID_AUTO, hilast, CTLTYPE_INT|CTLFLAG_RW,
195 &ipport_hilastauto, 0, &sysctl_net_ipport_check, "I", "");
196 SYSCTL_INT(_net_inet_ip_portrange, OID_AUTO, randomized, CTLFLAG_RW,
197 &ipport_randomized, 0, "");
200 * in_pcb.c: manage the Protocol Control Blocks.
202 * NOTE: It is assumed that most of these functions will be called at
203 * splnet(). XXX - There are, unfortunately, a few exceptions to this
204 * rule that should be fixed.
208 in_pcbinfo_init(struct inpcbinfo *pcbinfo)
210 LIST_INIT(&pcbinfo->pcblisthead);
214 * Allocate a PCB and associate it with the socket.
217 in_pcballoc(struct socket *so, struct inpcbinfo *pcbinfo)
224 inp = zalloc(pcbinfo->ipi_zone);
227 bzero((caddr_t)inp, sizeof *inp);
228 inp->inp_gencnt = ++pcbinfo->ipi_gencnt;
229 inp->inp_pcbinfo = inp->inp_cpcbinfo = pcbinfo;
230 inp->inp_socket = so;
232 error = ipsec_init_policy(so, &inp->inp_sp);
234 zfree(pcbinfo->ipi_zone, inp);
239 if (INP_SOCKAF(so) == AF_INET6 && ip6_v6only)
240 inp->inp_flags |= IN6P_IPV6_V6ONLY;
241 if (ip6_auto_flowlabel)
242 inp->inp_flags |= IN6P_AUTOFLOWLABEL;
244 so->so_pcb = (caddr_t)inp;
245 LIST_INSERT_HEAD(&pcbinfo->pcblisthead, inp, inp_list);
246 pcbinfo->ipi_count++;
251 in_pcbbind(struct inpcb *inp, struct sockaddr *nam, struct thread *td)
253 struct socket *so = inp->inp_socket;
254 struct proc *p = td->td_proc;
255 unsigned short *lastport;
256 struct sockaddr_in *sin;
257 struct inpcbinfo *pcbinfo = inp->inp_pcbinfo;
259 int wild = 0, reuseport = (so->so_options & SO_REUSEPORT);
260 int error, prison = 0;
264 if (TAILQ_EMPTY(&in_ifaddrhead)) /* XXX broken! */
265 return (EADDRNOTAVAIL);
266 if (inp->inp_lport != 0 || inp->inp_laddr.s_addr != INADDR_ANY)
267 return (EINVAL); /* already bound */
268 if (!(so->so_options & (SO_REUSEADDR|SO_REUSEPORT)))
269 wild = 1; /* neither SO_REUSEADDR nor SO_REUSEPORT is set */
271 sin = (struct sockaddr_in *)nam;
272 if (nam->sa_len != sizeof *sin)
276 * We should check the family, but old programs
277 * incorrectly fail to initialize it.
279 if (sin->sin_family != AF_INET)
280 return (EAFNOSUPPORT);
282 if (sin->sin_addr.s_addr != INADDR_ANY &&
283 prison_ip(td, 0, &sin->sin_addr.s_addr))
285 lport = sin->sin_port;
286 if (IN_MULTICAST(ntohl(sin->sin_addr.s_addr))) {
288 * Treat SO_REUSEADDR as SO_REUSEPORT for multicast;
289 * allow complete duplication of binding if
290 * SO_REUSEPORT is set, or if SO_REUSEADDR is set
291 * and a multicast address is bound on both
292 * new and duplicated sockets.
