]> git.saurik.com Git - apple/xnu.git/blob - bsd/netinet/raw_ip.c
xnu-1228.12.14.tar.gz
[apple/xnu.git] / bsd / netinet / raw_ip.c
1 /*
2 * Copyright (c) 2000-2008 Apple Inc. All rights reserved.
3 *
4 * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
5 *
6 * This file contains Original Code and/or Modifications of Original Code
7 * as defined in and that are subject to the Apple Public Source License
8 * Version 2.0 (the 'License'). You may not use this file except in
9 * compliance with the License. The rights granted to you under the License
10 * may not be used to create, or enable the creation or redistribution of,
11 * unlawful or unlicensed copies of an Apple operating system, or to
12 * circumvent, violate, or enable the circumvention or violation of, any
13 * terms of an Apple operating system software license agreement.
14 *
15 * Please obtain a copy of the License at
16 * http://www.opensource.apple.com/apsl/ and read it before using this file.
17 *
18 * The Original Code and all software distributed under the License are
19 * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
20 * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
21 * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
22 * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
23 * Please see the License for the specific language governing rights and
24 * limitations under the License.
25 *
26 * @APPLE_OSREFERENCE_LICENSE_HEADER_END@
27 */
28 /*
29 * Copyright (c) 1982, 1986, 1988, 1993
30 * The Regents of the University of California. All rights reserved.
31 *
32 * Redistribution and use in source and binary forms, with or without
33 * modification, are permitted provided that the following conditions
34 * are met:
35 * 1. Redistributions of source code must retain the above copyright
36 * notice, this list of conditions and the following disclaimer.
37 * 2. Redistributions in binary form must reproduce the above copyright
38 * notice, this list of conditions and the following disclaimer in the
39 * documentation and/or other materials provided with the distribution.
40 * 3. All advertising materials mentioning features or use of this software
41 * must display the following acknowledgement:
42 * This product includes software developed by the University of
43 * California, Berkeley and its contributors.
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 * @(#)raw_ip.c 8.7 (Berkeley) 5/15/95
61 */
62 /*
63 * NOTICE: This file was modified by SPARTA, Inc. in 2005 to introduce
64 * support for mandatory and extensible security protections. This notice
65 * is included in support of clause 2.2 (b) of the Apple Public License,
66 * Version 2.0.
67 */
68
69 #include <sys/param.h>
70 #include <sys/systm.h>
71 #include <sys/kernel.h>
72 #include <sys/malloc.h>
73 #include <sys/mbuf.h>
74 #include <sys/proc.h>
75 #include <sys/domain.h>
76 #include <sys/protosw.h>
77 #include <sys/socket.h>
78 #include <sys/socketvar.h>
79 #include <sys/sysctl.h>
80 #include <libkern/OSAtomic.h>
81 #include <kern/zalloc.h>
82
83 #include <pexpert/pexpert.h>
84
85 #include <net/if.h>
86 #include <net/route.h>
87
88 #define _IP_VHL
89 #include <netinet/in.h>
90 #include <netinet/in_systm.h>
91 #include <netinet/ip.h>
92 #include <netinet/in_pcb.h>
93 #include <netinet/in_var.h>
94 #include <netinet/ip_var.h>
95 #include <netinet/ip_mroute.h>
96
97 #include <netinet/ip_fw.h>
98
99 #if IPSEC
100 #include <netinet6/ipsec.h>
101 #endif /*IPSEC*/
102
103 #if DUMMYNET
104 #include <netinet/ip_dummynet.h>
105 #endif
106
107 #if CONFIG_MACF_NET
108 #include <security/mac_framework.h>
109 #endif /* MAC_NET */
110
111 int load_ipfw(void);
112 int rip_detach(struct socket *);
113 int rip_abort(struct socket *);
114 int rip_disconnect(struct socket *);
115 int rip_bind(struct socket *, struct sockaddr *, struct proc *);
116 int rip_connect(struct socket *, struct sockaddr *, struct proc *);
117 int rip_shutdown(struct socket *);
118
119 #if IPSEC
120 extern int ipsec_bypass;
121 #endif
122
123 extern u_long route_generation;
124 struct inpcbhead ripcb;
125 struct inpcbinfo ripcbinfo;
126
127 /* control hooks for ipfw and dummynet */
128 #if IPFIREWALL
129 ip_fw_ctl_t *ip_fw_ctl_ptr;
130 #if DUMMYNET
131 ip_dn_ctl_t *ip_dn_ctl_ptr;
132 #endif /* DUMMYNET */
133 #endif /* IPFIREWALL */
134
135 /*
136 * Nominal space allocated to a raw ip socket.
