1/*
2 * Copyright (c) 1982, 1986, 1988, 1993
3 * The Regents of the University of California. 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. All advertising materials mentioning features or use of this software
14 * must display the following acknowledgement:
15 * This product includes software developed by the University of
16 * California, Berkeley and its contributors.
17 * 4. Neither the name of the University nor the names of its contributors
18 * may be used to endorse or promote products derived from this software
19 * without specific prior written permission.
20 *
21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
31 * SUCH DAMAGE.
32 *
33 * @(#)raw_ip.c 8.7 (Berkeley) 5/15/95
34 * $FreeBSD: head/sys/netinet/raw_ip.c 62573 2000-07-04 11:25:35Z phk $
35 */
36
37#include "opt_inet6.h"
38#include "opt_ipsec.h"
39
40#include <sys/param.h>
41#include <sys/systm.h>
42#include <sys/kernel.h>
43#include <sys/malloc.h>
44#include <sys/mbuf.h>
45#include <sys/proc.h>
46#include <sys/protosw.h>
47#include <sys/socket.h>
48#include <sys/socketvar.h>
49#include <sys/sysctl.h>
50
51#include <vm/vm_zone.h>
52
53#include <net/if.h>
54#include <net/route.h>
55
56#define _IP_VHL
57#include <netinet/in.h>
58#include <netinet/in_systm.h>
59#include <netinet/ip.h>
60#include <netinet/in_pcb.h>
61#include <netinet/in_var.h>
62#include <netinet/ip_var.h>
63#include <netinet/ip_mroute.h>
64
65#include <netinet/ip_fw.h>
66
67#ifdef IPSEC
68#include <netinet6/ipsec.h>
69#endif /*IPSEC*/
70
71#include "opt_ipdn.h"
72#ifdef DUMMYNET
73#include <netinet/ip_dummynet.h>
74#endif
75
76struct inpcbhead ripcb;
77struct inpcbinfo ripcbinfo;
78
79/*
80 * Nominal space allocated to a raw ip socket.
81 */
82#define RIPSNDQ 8192
83#define RIPRCVQ 8192
84
85/*
86 * Raw interface to IP protocol.
87 */
88
89/*
90 * Initialize raw connection block q.
91 */
92void
93rip_init()
94{
95 LIST_INIT(&ripcb);
96 ripcbinfo.listhead = &ripcb;
97 /*
98 * XXX We don't use the hash list for raw IP, but it's easier
99 * to allocate a one entry hash list than it is to check all
100 * over the place for hashbase == NULL.
101 */
102 ripcbinfo.hashbase = hashinit(1, M_PCB, &ripcbinfo.hashmask);
103 ripcbinfo.porthashbase = hashinit(1, M_PCB, &ripcbinfo.porthashmask);
104 ripcbinfo.ipi_zone = zinit("ripcb", sizeof(struct inpcb),
105 maxsockets, ZONE_INTERRUPT, 0);
106}
107
108static struct sockaddr_in ripsrc = { sizeof(ripsrc), AF_INET };
109/*
110 * Setup generic address and protocol structures
111 * for raw_input routine, then pass them along with
112 * mbuf chain.
113 */
114void
115rip_input(m, off, proto)
116 struct mbuf *m;
117 int off, proto;
118{
119 register struct ip *ip = mtod(m, struct ip *);
120 register struct inpcb *inp;
121 struct inpcb *last = 0;
122 struct mbuf *opts = 0;
123
124 ripsrc.sin_addr = ip->ip_src;
125 LIST_FOREACH(inp, &ripcb, inp_list) {
126#ifdef INET6
127 if ((inp->inp_vflag & INP_IPV4) == 0)
128 continue;
129#endif
130 if (inp->inp_ip_p && inp->inp_ip_p != proto)
131 continue;
132 if (inp->inp_laddr.s_addr &&
133 inp->inp_laddr.s_addr != ip->ip_dst.s_addr)
134 continue;
135 if (inp->inp_faddr.s_addr &&
136 inp->inp_faddr.s_addr != ip->ip_src.s_addr)
137 continue;
138 if (last) {
139 struct mbuf *n = m_copy(m, 0, (int)M_COPYALL);
140 if (n) {
141 if (last->inp_flags & INP_CONTROLOPTS ||
142 last->inp_socket->so_options & SO_TIMESTAMP)
143 ip_savecontrol(last, &opts, ip, n);
144 if (sbappendaddr(&last->inp_socket->so_rcv,
145 (struct sockaddr *)&ripsrc, n,
146 opts) == 0) {
147 /* should notify about lost packet */
148 m_freem(n);
149 if (opts)
150 m_freem(opts);
151 } else
152 sorwakeup(last->inp_socket);
153 opts = 0;
154 }
155 }
156 last = inp;
157 }
158 if (last) {
159 if (last->inp_flags & INP_CONTROLOPTS ||
160 last->inp_socket->so_options & SO_TIMESTAMP)
161 ip_savecontrol(last, &opts, ip, m);
162 if (sbappendaddr(&last->inp_socket->so_rcv,
163 (struct sockaddr *)&ripsrc, m, opts) == 0) {
164 m_freem(m);
165 if (opts)
166 m_freem(opts);
167 } else
168 sorwakeup(last->inp_socket);
169 } else {
170 m_freem(m);
171 ipstat.ips_noproto++;
172 ipstat.ips_delivered--;
173 }
174}
175
176/*
177 * Generate IP header and pass packet to ip_output.