294 if (so->so_options & SO_REUSEADDR)
295 reuseport = SO_REUSEADDR | SO_REUSEPORT;
296 } else if (sin->sin_addr.s_addr != INADDR_ANY) {
297 sin->sin_port = 0; /* yech... */
298 bzero(&sin->sin_zero, sizeof sin->sin_zero);
299 if (ifa_ifwithaddr((struct sockaddr *)sin) == NULL)
300 return (EADDRNOTAVAIL);
306 if (ntohs(lport) < IPPORT_RESERVED &&
307 p && suser_cred(p->p_ucred, PRISON_ROOT))
309 if (p && p->p_ucred->cr_prison)
311 if (so->so_cred->cr_uid != 0 &&
312 !IN_MULTICAST(ntohl(sin->sin_addr.s_addr))) {
313 t = in_pcblookup_local(inp->inp_pcbinfo,
314 sin->sin_addr, lport,
315 prison ? 0 : INPLOOKUP_WILDCARD);
317 (!in_nullhost(sin->sin_addr) ||
318 !in_nullhost(t->inp_laddr) ||
319 (t->inp_socket->so_options &
320 SO_REUSEPORT) == 0) &&
321 (so->so_cred->cr_uid !=
322 t->inp_socket->so_cred->cr_uid)) {
324 if (!in_nullhost(sin->sin_addr) ||
325 !in_nullhost(t->inp_laddr) ||
327 INP_SOCKAF(t->inp_socket))
332 if (prison && prison_ip(td, 0, &sin->sin_addr.s_addr))
333 return (EADDRNOTAVAIL);
334 t = in_pcblookup_local(pcbinfo, sin->sin_addr,
335 lport, prison ? 0 : wild);
336 if (t && !(reuseport & t->inp_socket->so_options)) {
338 if (!in_nullhost(sin->sin_addr) ||
339 !in_nullhost(t->inp_laddr) ||
340 INP_SOCKAF(so) == INP_SOCKAF(t->inp_socket))
345 inp->inp_laddr = sin->sin_addr;
351 if (inp->inp_laddr.s_addr != INADDR_ANY &&
352 prison_ip(td, 0, &inp->inp_laddr.s_addr )) {
353 inp->inp_laddr.s_addr = INADDR_ANY;
356 inp->inp_flags |= INP_ANONPORT;
358 if (inp->inp_flags & INP_HIGHPORT) {
359 first = ipport_hifirstauto; /* sysctl */
360 last = ipport_hilastauto;
361 lastport = &pcbinfo->lasthi;
362 } else if (inp->inp_flags & INP_LOWPORT) {
364 (error = suser_cred(p->p_ucred, PRISON_ROOT))) {
365 inp->inp_laddr.s_addr = INADDR_ANY;
368 first = ipport_lowfirstauto; /* 1023 */
369 last = ipport_lowlastauto; /* 600 */
370 lastport = &pcbinfo->lastlow;
372 first = ipport_firstauto; /* sysctl */
373 last = ipport_lastauto;
374 lastport = &pcbinfo->lastport;
377 * Simple check to ensure all ports are not used up causing
380 * We split the two cases (up and down) so that the direction
381 * is not being tested on each round of the loop.
387 if (ipport_randomized)
389 (arc4random() % (first - last));
390 count = first - last;
393 if (count-- < 0) { /* completely used? */
394 inp->inp_laddr.s_addr = INADDR_ANY;
395 return (EADDRNOTAVAIL);
398 if (*lastport > first || *lastport < last)
400 lport = htons(*lastport);
401 } while (in_pcblookup_local(pcbinfo,
402 inp->inp_laddr, lport, wild));
407 if (ipport_randomized)
409 (arc4random() % (last - first));
410 count = last - first;
413 if (count-- < 0) { /* completely used? */
414 inp->inp_laddr.s_addr = INADDR_ANY;
415 return (EADDRNOTAVAIL);
418 if (*lastport < first || *lastport > last)
420 lport = htons(*lastport);
421 } while (in_pcblookup_local(pcbinfo,
422 inp->inp_laddr, lport, wild));
425 inp->inp_lport = lport;
426 if (prison_ip(td, 0, &inp->inp_laddr.s_addr)) {
427 inp->inp_laddr.s_addr = INADDR_ANY;
431 if (in_pcbinsporthash(inp) != 0) {
432 inp->inp_laddr.s_addr = INADDR_ANY;
440 * Transform old in_pcbconnect() into an inner subroutine for new
441 * in_pcbconnect(): Do some validity-checking on the remote
442 * address (in mbuf 'nam') and then determine local host address
443 * (i.e., which interface) to use to access that remote host.