137 */
138 #define RIPSNDQ 8192
139 #define RIPRCVQ 8192
140
141 /*
142 * Raw interface to IP protocol.
143 */
144
145 /*
146 * Initialize raw connection block q.
147 */
148 void
149 rip_init()
150 {
151 struct inpcbinfo *pcbinfo;
152
153 LIST_INIT(&ripcb);
154 ripcbinfo.listhead = &ripcb;
155 /*
156 * XXX We don't use the hash list for raw IP, but it's easier
157 * to allocate a one entry hash list than it is to check all
158 * over the place for hashbase == NULL.
159 */
160 ripcbinfo.hashbase = hashinit(1, M_PCB, &ripcbinfo.hashmask);
161 ripcbinfo.porthashbase = hashinit(1, M_PCB, &ripcbinfo.porthashmask);
162
163 ripcbinfo.ipi_zone = (void *) zinit(sizeof(struct inpcb),
164 (4096 * sizeof(struct inpcb)),
165 4096, "ripzone");
166
167 pcbinfo = &ripcbinfo;
168 /*
169 * allocate lock group attribute and group for udp pcb mutexes
170 */
171 pcbinfo->mtx_grp_attr = lck_grp_attr_alloc_init();
172
173 pcbinfo->mtx_grp = lck_grp_alloc_init("ripcb", pcbinfo->mtx_grp_attr);
174
175 /*
176 * allocate the lock attribute for udp pcb mutexes
177 */
178 pcbinfo->mtx_attr = lck_attr_alloc_init();
179
180 if ((pcbinfo->mtx = lck_rw_alloc_init(pcbinfo->mtx_grp, pcbinfo->mtx_attr)) == NULL)
181 return; /* pretty much dead if this fails... */
182
183 }
184
185 static struct sockaddr_in ripsrc = { sizeof(ripsrc), AF_INET , 0, {0}, {0,0,0,0,0,0,0,0,} };
186 /*
187 * Setup generic address and protocol structures
188 * for raw_input routine, then pass them along with
189 * mbuf chain.
190 */
191 void
192 rip_input(m, iphlen)
193 struct mbuf *m;
194 int iphlen;
195 {
196 register struct ip *ip = mtod(m, struct ip *);
197 register struct inpcb *inp;
198 struct inpcb *last = 0;
199 struct mbuf *opts = 0;
200 int skipit;
201
202 ripsrc.sin_addr = ip->ip_src;
203 lck_rw_lock_shared(ripcbinfo.mtx);
204 LIST_FOREACH(inp, &ripcb, inp_list) {
205 #if INET6
206 if ((inp->inp_vflag & INP_IPV4) == 0)
207 continue;
208 #endif
209 if (inp->inp_ip_p && (inp->inp_ip_p != ip->ip_p))
210 continue;
211 if (inp->inp_laddr.s_addr &&
212 inp->inp_laddr.s_addr != ip->ip_dst.s_addr)
213 continue;
214 if (inp->inp_faddr.s_addr &&
215 inp->inp_faddr.s_addr != ip->ip_src.s_addr)
216 continue;
217 if (last) {
218 struct mbuf *n = m_copy(m, 0, (int)M_COPYALL);
219
220 #if IPSEC
221 /* check AH/ESP integrity. */
222 skipit = 0;
223 if (ipsec_bypass == 0 && n) {
224 if (ipsec4_in_reject_so(n, last->inp_socket)) {
225 m_freem(n);
226 IPSEC_STAT_INCREMENT(ipsecstat.in_polvio);
227 /* do not inject data to pcb */
228 skipit = 1;
229 }
230 }
231 #endif /*IPSEC*/
232 #if CONFIG_MACF_NET
233 if (n && skipit == 0) {
234 if (mac_inpcb_check_deliver(last, n, AF_INET,
235 SOCK_RAW) != 0)
236 skipit = 1;
237 }
238 #endif
239 if (n && skipit == 0) {
240 int error = 0;
241 if (last->inp_flags & INP_CONTROLOPTS ||
242 last->inp_socket->so_options & SO_TIMESTAMP)
243 ip_savecontrol(last, &opts, ip, n);
244 if (last->inp_flags & INP_STRIPHDR) {
245 n->m_len -= iphlen;
246 n->m_pkthdr.len -= iphlen;
247 n->m_data += iphlen;
248 }
249 // ###LOCK need to lock that socket?