178 * Tack on options user may have setup with control call.
179 */
180int
181rip_output(m, so, dst)
182 struct mbuf *m;
183 struct socket *so;
184 u_long dst;
185{
186 register struct ip *ip;
187 register struct inpcb *inp = sotoinpcb(so);
188 int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST;
189
190 /*
191 * If the user handed us a complete IP packet, use it.
192 * Otherwise, allocate an mbuf for a header and fill it in.
193 */
194 if ((inp->inp_flags & INP_HDRINCL) == 0) {
195 if (m->m_pkthdr.len + sizeof(struct ip) > IP_MAXPACKET) {
196 m_freem(m);
197 return(EMSGSIZE);
198 }
199 M_PREPEND(m, sizeof(struct ip), M_WAIT);
200 ip = mtod(m, struct ip *);
201 ip->ip_tos = 0;
202 ip->ip_off = 0;
203 ip->ip_p = inp->inp_ip_p;
204 ip->ip_len = m->m_pkthdr.len;
205 ip->ip_src = inp->inp_laddr;
206 ip->ip_dst.s_addr = dst;
207 ip->ip_ttl = MAXTTL;
208 } else {
209 if (m->m_pkthdr.len > IP_MAXPACKET) {
210 m_freem(m);
211 return(EMSGSIZE);
212 }
213 ip = mtod(m, struct ip *);
214 /* don't allow both user specified and setsockopt options,
215 and don't allow packet length sizes that will crash */
216 if (((IP_VHL_HL(ip->ip_vhl) != (sizeof (*ip) >> 2))
217 && inp->inp_options)
218 || (ip->ip_len > m->m_pkthdr.len)
219 || (ip->ip_len < (IP_VHL_HL(ip->ip_vhl) << 2))) {
220 m_freem(m);
221 return EINVAL;
222 }
223 if (ip->ip_id == 0)
224 ip->ip_id = htons(ip_id++);
225 /* XXX prevent ip_output from overwriting header fields */
226 flags |= IP_RAWOUTPUT;
227 ipstat.ips_rawout++;
228 }
229
230#ifdef IPSEC
231 m->m_pkthdr.rcvif = (struct ifnet *)so; /*XXX*/
232#endif /*IPSEC*/
233
234 return (ip_output(m, inp->inp_options, &inp->inp_route,
235 flags | IP_SOCKINMRCVIF,
236 inp->inp_moptions));
237}
238
239/*
240 * Raw IP socket option processing.