445 * This preserves definition of in_pcbconnect(), while supporting a
446 * slightly different version for T/TCP. (This is more than
447 * a bit of a kludge, but cleaning up the internal interfaces would
448 * have forced minor changes in every protocol).
451 in_pcbladdr(inp, nam, plocal_sin)
453 struct sockaddr *nam;
454 struct sockaddr_in **plocal_sin;
456 struct in_ifaddr *ia;
457 struct sockaddr_in *sin = (struct sockaddr_in *)nam;
459 if (nam->sa_len != sizeof *sin)
461 if (sin->sin_family != AF_INET)
462 return (EAFNOSUPPORT);
463 if (sin->sin_port == 0)
464 return (EADDRNOTAVAIL);
465 if (!TAILQ_EMPTY(&in_ifaddrhead)) {
466 ia = TAILQ_FIRST(&in_ifaddrhead);
468 * If the destination address is INADDR_ANY,
469 * use the primary local address.
470 * If the supplied address is INADDR_BROADCAST,
471 * and the primary interface supports broadcast,
472 * choose the broadcast address for that interface.
474 if (sin->sin_addr.s_addr == INADDR_ANY)
475 sin->sin_addr = IA_SIN(ia)->sin_addr;
476 else if (sin->sin_addr.s_addr == (u_long)INADDR_BROADCAST &&
477 (ia->ia_ifp->if_flags & IFF_BROADCAST))
478 sin->sin_addr = satosin(&ia->ia_broadaddr)->sin_addr;
480 if (inp->inp_laddr.s_addr == INADDR_ANY) {
483 ia = (struct in_ifaddr *)NULL;
485 * If route is known or can be allocated now,
486 * our src addr is taken from the i/f, else punt.
487 * Note that we should check the address family of the cached
488 * destination, in case of sharing the cache with IPv6.
490 ro = &inp->inp_route;
492 (!(ro->ro_rt->rt_flags & RTF_UP) ||
493 ro->ro_dst.sa_family != AF_INET ||
494 satosin(&ro->ro_dst)->sin_addr.s_addr !=
495 sin->sin_addr.s_addr ||
496 inp->inp_socket->so_options & SO_DONTROUTE)) {
498 ro->ro_rt = (struct rtentry *)NULL;
500 if (!(inp->inp_socket->so_options & SO_DONTROUTE) && /*XXX*/
501 (ro->ro_rt == (struct rtentry *)NULL ||
502 ro->ro_rt->rt_ifp == (struct ifnet *)NULL)) {
503 /* No route yet, so try to acquire one */
504 bzero(&ro->ro_dst, sizeof(struct sockaddr_in));
505 ro->ro_dst.sa_family = AF_INET;
506 ro->ro_dst.sa_len = sizeof(struct sockaddr_in);
507 ((struct sockaddr_in *) &ro->ro_dst)->sin_addr =
512 * If we found a route, use the address
513 * corresponding to the outgoing interface
514 * unless it is the loopback (in case a route
515 * to our address on another net goes to loopback).
517 if (ro->ro_rt && !(ro->ro_rt->rt_ifp->if_flags & IFF_LOOPBACK))
518 ia = ifatoia(ro->ro_rt->rt_ifa);
520 u_short fport = sin->sin_port;
523 ia = ifatoia(ifa_ifwithdstaddr(sintosa(sin)));
525 ia = ifatoia(ifa_ifwithnet(sintosa(sin)));
526 sin->sin_port = fport;
528 ia = TAILQ_FIRST(&in_ifaddrhead);
530 return (EADDRNOTAVAIL);
533 * If the destination address is multicast and an outgoing
534 * interface has been set as a multicast option, use the
535 * address of that interface as our source address.
537 if (IN_MULTICAST(ntohl(sin->sin_addr.s_addr)) &&
538 inp->inp_moptions != NULL) {
539 struct ip_moptions *imo;
542 imo = inp->inp_moptions;
543 if (imo->imo_multicast_ifp != NULL) {
544 ifp = imo->imo_multicast_ifp;
545 TAILQ_FOREACH(ia, &in_ifaddrhead, ia_link)
546 if (ia->ia_ifp == ifp)
549 return (EADDRNOTAVAIL);
553 * Don't do pcblookup call here; return interface in plocal_sin
554 * and exit to caller, that will do the lookup.