250 if (sbappendaddr(&last->inp_socket->so_rcv,
251 (struct sockaddr *)&ripsrc, n,
252 opts, &error) != 0) {
253 sorwakeup(last->inp_socket);
254 }
255 else {
256 if (error) {
257 /* should notify about lost packet */
258 kprintf("rip_input can't append to socket\n");
259 }
260 }
261 opts = 0;
262 }
263 }
264 last = inp;
265 }
266 lck_rw_done(ripcbinfo.mtx);
267 #if IPSEC
268 /* check AH/ESP integrity. */
269 skipit = 0;
270 if (ipsec_bypass == 0 && last) {
271 if (ipsec4_in_reject_so(m, last->inp_socket)) {
272 m_freem(m);
273 IPSEC_STAT_INCREMENT(ipsecstat.in_polvio);
274 OSAddAtomic(1, (SInt32*)&ipstat.ips_delivered);
275 /* do not inject data to pcb */
276 skipit = 1;
277 }
278 }
279 #endif /*IPSEC*/
280 #if CONFIG_MACF_NET
281 if (last && skipit == 0) {
282 if (mac_inpcb_check_deliver(last, m, AF_INET, SOCK_RAW) != 0)
283 skipit = 1;
284 }
285 #endif
286 if (skipit == 0) {
287 if (last) {
288 if (last->inp_flags & INP_CONTROLOPTS ||
289 last->inp_socket->so_options & SO_TIMESTAMP)
290 ip_savecontrol(last, &opts, ip, m);
291 if (last->inp_flags & INP_STRIPHDR) {
292 m->m_len -= iphlen;
293 m->m_pkthdr.len -= iphlen;
294 m->m_data += iphlen;
295 }
296 if (sbappendaddr(&last->inp_socket->so_rcv,
297 (struct sockaddr *)&ripsrc, m, opts, NULL) != 0) {
298 sorwakeup(last->inp_socket);
299 } else {
300 kprintf("rip_input(2) can't append to socket\n");
301 }
302 } else {
303 m_freem(m);
304 OSAddAtomic(1, (SInt32*)&ipstat.ips_noproto);
305 OSAddAtomic(-1, (SInt32*)&ipstat.ips_delivered);
306 }
307 }
308 }
309
310 /*
311 * Generate IP header and pass packet to ip_output.
312 * Tack on options user may have setup with control call.
313 */
314 int
315 rip_output(m, so, dst)
316 register struct mbuf *m;
317 struct socket *so;
318 u_long dst;
319 {
320 register struct ip *ip;
321 register struct inpcb *inp = sotoinpcb(so);
322 int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST;
323 struct ip_out_args ipoa;
324
325 /* If socket was bound to an ifindex, tell ip_output about it */
326 ipoa.ipoa_ifscope = (inp->inp_flags & INP_BOUND_IF) ?
327 inp->inp_boundif : IFSCOPE_NONE;
328 flags |= IP_OUTARGS;
329
330 /*
331 * If the user handed us a complete IP packet, use it.
332 * Otherwise, allocate an mbuf for a header and fill it in.