241 */
242int
243rip_ctloutput(so, sopt)
244 struct socket *so;
245 struct sockopt *sopt;
246{
247 struct inpcb *inp = sotoinpcb(so);
248 int error, optval;
249
250 if (sopt->sopt_level != IPPROTO_IP)
251 return (EINVAL);
252
253 error = 0;
254
255 switch (sopt->sopt_dir) {
256 case SOPT_GET:
257 switch (sopt->sopt_name) {
258 case IP_HDRINCL:
259 optval = inp->inp_flags & INP_HDRINCL;
260 error = sooptcopyout(sopt, &optval, sizeof optval);
261 break;
262
263 case IP_FW_GET:
264 if (ip_fw_ctl_ptr == 0)
265 error = ENOPROTOOPT;
266 else
267 error = ip_fw_ctl_ptr(sopt);
268 break;
269
270#ifdef DUMMYNET
271 case IP_DUMMYNET_GET:
272 if (ip_dn_ctl_ptr == NULL)
273 error = ENOPROTOOPT ;
274 else
275 error = ip_dn_ctl_ptr(sopt);
276 break ;
277#endif /* DUMMYNET */
278
279 case MRT_INIT:
280 case MRT_DONE:
281 case MRT_ADD_VIF:
282 case MRT_DEL_VIF:
283 case MRT_ADD_MFC:
284 case MRT_DEL_MFC:
285 case MRT_VERSION:
286 case MRT_ASSERT:
287 error = ip_mrouter_get(so, sopt);
288 break;
289
290 default:
291 error = ip_ctloutput(so, sopt);
292 break;
293 }
294 break;
295
296 case SOPT_SET:
297 switch (sopt->sopt_name) {
298 case IP_HDRINCL:
299 error = sooptcopyin(sopt, &optval, sizeof optval,
300 sizeof optval);
301 if (error)
302 break;
303 if (optval)
304 inp->inp_flags |= INP_HDRINCL;
305 else
306 inp->inp_flags &= ~INP_HDRINCL;
307 break;
308
309 case IP_FW_ADD:
310 case IP_FW_DEL:
311 case IP_FW_FLUSH:
312 case IP_FW_ZERO:
313 case IP_FW_RESETLOG:
314 if (ip_fw_ctl_ptr == 0)
315 error = ENOPROTOOPT;
316 else
317 error = ip_fw_ctl_ptr(sopt);
318 break;
319
320#ifdef DUMMYNET
321 case IP_DUMMYNET_CONFIGURE:
322 case IP_DUMMYNET_DEL:
323 case IP_DUMMYNET_FLUSH:
324 if (ip_dn_ctl_ptr == NULL)
325 error = ENOPROTOOPT ;
326 else
327 error = ip_dn_ctl_ptr(sopt);
328 break ;
329#endif
330
331 case IP_RSVP_ON:
332 error = ip_rsvp_init(so);
333 break;
334
335 case IP_RSVP_OFF:
336 error = ip_rsvp_done();
337 break;
338
339 /* XXX - should be combined */
340 case IP_RSVP_VIF_ON:
341 error = ip_rsvp_vif_init(so, sopt);
342 break;
343
344 case IP_RSVP_VIF_OFF:
345 error = ip_rsvp_vif_done(so, sopt);
346 break;
347
348 case MRT_INIT:
349 case MRT_DONE:
350 case MRT_ADD_VIF:
351 case MRT_DEL_VIF:
352 case MRT_ADD_MFC:
353 case MRT_DEL_MFC:
354 case MRT_VERSION:
355 case MRT_ASSERT:
356 error = ip_mrouter_set(so, sopt);
357 break;
358
359 default:
360 error = ip_ctloutput(so, sopt);
361 break;
362 }
363 break;
364 }
365
366 return (error);
367}
368
369/*
370 * This function exists solely to receive the PRC_IFDOWN messages which
371 * are sent by if_down(). It looks for an ifaddr whose ifa_addr is sa,
372 * and calls in_ifadown() to remove all routes corresponding to that address.
373 * It also receives the PRC_IFUP messages from if_up() and reinstalls the
374 * interface routes.
375 */
376void
377rip_ctlinput(cmd, sa, vip)
378 int cmd;
379 struct sockaddr *sa;
380 void *vip;
381{
382 struct in_ifaddr *ia;
383 struct ifnet *ifp;
384 int err;
385 int flags;
386
387 switch (cmd) {
388 case PRC_IFDOWN:
389 for (ia = in_ifaddrhead.tqh_first; ia;
390 ia = ia->ia_link.tqe_next) {
391 if (ia->ia_ifa.ifa_addr == sa
392 && (ia->ia_flags & IFA_ROUTE)) {
393 /*
394 * in_ifscrub kills the interface route.
395 */
396 in_ifscrub(ia->ia_ifp, ia);
397 /*
398 * in_ifadown gets rid of all the rest of
399 * the routes. This is not quite the right
400 * thing to do, but at least if we are running
401 * a routing process they will come back.