556 *plocal_sin = &ia->ia_addr;
564 * Connect from a socket to a specified address.
565 * Both address and port must be specified in argument sin.
566 * If don't have a local address for this socket yet,
570 in_pcbconnect(struct inpcb *inp, struct sockaddr *nam, struct thread *td)
572 struct sockaddr_in *if_sin;
573 struct sockaddr_in *sin = (struct sockaddr_in *)nam;
574 struct sockaddr_in sa;
575 struct ucred *cr = td->td_proc ? td->td_proc->p_ucred : NULL;
578 if (cr && cr->cr_prison != NULL && in_nullhost(inp->inp_laddr)) {
579 bzero(&sa, sizeof sa);
580 sa.sin_addr.s_addr = htonl(cr->cr_prison->pr_ip);
581 sa.sin_len = sizeof sa;
582 sa.sin_family = AF_INET;
583 error = in_pcbbind(inp, (struct sockaddr *)&sa, td);
588 /* Call inner routine to assign local interface address. */
589 if ((error = in_pcbladdr(inp, nam, &if_sin)) != 0)
592 if (in_pcblookup_hash(inp->inp_cpcbinfo, sin->sin_addr, sin->sin_port,
593 inp->inp_laddr.s_addr ? inp->inp_laddr : if_sin->sin_addr,
594 inp->inp_lport, FALSE, NULL) != NULL) {
597 if (inp->inp_laddr.s_addr == INADDR_ANY) {
598 if (inp->inp_lport == 0) {
599 error = in_pcbbind(inp, (struct sockaddr *)NULL, td);
603 inp->inp_laddr = if_sin->sin_addr;
605 inp->inp_faddr = sin->sin_addr;
606 inp->inp_fport = sin->sin_port;
607 in_pcbinsconnhash(inp);
612 in_pcbdisconnect(inp)
616 inp->inp_faddr.s_addr = INADDR_ANY;
618 in_pcbremconnhash(inp);
619 if (inp->inp_socket->so_state & SS_NOFDREF)
627 struct socket *so = inp->inp_socket;
628 struct inpcbinfo *ipi = inp->inp_pcbinfo;
631 ipsec4_delete_pcbpolicy(inp);
633 inp->inp_gencnt = ++ipi->ipi_gencnt;
637 if (inp->inp_options)
638 (void)m_free(inp->inp_options);
639 if (inp->inp_route.ro_rt)
640 rtfree(inp->inp_route.ro_rt);
641 ip_freemoptions(inp->inp_moptions);
643 zfree(ipi->ipi_zone, inp);
647 * The calling convention of in_setsockaddr() and in_setpeeraddr() was
648 * modified to match the pru_sockaddr() and pru_peeraddr() entry points
649 * in struct pr_usrreqs, so that protocols can just reference then directly
650 * without the need for a wrapper function. The socket must have a valid
651 * (i.e., non-nil) PCB, but it should be impossible to get an invalid one
652 * except through a kernel programming error, so it is acceptable to panic
653 * (or in this case trap) if the PCB is invalid. (Actually, we don't trap
654 * because there actually /is/ a programming error somewhere... XXX)
657 in_setsockaddr(so, nam)
659 struct sockaddr **nam;
663 struct sockaddr_in *sin;
666 * Do the malloc first in case it blocks.
668 MALLOC(sin, struct sockaddr_in *, sizeof *sin, M_SONAME,
670 sin->sin_family = AF_INET;
671 sin->sin_len = sizeof *sin;
680 sin->sin_port = inp->inp_lport;
681 sin->sin_addr = inp->inp_laddr;
684 *nam = (struct sockaddr *)sin;
689 in_setpeeraddr(so, nam)
691 struct sockaddr **nam;
695 struct sockaddr_in *sin;
698 * Do the malloc first in case it blocks.