333 */
334 if ((inp->inp_flags & INP_HDRINCL) == 0) {
335 if (m->m_pkthdr.len + sizeof(struct ip) > IP_MAXPACKET) {
336 m_freem(m);
337 return(EMSGSIZE);
338 }
339 M_PREPEND(m, sizeof(struct ip), M_WAIT);
340 ip = mtod(m, struct ip *);
341 ip->ip_tos = inp->inp_ip_tos;
342 ip->ip_off = 0;
343 ip->ip_p = inp->inp_ip_p;
344 ip->ip_len = m->m_pkthdr.len;
345 ip->ip_src = inp->inp_laddr;
346 ip->ip_dst.s_addr = dst;
347 ip->ip_ttl = inp->inp_ip_ttl;
348 } else {
349 if (m->m_pkthdr.len > IP_MAXPACKET) {
350 m_freem(m);
351 return(EMSGSIZE);
352 }
353 ip = mtod(m, struct ip *);
354 /* don't allow both user specified and setsockopt options,
355 and don't allow packet length sizes that will crash */
356 if (((IP_VHL_HL(ip->ip_vhl) != (sizeof (*ip) >> 2))
357 && inp->inp_options)
358 || (ip->ip_len > m->m_pkthdr.len)
359 || (ip->ip_len < (IP_VHL_HL(ip->ip_vhl) << 2))) {
360 m_freem(m);
361 return EINVAL;
362 }
363 if (ip->ip_id == 0)
364 #if RANDOM_IP_ID
365 ip->ip_id = ip_randomid();
366 #else
367 ip->ip_id = htons(ip_id++);
368 #endif
369 /* XXX prevent ip_output from overwriting header fields */
370 flags |= IP_RAWOUTPUT;
371 OSAddAtomic(1, (SInt32*)&ipstat.ips_rawout);
372 }
373
374 #if IPSEC
375 if (ipsec_bypass == 0 && ipsec_setsocket(m, so) != 0) {
376 m_freem(m);
377 return ENOBUFS;
378 }
379 #endif /*IPSEC*/
380
381 if (inp->inp_route.ro_rt && inp->inp_route.ro_rt->generation_id != route_generation) {
382 rtfree(inp->inp_route.ro_rt);
383 inp->inp_route.ro_rt = (struct rtentry *)0;
384 }
385
386 #if CONFIG_MACF_NET
387 mac_mbuf_label_associate_inpcb(inp, m);
388 #endif
389
390 #if CONFIG_IP_EDGEHOLE
391 ip_edgehole_mbuf_tag(inp, m);
392 #endif
393 return (ip_output(m, inp->inp_options, &inp->inp_route, flags,
394 inp->inp_moptions, &ipoa));
395 }
396
397 #if IPFIREWALL
398 int
399 load_ipfw(void)
400 {
401 kern_return_t err;
402
403 ipfw_init();
404
405 #if DUMMYNET
406 if (!DUMMYNET_LOADED)
407 ip_dn_init();
408 #endif /* DUMMYNET */
409 err = 0;
410
411 return err == 0 && ip_fw_ctl_ptr == NULL ? -1 : err;
412 }
413 #endif /* IPFIREWALL */
414
415 /*
416 * Raw IP socket option processing.
417 */
418 int
419 rip_ctloutput(so, sopt)
420 struct socket *so;
421 struct sockopt *sopt;
422 {
423 struct inpcb *inp = sotoinpcb(so);
424 int error, optval;
425
426 if (sopt->sopt_level != IPPROTO_IP)
427 return (EINVAL);
428
429 error = 0;
430
431 switch (sopt->sopt_dir) {
432 case SOPT_GET:
433 switch (sopt->sopt_name) {
434 case IP_HDRINCL:
435 optval = inp->inp_flags & INP_HDRINCL;
436 error = sooptcopyout(sopt, &optval, sizeof optval);
437 break;
438
439 case IP_STRIPHDR:
440 optval = inp->inp_flags & INP_STRIPHDR;
441 error = sooptcopyout(sopt, &optval, sizeof optval);
442 break;
443
444 #if IPFIREWALL
445 case IP_FW_ADD:
446 case IP_FW_GET:
447 case IP_OLD_FW_ADD:
448 case IP_OLD_FW_GET:
449 if (ip_fw_ctl_ptr == 0)
450 error = load_ipfw();