402 */
403 in_ifadown(&ia->ia_ifa);
404 break;
405 }
406 }
407 break;
408
409 case PRC_IFUP:
410 for (ia = in_ifaddrhead.tqh_first; ia;
411 ia = ia->ia_link.tqe_next) {
412 if (ia->ia_ifa.ifa_addr == sa)
413 break;
414 }
415 if (ia == 0 || (ia->ia_flags & IFA_ROUTE))
416 return;
417 flags = RTF_UP;
418 ifp = ia->ia_ifa.ifa_ifp;
419
420 if ((ifp->if_flags & IFF_LOOPBACK)
421 || (ifp->if_flags & IFF_POINTOPOINT))
422 flags |= RTF_HOST;
423
424 err = rtinit(&ia->ia_ifa, RTM_ADD, flags);
425 if (err == 0)
426 ia->ia_flags |= IFA_ROUTE;
427 break;
428 }
429}
430
431u_long rip_sendspace = RIPSNDQ;
432u_long rip_recvspace = RIPRCVQ;
433
434SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW,
435 &rip_sendspace, 0, "Maximum outgoing raw IP datagram size");
436SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW,
437 &rip_recvspace, 0, "Maximum incoming raw IP datagram size");
438
439static int
440rip_attach(struct socket *so, int proto, struct proc *p)
441{
442 struct inpcb *inp;
443 int error, s;
444
445 inp = sotoinpcb(so);
446 if (inp)
447 panic("rip_attach");
448 if (p && (error = suser(p)) != 0)
449 return error;
450
451 error = soreserve(so, rip_sendspace, rip_recvspace);
452 if (error)
453 return error;
454 s = splnet();
455 error = in_pcballoc(so, &ripcbinfo, p);
456 splx(s);
457 if (error)
458 return error;
459 inp = (struct inpcb *)so->so_pcb;
460 inp->inp_vflag |= INP_IPV4;
461 inp->inp_ip_p = proto;
462#ifdef IPSEC
463 error = ipsec_init_policy(so, &inp->inp_sp);
464 if (error != 0) {
465 in_pcbdetach(inp);
466 return error;
467 }
468#endif /*IPSEC*/
469 return 0;
470}
471
472static int
473rip_detach(struct socket *so)
474{
475 struct inpcb *inp;
476
477 inp = sotoinpcb(so);
478 if (inp == 0)
479 panic("rip_detach");
480 if (so == ip_mrouter)
481 ip_mrouter_done();
482 ip_rsvp_force_done(so);
483 if (so == ip_rsvpd)
484 ip_rsvp_done();
485 in_pcbdetach(inp);
486 return 0;
487}
488
489static int
490rip_abort(struct socket *so)
491{
492 soisdisconnected(so);
493 return rip_detach(so);
494}
495
496static int
497rip_disconnect(struct socket *so)
498{
499 if ((so->so_state & SS_ISCONNECTED) == 0)
500 return ENOTCONN;
501 return rip_abort(so);
502}
503
504static int
505rip_bind(struct socket *so, struct sockaddr *nam, struct proc *p)
506{
507 struct inpcb *inp = sotoinpcb(so);
508 struct sockaddr_in *addr = (struct sockaddr_in *)nam;
509
510 if (nam->sa_len != sizeof(*addr))
511 return EINVAL;
512
513 if (TAILQ_EMPTY(&ifnet) || ((addr->sin_family != AF_INET) &&
514 (addr->sin_family != AF_IMPLINK)) ||
515 (addr->sin_addr.s_addr &&
516 ifa_ifwithaddr((struct sockaddr *)addr) == 0))
517 return EADDRNOTAVAIL;
518 inp->inp_laddr = addr->sin_addr;
519 return 0;
520}
521
522static int
523rip_connect(struct socket *so, struct sockaddr *nam, struct proc *p)
524{
525 struct inpcb *inp = sotoinpcb(so);
526 struct sockaddr_in *addr = (struct sockaddr_in *)nam;
527
528 if (nam->sa_len != sizeof(*addr))
529 return EINVAL;
530 if (TAILQ_EMPTY(&ifnet))
531 return EADDRNOTAVAIL;
532 if ((addr->sin_family != AF_INET) &&
533 (addr->sin_family != AF_IMPLINK))
534 return EAFNOSUPPORT;
535 inp->inp_faddr = addr->sin_addr;
536 soisconnected(so);
537 return 0;
538}
539
540static int
541rip_shutdown(struct socket *so)
542{
543 socantsendmore(so);
544 return 0;
545}
546
547static int
548rip_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam,
549 struct mbuf *control, struct proc *p)
550{
551 struct inpcb *inp = sotoinpcb(so);
552 register u_long dst;
553
554 if (so->so_state & SS_ISCONNECTED) {
555 if (nam) {
556 m_freem(m);
557 return EISCONN;
558 }
559 dst = inp->inp_faddr.s_addr;
560 } else {
561 if (nam == NULL) {
562 m_freem(m);
563 return ENOTCONN;
564 }
565 dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr;
566 }
567 return rip_output(m, so, dst);
568}
569
570static int
571rip_pcblist(SYSCTL_HANDLER_ARGS)
572{
573 int error, i, n, s;
574 struct inpcb *inp, **inp_list;
575 inp_gen_t gencnt;
576 struct xinpgen xig;
577
578 /*
579 * The process of preparing the TCB list is too time-consuming and
580 * resource-intensive to repeat twice on every request.