700 MALLOC(sin, struct sockaddr_in *, sizeof *sin, M_SONAME,
702 sin->sin_family = AF_INET;
703 sin->sin_len = sizeof *sin;
712 sin->sin_port = inp->inp_fport;
713 sin->sin_addr = inp->inp_faddr;
716 *nam = (struct sockaddr *)sin;
721 in_pcbnotifyall(head, faddr, errno, notify)
722 struct inpcbhead *head;
723 struct in_addr faddr;
724 void (*notify) (struct inpcb *, int);
726 struct inpcb *inp, *ninp;
730 * note: if INP_PLACEMARKER is set we must ignore the rest of
731 * the structure and skip it.
734 for (inp = LIST_FIRST(head); inp != NULL; inp = ninp) {
735 ninp = LIST_NEXT(inp, inp_list);
736 if (inp->inp_flags & INP_PLACEMARKER)
739 if (!(inp->inp_vflag & INP_IPV4))
742 if (inp->inp_faddr.s_addr != faddr.s_addr ||
743 inp->inp_socket == NULL)
745 (*notify)(inp, errno);
751 in_pcbpurgeif0(head, ifp)
756 struct ip_moptions *imo;
759 for (inp = head; inp != NULL; inp = LIST_NEXT(inp, inp_list)) {
760 if (inp->inp_flags & INP_PLACEMARKER)
762 imo = inp->inp_moptions;
763 if ((inp->inp_vflag & INP_IPV4) && imo != NULL) {
765 * Unselect the outgoing interface if it is being
768 if (imo->imo_multicast_ifp == ifp)
769 imo->imo_multicast_ifp = NULL;
772 * Drop multicast group membership if we joined
773 * through the interface being detached.
775 for (i = 0, gap = 0; i < imo->imo_num_memberships;
777 if (imo->imo_membership[i]->inm_ifp == ifp) {
778 in_delmulti(imo->imo_membership[i]);
781 imo->imo_membership[i - gap] =
782 imo->imo_membership[i];
784 imo->imo_num_memberships -= gap;
790 * Check for alternatives when higher level complains
791 * about service problems. For now, invalidate cached
792 * routing information. If the route was created dynamically
793 * (by a redirect), time to try a default gateway again.
800 struct rt_addrinfo info;
802 if ((rt = inp->inp_route.ro_rt)) {
803 bzero((caddr_t)&info, sizeof info);
804 info.rti_flags = rt->rt_flags;
805 info.rti_info[RTAX_DST] = rt_key(rt);
806 info.rti_info[RTAX_GATEWAY] = rt->rt_gateway;
807 info.rti_info[RTAX_NETMASK] = rt_mask(rt);
808 rt_missmsg(RTM_LOSING, &info, rt->rt_flags, 0);
809 if (rt->rt_flags & RTF_DYNAMIC)
810 (void) rtrequest1(RTM_DELETE, &info, NULL);
811 inp->inp_route.ro_rt = NULL;
814 * A new route can be allocated
815 * the next time output is attempted.
821 * After a routing change, flush old routing
822 * and allocate a (hopefully) better one.
825 in_rtchange(inp, errno)
829 if (inp->inp_route.ro_rt) {
830 rtfree(inp->inp_route.ro_rt);
831 inp->inp_route.ro_rt = 0;
833 * A new route can be allocated the next time
834 * output is attempted.
840 * Lookup a PCB based on the local address and port.
843 in_pcblookup_local(pcbinfo, laddr, lport_arg, wild_okay)
844 struct inpcbinfo *pcbinfo;
845 struct in_addr laddr;
850 int matchwild = 3, wildcard;
851 u_short lport = lport_arg;
853 struct inpcbporthead *porthash;
854 struct inpcbport *phd;
855 struct inpcb *match = NULL;
858 * Best fit PCB lookup.