451 if (ip_fw_ctl_ptr && error == 0)
452 error = ip_fw_ctl_ptr(sopt);
453 else
454 error = ENOPROTOOPT;
455 break;
456 #endif IPFIREWALL
457
458 #if DUMMYNET
459 case IP_DUMMYNET_GET:
460 if (DUMMYNET_LOADED)
461 error = ip_dn_ctl_ptr(sopt);
462 else
463 error = ENOPROTOOPT;
464 break ;
465 #endif /* DUMMYNET */
466
467 #if MROUTING
468 case MRT_INIT:
469 case MRT_DONE:
470 case MRT_ADD_VIF:
471 case MRT_DEL_VIF:
472 case MRT_ADD_MFC:
473 case MRT_DEL_MFC:
474 case MRT_VERSION:
475 case MRT_ASSERT:
476 error = ip_mrouter_get(so, sopt);
477 break;
478 #endif /* MROUTING */
479
480 default:
481 error = ip_ctloutput(so, sopt);
482 break;
483 }
484 break;
485
486 case SOPT_SET:
487 switch (sopt->sopt_name) {
488 case IP_HDRINCL:
489 error = sooptcopyin(sopt, &optval, sizeof optval,
490 sizeof optval);
491 if (error)
492 break;
493 if (optval)
494 inp->inp_flags |= INP_HDRINCL;
495 else
496 inp->inp_flags &= ~INP_HDRINCL;
497 break;
498
499 case IP_STRIPHDR:
500 error = sooptcopyin(sopt, &optval, sizeof optval,
501 sizeof optval);
502 if (error)
503 break;
504 if (optval)
505 inp->inp_flags |= INP_STRIPHDR;
506 else
507 inp->inp_flags &= ~INP_STRIPHDR;
508 break;
509
510
511 #if IPFIREWALL
512 case IP_FW_ADD:
513 case IP_FW_DEL:
514 case IP_FW_FLUSH:
515 case IP_FW_ZERO:
516 case IP_FW_RESETLOG:
517 case IP_OLD_FW_ADD:
518 case IP_OLD_FW_DEL:
519 case IP_OLD_FW_FLUSH:
520 case IP_OLD_FW_ZERO:
521 case IP_OLD_FW_RESETLOG:
522 if (ip_fw_ctl_ptr == 0)
523 error = load_ipfw();
524 if (ip_fw_ctl_ptr && error == 0)
525 error = ip_fw_ctl_ptr(sopt);
526 else
527 error = ENOPROTOOPT;
528 break;
529 #endif /* IPFIREWALL */
530
531 #if DUMMYNET
532 case IP_DUMMYNET_CONFIGURE:
533 case IP_DUMMYNET_DEL:
534 case IP_DUMMYNET_FLUSH:
535 if (DUMMYNET_LOADED)
536 error = ip_dn_ctl_ptr(sopt);
537 else
538 error = ENOPROTOOPT ;
539 break ;
540 #endif
541
542 #if MROUTING
543 case IP_RSVP_ON:
544 error = ip_rsvp_init(so);
545 break;
546
547 case IP_RSVP_OFF:
548 error = ip_rsvp_done();
549 break;
550
551 /* XXX - should be combined */
552 case IP_RSVP_VIF_ON:
553 error = ip_rsvp_vif_init(so, sopt);
554 break;
555
556 case IP_RSVP_VIF_OFF:
557 error = ip_rsvp_vif_done(so, sopt);
558 break;
559
560 case MRT_INIT:
561 case MRT_DONE:
562 case MRT_ADD_VIF:
563 case MRT_DEL_VIF:
564 case MRT_ADD_MFC:
565 case MRT_DEL_MFC:
566 case MRT_VERSION:
567 case MRT_ASSERT:
568 error = ip_mrouter_set(so, sopt);
569 break;
570 #endif /* MROUTING */
571
572 default:
573 error = ip_ctloutput(so, sopt);
574 break;
575 }
576 break;
577 }
578
579 return (error);
580 }
581
582 /*
583 * This function exists solely to receive the PRC_IFDOWN messages which
584 * are sent by if_down(). It looks for an ifaddr whose ifa_addr is sa,
585 * and calls in_ifadown() to remove all routes corresponding to that address.
586 * It also receives the PRC_IFUP messages from if_up() and reinstalls the
587 * interface routes.