581 */
582 if (req->oldptr == 0) {
583 n = ripcbinfo.ipi_count;
584 req->oldidx = 2 * (sizeof xig)
585 + (n + n/8) * sizeof(struct xinpcb);
586 return 0;
587 }
588
589 if (req->newptr != 0)
590 return EPERM;
591
592 /*
593 * OK, now we're committed to doing something.
594 */
595 s = splnet();
596 gencnt = ripcbinfo.ipi_gencnt;
597 n = ripcbinfo.ipi_count;
598 splx(s);
599
600 xig.xig_len = sizeof xig;
601 xig.xig_count = n;
602 xig.xig_gen = gencnt;
603 xig.xig_sogen = so_gencnt;
604 error = SYSCTL_OUT(req, &xig, sizeof xig);
605 if (error)
606 return error;
607
608 inp_list = malloc(n * sizeof *inp_list, M_TEMP, M_WAITOK);
609 if (inp_list == 0)
610 return ENOMEM;
611
612 s = splnet();
613 for (inp = ripcbinfo.listhead->lh_first, i = 0; inp && i < n;
614 inp = inp->inp_list.le_next) {
615 if (inp->inp_gencnt <= gencnt)
616 inp_list[i++] = inp;
617 }
618 splx(s);
619 n = i;
620
621 error = 0;
622 for (i = 0; i < n; i++) {
623 inp = inp_list[i];
624 if (inp->inp_gencnt <= gencnt) {
625 struct xinpcb xi;
626 xi.xi_len = sizeof xi;
627 /* XXX should avoid extra copy */
628 bcopy(inp, &xi.xi_inp, sizeof *inp);
629 if (inp->inp_socket)
630 sotoxsocket(inp->inp_socket, &xi.xi_socket);
631 error = SYSCTL_OUT(req, &xi, sizeof xi);
632 }
633 }
634 if (!error) {
635 /*
636 * Give the user an updated idea of our state.
637 * If the generation differs from what we told
638 * her before, she knows that something happened
639 * while we were processing this request, and it
640 * might be necessary to retry.
641 */
642 s = splnet();
643 xig.xig_gen = ripcbinfo.ipi_gencnt;
644 xig.xig_sogen = so_gencnt;
645 xig.xig_count = ripcbinfo.ipi_count;
646 splx(s);
647 error = SYSCTL_OUT(req, &xig, sizeof xig);
648 }
649 free(inp_list, M_TEMP);
650 return error;
651}
652
653SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, 0, 0,
654 rip_pcblist, "S,xinpcb", "List of active raw IP sockets");
655
656struct pr_usrreqs rip_usrreqs = {
657 rip_abort, pru_accept_notsupp, rip_attach, rip_bind, rip_connect,
658 pru_connect2_notsupp, in_control, rip_detach, rip_disconnect,
659 pru_listen_notsupp, in_setpeeraddr, pru_rcvd_notsupp,
660 pru_rcvoob_notsupp, rip_send, pru_sense_null, rip_shutdown,
661 in_setsockaddr, sosend, soreceive, sopoll
662};
2 * Copyright (c) 1982, 1986, 1988, 1993
3 * The Regents of the University of California. 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. All advertising materials mentioning features or use of this software
14 * must display the following acknowledgement:
15 * This product includes software developed by the University of
16 * California, Berkeley and its contributors.
17 * 4. Neither the name of the University nor the names of its contributors
18 * may be used to endorse or promote products derived from this software
19 * without specific prior written permission.
20 *
21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
31 * SUCH DAMAGE.
32 *
33 * @(#)raw_ip.c 8.7 (Berkeley) 5/15/95
34 * $FreeBSD: head/sys/netinet/raw_ip.c 62573 2000-07-04 11:25:35Z phk $
35 */
36
37#include "opt_inet6.h"
38#include "opt_ipsec.h"
39
40#include <sys/param.h>
41#include <sys/systm.h>
42#include <sys/kernel.h>
43#include <sys/malloc.h>
44#include <sys/mbuf.h>
45#include <sys/proc.h>
46#include <sys/protosw.h>
47#include <sys/socket.h>
48#include <sys/socketvar.h>
49#include <sys/sysctl.h>
50
51#include <vm/vm_zone.h>
52
53#include <net/if.h>
54#include <net/route.h>
55
56#define _IP_VHL
57#include <netinet/in.h>
58#include <netinet/in_systm.h>
59#include <netinet/ip.h>
60#include <netinet/in_pcb.h>
61#include <netinet/in_var.h>
62#include <netinet/ip_var.h>
63#include <netinet/ip_mroute.h>
64
65#include <netinet/ip_fw.h>
66
67#ifdef IPSEC
68#include <netinet6/ipsec.h>
69#endif /*IPSEC*/
70
71#include "opt_ipdn.h"
72#ifdef DUMMYNET
73#include <netinet/ip_dummynet.h>
74#endif
75
76struct inpcbhead ripcb;
77struct inpcbinfo ripcbinfo;
78
79/*
80 * Nominal space allocated to a raw ip socket.