860 * First see if this local port is in use by looking on the
863 porthash = &pcbinfo->porthashbase[INP_PCBPORTHASH(lport,
864 pcbinfo->porthashmask)];
865 LIST_FOREACH(phd, porthash, phd_hash) {
866 if (phd->phd_port == lport)
871 * Port is in use by one or more PCBs. Look for best
874 LIST_FOREACH(inp, &phd->phd_pcblist, inp_portlist) {
877 if ((inp->inp_vflag & INP_IPV4) == 0)
880 if (inp->inp_faddr.s_addr != INADDR_ANY)
882 if (inp->inp_laddr.s_addr != INADDR_ANY) {
883 if (laddr.s_addr == INADDR_ANY)
885 else if (inp->inp_laddr.s_addr != laddr.s_addr)
888 if (laddr.s_addr != INADDR_ANY)
891 if (wildcard && !wild_okay)
893 if (wildcard < matchwild) {
895 matchwild = wildcard;
896 if (matchwild == 0) {
906 * Lookup PCB in hash list.
909 in_pcblookup_hash(pcbinfo, faddr, fport_arg, laddr, lport_arg, wildcard, ifp)
910 struct inpcbinfo *pcbinfo;
911 struct in_addr faddr, laddr;
912 u_int fport_arg, lport_arg;
916 struct inpcbhead *head;
918 u_short fport = fport_arg, lport = lport_arg;
921 * First look for an exact match.
923 head = &pcbinfo->hashbase[INP_PCBCONNHASH(faddr.s_addr, fport,
924 laddr.s_addr, lport, pcbinfo->hashmask)];
925 LIST_FOREACH(inp, head, inp_hash) {
927 if (!(inp->inp_vflag & INP_IPV4))
930 if (in_hosteq(inp->inp_faddr, faddr) &&
931 in_hosteq(inp->inp_laddr, laddr) &&
932 inp->inp_fport == fport && inp->inp_lport == lport) {
939 struct inpcb *local_wild = NULL;
941 struct inpcb *local_wild_mapped = NULL;
943 struct inpcontainer *ic;
944 struct inpcontainerhead *chead;
946 chead = &pcbinfo->wildcardhashbase[
947 INP_PCBWILDCARDHASH(lport, pcbinfo->wildcardhashmask)];
948 LIST_FOREACH(ic, chead, ic_list) {
951 if (!(inp->inp_vflag & INP_IPV4))
954 if (inp->inp_lport == lport) {
955 if (ifp && ifp->if_type == IFT_FAITH &&
956 !(inp->inp_flags & INP_FAITH))
958 if (inp->inp_laddr.s_addr == laddr.s_addr)
960 if (inp->inp_laddr.s_addr == INADDR_ANY) {
962 if (INP_CHECK_SOCKAF(inp->inp_socket,
964 local_wild_mapped = inp;
972 if (local_wild == NULL)
973 return (local_wild_mapped);
985 * Insert PCB into connection hash table.
988 in_pcbinsconnhash(struct inpcb *inp)
990 struct inpcbinfo *pcbinfo = inp->inp_cpcbinfo;
991 struct inpcbhead *bucket;
992 u_int32_t hashkey_faddr, hashkey_laddr;
995 if (inp->inp_vflag & INP_IPV6) {
996 hashkey_faddr = inp->in6p_faddr.s6_addr32[3] /* XXX JH */;
997 hashkey_laddr = inp->in6p_laddr.s6_addr32[3] /* XXX JH */;
1000 hashkey_faddr = inp->inp_faddr.s_addr;
1001 hashkey_laddr = inp->inp_laddr.s_addr;
1006 KASSERT(!(inp->inp_flags & INP_CONNECTED), ("already on hash list"));
1007 inp->inp_flags |= INP_CONNECTED;
1010 * Insert into the connection hash table.
1012 bucket = &pcbinfo->hashbase[INP_PCBCONNHASH(hashkey_faddr,
1013 inp->inp_fport, hashkey_laddr, inp->inp_lport, pcbinfo->hashmask)];
1014 LIST_INSERT_HEAD(bucket, inp, inp_hash);
1018 * Remove PCB from connection hash table.
1021 in_pcbremconnhash(struct inpcb *inp)
1023 KASSERT(inp->inp_flags & INP_CONNECTED, ("inp not connected"));
1024 LIST_REMOVE(inp, inp_hash);
1025 inp->inp_flags &= ~INP_CONNECTED;
1029 * Insert PCB into port hash table.
1032 in_pcbinsporthash(struct inpcb *inp)
1034 struct inpcbinfo *pcbinfo = inp->inp_pcbinfo;
1035 struct inpcbporthead *pcbporthash;
1036 struct inpcbport *phd;
1039 * Insert into the port hash table.