588 */
589 void
590 rip_ctlinput(
591 int cmd,
592 struct sockaddr *sa,
593 __unused void *vip)
594 {
595 struct in_ifaddr *ia;
596 struct ifnet *ifp;
597 int err;
598 int flags;
599
600 switch (cmd) {
601 case PRC_IFDOWN:
602 lck_mtx_lock(rt_mtx);
603 for (ia = in_ifaddrhead.tqh_first; ia;
604 ia = ia->ia_link.tqe_next) {
605 if (ia->ia_ifa.ifa_addr == sa
606 && (ia->ia_flags & IFA_ROUTE)) {
607 /*
608 * in_ifscrub kills the interface route.
609 */
610 in_ifscrub(ia->ia_ifp, ia, 1);
611 /*
612 * in_ifadown gets rid of all the rest of
613 * the routes. This is not quite the right
614 * thing to do, but at least if we are running
615 * a routing process they will come back.
616 */
617 in_ifadown(&ia->ia_ifa, 1);
618 break;
619 }
620 }
621 lck_mtx_unlock(rt_mtx);
622 break;
623
624 case PRC_IFUP:
625 lck_mtx_lock(rt_mtx);
626 for (ia = in_ifaddrhead.tqh_first; ia;
627 ia = ia->ia_link.tqe_next) {
628 if (ia->ia_ifa.ifa_addr == sa)
629 break;
630 }
631 if (ia == 0 || (ia->ia_flags & IFA_ROUTE)) {
632 lck_mtx_unlock(rt_mtx);
633 return;
634 }
635 flags = RTF_UP;
636 ifp = ia->ia_ifa.ifa_ifp;
637
638 if ((ifp->if_flags & IFF_LOOPBACK)
639 || (ifp->if_flags & IFF_POINTOPOINT))
640 flags |= RTF_HOST;
641
642 err = rtinit_locked(&ia->ia_ifa, RTM_ADD, flags);
643 lck_mtx_unlock(rt_mtx);
644 if (err == 0)
645 ia->ia_flags |= IFA_ROUTE;
646 break;
647 }
648 }
649
650 u_long rip_sendspace = RIPSNDQ;
651 u_long rip_recvspace = RIPRCVQ;
652
653 SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW,
654 &rip_sendspace, 0, "Maximum outgoing raw IP datagram size");
655 SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW,
656 &rip_recvspace, 0, "Maximum incoming raw IP datagram size");
657
658 static int
659 rip_attach(struct socket *so, int proto, struct proc *p)
660 {
661 struct inpcb *inp;
662 int error;
663
664 inp = sotoinpcb(so);
665 if (inp)
666 panic("rip_attach");
667 if ((so->so_state & SS_PRIV) == 0)
668 return (EPERM);
669
670 error = soreserve(so, rip_sendspace, rip_recvspace);
671 if (error)
672 return error;
673 error = in_pcballoc(so, &ripcbinfo, p);
674 if (error)
675 return error;
676 inp = (struct inpcb *)so->so_pcb;
677 inp->inp_vflag |= INP_IPV4;
678 inp->inp_ip_p = proto;
679 inp->inp_ip_ttl = ip_defttl;
680 return 0;
681 }
682
683 __private_extern__ int
684 rip_detach(struct socket *so)
685 {
686 struct inpcb *inp;
687
688 inp = sotoinpcb(so);
689 if (inp == 0)
690 panic("rip_detach");
691 #if MROUTING
692 if (so == ip_mrouter)
693 ip_mrouter_done();
694 ip_rsvp_force_done(so);
695 if (so == ip_rsvpd)
696 ip_rsvp_done();
697 #endif /* MROUTING */
698 in_pcbdetach(inp);
699 return 0;
700 }
701
702 __private_extern__ int
703 rip_abort(struct socket *so)
704 {
705 soisdisconnected(so);
706 return rip_detach(so);
707 }
708
709 __private_extern__ int
710 rip_disconnect(struct socket *so)
711 {
712 if ((so->so_state & SS_ISCONNECTED) == 0)
713 return ENOTCONN;
714 return rip_abort(so);
715 }
716
717 __private_extern__ int
718 rip_bind(struct socket *so, struct sockaddr *nam, __unused struct proc *p)
719 {
720 struct inpcb *inp = sotoinpcb(so);