81 */
82#define RIPSNDQ 8192
83#define RIPRCVQ 8192
84
85/*
86 * Raw interface to IP protocol.
87 */
88
89/*
90 * Initialize raw connection block q.
91 */
92void
93rip_init()
94{
95 LIST_INIT(&ripcb);
96 ripcbinfo.listhead = &ripcb;
97 /*
98 * XXX We don't use the hash list for raw IP, but it's easier
99 * to allocate a one entry hash list than it is to check all
100 * over the place for hashbase == NULL.
101 */
102 ripcbinfo.hashbase = hashinit(1, M_PCB, &ripcbinfo.hashmask);
103 ripcbinfo.porthashbase = hashinit(1, M_PCB, &ripcbinfo.porthashmask);
104 ripcbinfo.ipi_zone = zinit("ripcb", sizeof(struct inpcb),
105 maxsockets, ZONE_INTERRUPT, 0);
106}
107
108static struct sockaddr_in ripsrc = { sizeof(ripsrc), AF_INET };
109/*
110 * Setup generic address and protocol structures
111 * for raw_input routine, then pass them along with
112 * mbuf chain.
113 */
114void
115rip_input(m, off, proto)
116 struct mbuf *m;
117 int off, proto;
118{
119 register struct ip *ip = mtod(m, struct ip *);
120 register struct inpcb *inp;
121 struct inpcb *last = 0;
122 struct mbuf *opts = 0;
123
124 ripsrc.sin_addr = ip->ip_src;
125 LIST_FOREACH(inp, &ripcb, inp_list) {
126#ifdef INET6
127 if ((inp->inp_vflag & INP_IPV4) == 0)
128 continue;
129#endif
130 if (inp->inp_ip_p && inp->inp_ip_p != proto)
131 continue;
132 if (inp->inp_laddr.s_addr &&
133 inp->inp_laddr.s_addr != ip->ip_dst.s_addr)
134 continue;
135 if (inp->inp_faddr.s_addr &&
136 inp->inp_faddr.s_addr != ip->ip_src.s_addr)
137 continue;
138 if (last) {
139 struct mbuf *n = m_copy(m, 0, (int)M_COPYALL);
140 if (n) {
141 if (last->inp_flags & INP_CONTROLOPTS ||
142 last->inp_socket->so_options & SO_TIMESTAMP)
143 ip_savecontrol(last, &opts, ip, n);
144 if (sbappendaddr(&last->inp_socket->so_rcv,
145 (struct sockaddr *)&ripsrc, n,
146 opts) == 0) {
147 /* should notify about lost packet */
148 m_freem(n);
149 if (opts)
150 m_freem(opts);
151 } else
152 sorwakeup(last->inp_socket);
153 opts = 0;
154 }
155 }
156 last = inp;
157 }
158 if (last) {
159 if (last->inp_flags & INP_CONTROLOPTS ||
160 last->inp_socket->so_options & SO_TIMESTAMP)
161 ip_savecontrol(last, &opts, ip, m);
162 if (sbappendaddr(&last->inp_socket->so_rcv,
163 (struct sockaddr *)&ripsrc, m, opts) == 0) {
164 m_freem(m);
165 if (opts)
166 m_freem(opts);
167 } else
168 sorwakeup(last->inp_socket);
169 } else {
170 m_freem(m);
171 ipstat.ips_noproto++;
172 ipstat.ips_delivered--;
173 }
174}
175
176/*
177 * Generate IP header and pass packet to ip_output.
178 * Tack on options user may have setup with control call.
179 */
180int
181rip_output(m, so, dst)
182 struct mbuf *m;
183 struct socket *so;
184 u_long dst;
185{
186 register struct ip *ip;
187 register struct inpcb *inp = sotoinpcb(so);
188 int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST;
189
190 /*
191 * If the user handed us a complete IP packet, use it.