1041 pcbporthash = &pcbinfo->porthashbase[
1042 INP_PCBPORTHASH(inp->inp_lport, pcbinfo->porthashmask)];
1044 /* Go through port list and look for a head for this lport. */
1045 LIST_FOREACH(phd, pcbporthash, phd_hash)
1046 if (phd->phd_port == inp->inp_lport)
1049 /* If none exists, malloc one and tack it on. */
1051 MALLOC(phd, struct inpcbport *, sizeof(struct inpcbport),
1052 M_PCB, M_INTWAIT | M_NULLOK);
1054 return (ENOBUFS); /* XXX */
1055 phd->phd_port = inp->inp_lport;
1056 LIST_INIT(&phd->phd_pcblist);
1057 LIST_INSERT_HEAD(pcbporthash, phd, phd_hash);
1061 LIST_INSERT_HEAD(&phd->phd_pcblist, inp, inp_portlist);
1067 in_pcbinswildcardhash_oncpu(struct inpcb *inp, struct inpcbinfo *pcbinfo)
1069 struct inpcontainer *ic;
1070 struct inpcontainerhead *bucket;
1072 bucket = &pcbinfo->wildcardhashbase[
1073 INP_PCBWILDCARDHASH(inp->inp_lport, pcbinfo->wildcardhashmask)];
1075 ic = malloc(sizeof(struct inpcontainer), M_TEMP, M_INTWAIT);
1077 LIST_INSERT_HEAD(bucket, ic, ic_list);
1081 * Insert PCB into wildcard hash table.
1084 in_pcbinswildcardhash(struct inpcb *inp)
1086 struct inpcbinfo *pcbinfo = inp->inp_pcbinfo;
1088 KKASSERT(pcbinfo != NULL);
1090 in_pcbinswildcardhash_oncpu(inp, pcbinfo);
1091 inp->inp_flags |= INP_WILDCARD;
1095 in_pcbremwildcardhash_oncpu(struct inpcb *inp, struct inpcbinfo *pcbinfo)
1097 struct inpcontainer *ic;
1098 struct inpcontainerhead *head;
1101 head = &pcbinfo->wildcardhashbase[
1102 INP_PCBWILDCARDHASH(inp->inp_lport, pcbinfo->wildcardhashmask)];
1104 LIST_FOREACH(ic, head, ic_list) {
1105 if (ic->ic_inp == inp)
1108 return; /* not found! */
1111 LIST_REMOVE(ic, ic_list); /* remove container from bucket chain */
1112 free(ic, M_TEMP); /* deallocate container */
1116 * Remove PCB from wildcard hash table.
1119 in_pcbremwildcardhash(struct inpcb *inp)
1121 struct inpcbinfo *pcbinfo = inp->inp_pcbinfo;
1123 KASSERT(inp->inp_flags & INP_WILDCARD, ("inp not wildcard"));
1124 in_pcbremwildcardhash_oncpu(inp, pcbinfo);
1125 inp->inp_flags &= ~INP_WILDCARD;
1129 * Remove PCB from various lists.
1135 if (inp->inp_lport) {
1136 struct inpcbport *phd = inp->inp_phd;
1138 LIST_REMOVE(inp, inp_portlist);
1139 if (LIST_FIRST(&phd->phd_pcblist) == NULL) {
1140 LIST_REMOVE(phd, phd_hash);
1144 if (inp->inp_flags & INP_WILDCARD) {
1145 in_pcbremwildcardhash(inp);
1146 } else if (inp->inp_flags & INP_CONNECTED) {
1147 in_pcbremconnhash(inp);
1149 LIST_REMOVE(inp, inp_list);
1150 inp->inp_pcbinfo->ipi_count--;
1154 prison_xinpcb(struct thread *td, struct inpcb *inp)
1158 if (td->td_proc == NULL)
1160 cr = td->td_proc->p_ucred;
1161 if (cr->cr_prison == NULL)
1163 if (ntohl(inp->inp_laddr.s_addr) == cr->cr_prison->pr_ip)