721 struct sockaddr_in *addr = (struct sockaddr_in *)nam;
722 struct ifaddr *ifa = NULL;
723
724 if (nam->sa_len != sizeof(*addr))
725 return EINVAL;
726
727 if (TAILQ_EMPTY(&ifnet_head) || ((addr->sin_family != AF_INET) &&
728 (addr->sin_family != AF_IMPLINK)) ||
729 (addr->sin_addr.s_addr &&
730 (ifa = ifa_ifwithaddr((struct sockaddr *)addr)) == 0)) {
731 return EADDRNOTAVAIL;
732 }
733 else if (ifa) {
734 ifafree(ifa);
735 ifa = NULL;
736 }
737 inp->inp_laddr = addr->sin_addr;
738 return 0;
739 }
740
741 __private_extern__ int
742 rip_connect(struct socket *so, struct sockaddr *nam, __unused struct proc *p)
743 {
744 struct inpcb *inp = sotoinpcb(so);
745 struct sockaddr_in *addr = (struct sockaddr_in *)nam;
746
747 if (nam->sa_len != sizeof(*addr))
748 return EINVAL;
749 if (TAILQ_EMPTY(&ifnet_head))
750 return EADDRNOTAVAIL;
751 if ((addr->sin_family != AF_INET) &&
752 (addr->sin_family != AF_IMPLINK))
753 return EAFNOSUPPORT;
754 inp->inp_faddr = addr->sin_addr;
755 soisconnected(so);
756 return 0;
757 }
758
759 __private_extern__ int
760 rip_shutdown(struct socket *so)
761 {
762 socantsendmore(so);
763 return 0;
764 }
765
766 __private_extern__ int
767 rip_send(struct socket *so, __unused int flags, struct mbuf *m, struct sockaddr *nam,
768 __unused struct mbuf *control, __unused struct proc *p)
769 {
770 struct inpcb *inp = sotoinpcb(so);
771 register u_long dst;
772
773 if (so->so_state & SS_ISCONNECTED) {
774 if (nam) {
775 m_freem(m);
776 return EISCONN;
777 }
778 dst = inp->inp_faddr.s_addr;
779 } else {
780 if (nam == NULL) {
781 m_freem(m);
782 return ENOTCONN;
783 }
784 dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr;
785 }
786 return rip_output(m, so, dst);
787 }
788
789 /* note: rip_unlock is called from different protos instead of the generic socket_unlock,
790 * it will handle the socket dealloc on last reference
791 * */
792 int
793 rip_unlock(struct socket *so, int refcount, int debug)
794 {
795 int lr_saved;
796 struct inpcb *inp = sotoinpcb(so);
797
798 if (debug == 0)
799 lr_saved = (unsigned int) __builtin_return_address(0);
800 else lr_saved = debug;
801
802 if (refcount) {
803 if (so->so_usecount <= 0)
804 panic("rip_unlock: bad refoucnt so=%p val=%x\n", so, so->so_usecount);
805 so->so_usecount--;
806 if (so->so_usecount == 0 && (inp->inp_wantcnt == WNT_STOPUSING)) {
807 /* cleanup after last reference */
808 lck_mtx_unlock(so->so_proto->pr_domain->dom_mtx);
809 lck_rw_lock_exclusive(ripcbinfo.mtx);
810 in_pcbdispose(inp);
811 lck_rw_done(ripcbinfo.mtx);
812 return(0);
813 }
814 }
815 so->unlock_lr[so->next_unlock_lr] = (u_int32_t)lr_saved;
816 so->next_unlock_lr = (so->next_unlock_lr+1) % SO_LCKDBG_MAX;
817 lck_mtx_unlock(so->so_proto->pr_domain->dom_mtx);
818 return(0);
819 }
820
821 static int
822 rip_pcblist SYSCTL_HANDLER_ARGS
823 {
824 #pragma unused(oidp, arg1, arg2)
825 int error, i, n;
826 struct inpcb *inp, **inp_list;
827 inp_gen_t gencnt;
828 struct xinpgen xig;
829
830 /*
831 * The process of preparing the TCB list is too time-consuming and
832 * resource-intensive to repeat twice on every request.