192 * Otherwise, allocate an mbuf for a header and fill it in.
193 */
194 if ((inp->inp_flags & INP_HDRINCL) == 0) {
195 if (m->m_pkthdr.len + sizeof(struct ip) > IP_MAXPACKET) {
196 m_freem(m);
197 return(EMSGSIZE);
198 }
199 M_PREPEND(m, sizeof(struct ip), M_WAIT);
200 ip = mtod(m, struct ip *);
201 ip->ip_tos = 0;
202 ip->ip_off = 0;
203 ip->ip_p = inp->inp_ip_p;
204 ip->ip_len = m->m_pkthdr.len;
205 ip->ip_src = inp->inp_laddr;
206 ip->ip_dst.s_addr = dst;
207 ip->ip_ttl = MAXTTL;
208 } else {
209 if (m->m_pkthdr.len > IP_MAXPACKET) {
210 m_freem(m);
211 return(EMSGSIZE);
212 }
213 ip = mtod(m, struct ip *);
214 /* don't allow both user specified and setsockopt options,
215 and don't allow packet length sizes that will crash */
216 if (((IP_VHL_HL(ip->ip_vhl) != (sizeof (*ip) >> 2))
217 && inp->inp_options)
218 || (ip->ip_len > m->m_pkthdr.len)
219 || (ip->ip_len < (IP_VHL_HL(ip->ip_vhl) << 2))) {
220 m_freem(m);
221 return EINVAL;
222 }
223 if (ip->ip_id == 0)
224 ip->ip_id = htons(ip_id++);
225 /* XXX prevent ip_output from overwriting header fields */
226 flags |= IP_RAWOUTPUT;
227 ipstat.ips_rawout++;
228 }
229
230#ifdef IPSEC
231 m->m_pkthdr.rcvif = (struct ifnet *)so; /*XXX*/
232#endif /*IPSEC*/
233
234 return (ip_output(m, inp->inp_options, &inp->inp_route,
235 flags | IP_SOCKINMRCVIF,
236 inp->inp_moptions));
237}
238
239/*
240 * Raw IP socket option processing.
241 */
242int
243rip_ctloutput(so, sopt)
244 struct socket *so;
245 struct sockopt *sopt;
246{
247 struct inpcb *inp = sotoinpcb(so);
248 int error, optval;
249
250 if (sopt->sopt_level != IPPROTO_IP)
251 return (EINVAL);
252
253 error = 0;
254
255 switch (sopt->sopt_dir) {
256 case SOPT_GET:
257 switch (sopt->sopt_name) {
258 case IP_HDRINCL:
259 optval = inp->inp_flags & INP_HDRINCL;
260 error = sooptcopyout(sopt, &optval, sizeof optval);
261 break;
262
263 case IP_FW_GET:
264 if (ip_fw_ctl_ptr == 0)
265 error = ENOPROTOOPT;
266 else
267 error = ip_fw_ctl_ptr(sopt);
268 break;
269
270#ifdef DUMMYNET
271 case IP_DUMMYNET_GET:
272 if (ip_dn_ctl_ptr == NULL)
273 error = ENOPROTOOPT ;
274 else
275 error = ip_dn_ctl_ptr(sopt);
276 break ;
277#endif /* DUMMYNET */
278
279 case MRT_INIT:
280 case MRT_DONE:
281 case MRT_ADD_VIF:
282 case MRT_DEL_VIF:
283 case MRT_ADD_MFC:
284 case MRT_DEL_MFC:
285 case MRT_VERSION:
286 case MRT_ASSERT:
287 error = ip_mrouter_get(so, sopt);
288 break;
289
290 default:
291 error = ip_ctloutput(so, sopt);
292 break;
293 }
294 break;
295
296 case SOPT_SET:
297 switch (sopt->sopt_name) {
298 case IP_HDRINCL:
299 error = sooptcopyin(sopt, &optval, sizeof optval,
300 sizeof optval);
301 if (error)
302 break;
303 if (optval)
304 inp->inp_flags |= INP_HDRINCL;
305 else
306 inp->inp_flags &= ~INP_HDRINCL;
307 break;
308
309 case IP_FW_ADD:
310 case IP_FW_DEL:
311 case IP_FW_FLUSH:
312 case IP_FW_ZERO:
313 case IP_FW_RESETLOG:
314 if (ip_fw_ctl_ptr == 0)
315 error = ENOPROTOOPT;
316 else
317 error = ip_fw_ctl_ptr(sopt);
318 break;
319
320#ifdef DUMMYNET
321 case IP_DUMMYNET_CONFIGURE:
322 case IP_DUMMYNET_DEL:
323 case IP_DUMMYNET_FLUSH:
324 if (ip_dn_ctl_ptr == NULL)
325 error = ENOPROTOOPT ;
326 else