833 */
834 lck_rw_lock_exclusive(ripcbinfo.mtx);
835 if (req->oldptr == USER_ADDR_NULL) {
836 n = ripcbinfo.ipi_count;
837 req->oldidx = 2 * (sizeof xig)
838 + (n + n/8) * sizeof(struct xinpcb);
839 lck_rw_done(ripcbinfo.mtx);
840 return 0;
841 }
842
843 if (req->newptr != USER_ADDR_NULL) {
844 lck_rw_done(ripcbinfo.mtx);
845 return EPERM;
846 }
847
848 /*
849 * OK, now we're committed to doing something.
850 */
851 gencnt = ripcbinfo.ipi_gencnt;
852 n = ripcbinfo.ipi_count;
853
854 bzero(&xig, sizeof(xig));
855 xig.xig_len = sizeof xig;
856 xig.xig_count = n;
857 xig.xig_gen = gencnt;
858 xig.xig_sogen = so_gencnt;
859 error = SYSCTL_OUT(req, &xig, sizeof xig);
860 if (error) {
861 lck_rw_done(ripcbinfo.mtx);
862 return error;
863 }
864 /*
865 * We are done if there is no pcb
866 */
867 if (n == 0) {
868 lck_rw_done(ripcbinfo.mtx);
869 return 0;
870 }
871
872 inp_list = _MALLOC(n * sizeof *inp_list, M_TEMP, M_WAITOK);
873 if (inp_list == 0) {
874 lck_rw_done(ripcbinfo.mtx);
875 return ENOMEM;
876 }
877
878 for (inp = ripcbinfo.listhead->lh_first, i = 0; inp && i < n;
879 inp = inp->inp_list.le_next) {
880 if (inp->inp_gencnt <= gencnt && inp->inp_state != INPCB_STATE_DEAD)
881 inp_list[i++] = inp;
882 }
883 n = i;
884
885 error = 0;
886 for (i = 0; i < n; i++) {
887 inp = inp_list[i];
888 if (inp->inp_gencnt <= gencnt && inp->inp_state != INPCB_STATE_DEAD) {
889 struct xinpcb xi;
890
891 bzero(&xi, sizeof(xi));
892 xi.xi_len = sizeof xi;
893 /* XXX should avoid extra copy */
894 inpcb_to_compat(inp, &xi.xi_inp);
895 if (inp->inp_socket)
896 sotoxsocket(inp->inp_socket, &xi.xi_socket);
897 error = SYSCTL_OUT(req, &xi, sizeof xi);
898 }
899 }
900 if (!error) {
901 /*
902 * Give the user an updated idea of our state.
903 * If the generation differs from what we told
904 * her before, she knows that something happened
905 * while we were processing this request, and it
906 * might be necessary to retry.
907 */
908 bzero(&xig, sizeof(xig));
909 xig.xig_len = sizeof xig;
910 xig.xig_gen = ripcbinfo.ipi_gencnt;
911 xig.xig_sogen = so_gencnt;
912 xig.xig_count = ripcbinfo.ipi_count;
913 error = SYSCTL_OUT(req, &xig, sizeof xig);
914 }
915 FREE(inp_list, M_TEMP);
916 lck_rw_done(ripcbinfo.mtx);
917 return error;
918 }
919
920 SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, 0, 0,
921 rip_pcblist, "S,xinpcb", "List of active raw IP sockets");
922
923 struct pr_usrreqs rip_usrreqs = {
924 rip_abort, pru_accept_notsupp, rip_attach, rip_bind, rip_connect,
925 pru_connect2_notsupp, in_control, rip_detach, rip_disconnect,
926 pru_listen_notsupp, in_setpeeraddr, pru_rcvd_notsupp,
927 pru_rcvoob_notsupp, rip_send, pru_sense_null, rip_shutdown,
928 in_setsockaddr, sosend, soreceive, pru_sopoll_notsupp
929 };
930 /* DSEP Review Done pl-20051213-v02 @3253 */