327 error = ip_dn_ctl_ptr(sopt);
328 break ;
329#endif
330
331 case IP_RSVP_ON:
332 error = ip_rsvp_init(so);
333 break;
334
335 case IP_RSVP_OFF:
336 error = ip_rsvp_done();
337 break;
338
339 /* XXX - should be combined */
340 case IP_RSVP_VIF_ON:
341 error = ip_rsvp_vif_init(so, sopt);
342 break;
343
344 case IP_RSVP_VIF_OFF:
345 error = ip_rsvp_vif_done(so, sopt);
346 break;
347
348 case MRT_INIT:
349 case MRT_DONE:
350 case MRT_ADD_VIF:
351 case MRT_DEL_VIF:
352 case MRT_ADD_MFC:
353 case MRT_DEL_MFC:
354 case MRT_VERSION:
355 case MRT_ASSERT:
356 error = ip_mrouter_set(so, sopt);
357 break;
358
359 default:
360 error = ip_ctloutput(so, sopt);
361 break;
362 }
363 break;
364 }
365
366 return (error);
367}
368
369/*
370 * This function exists solely to receive the PRC_IFDOWN messages which
371 * are sent by if_down(). It looks for an ifaddr whose ifa_addr is sa,
372 * and calls in_ifadown() to remove all routes corresponding to that address.
373 * It also receives the PRC_IFUP messages from if_up() and reinstalls the
374 * interface routes.
375 */
376void
377rip_ctlinput(cmd, sa, vip)
378 int cmd;
379 struct sockaddr *sa;
380 void *vip;
381{
382 struct in_ifaddr *ia;
383 struct ifnet *ifp;
384 int err;
385 int flags;
386
387 switch (cmd) {
388 case PRC_IFDOWN:
389 for (ia = in_ifaddrhead.tqh_first; ia;
390 ia = ia->ia_link.tqe_next) {
391 if (ia->ia_ifa.ifa_addr == sa
392 && (ia->ia_flags & IFA_ROUTE)) {
393 /*
394 * in_ifscrub kills the interface route.
395 */
396 in_ifscrub(ia->ia_ifp, ia);
397 /*
398 * in_ifadown gets rid of all the rest of
399 * the routes. This is not quite the right
400 * thing to do, but at least if we are running
401 * a routing process they will come back.
402 */
403 in_ifadown(&ia->ia_ifa);
404 break;
405 }
406 }
407 break;
408
409 case PRC_IFUP:
410 for (ia = in_ifaddrhead.tqh_first; ia;
411 ia = ia->ia_link.tqe_next) {
412 if (ia->ia_ifa.ifa_addr == sa)
413 break;
414 }
415 if (ia == 0 || (ia->ia_flags & IFA_ROUTE))
416 return;
417 flags = RTF_UP;
418 ifp = ia->ia_ifa.ifa_ifp;
419
420 if ((ifp->if_flags & IFF_LOOPBACK)
421 || (ifp->if_flags & IFF_POINTOPOINT))
422 flags |= RTF_HOST;
423
424 err = rtinit(&ia->ia_ifa, RTM_ADD, flags);
425 if (err == 0)
426 ia->ia_flags |= IFA_ROUTE;
427 break;
428 }
429}
430
431u_long rip_sendspace = RIPSNDQ;
432u_long rip_recvspace = RIPRCVQ;
433
434SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW,
435 &rip_sendspace, 0, "Maximum outgoing raw IP datagram size");
436SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW,
437 &rip_recvspace, 0, "Maximum incoming raw IP datagram size");
438
439static int
440rip_attach(struct socket *so, int proto, struct proc *p)
441{
442 struct inpcb *inp;
443 int error, s;
444
445 inp = sotoinpcb(so);
446 if (inp)
447 panic("rip_attach");
448 if (p && (error = suser(p)) != 0)
449 return error;
450
451 error = soreserve(so, rip_sendspace, rip_recvspace);
452 if (error)
453 return error;
454 s = splnet();
455 error = in_pcballoc(so, &ripcbinfo, p);
456 splx(s);
457 if (error)
458 return error;
459 inp = (struct inpcb *)so->so_pcb;
460 inp->inp_vflag |= INP_IPV4;
461 inp->inp_ip_p = proto;
462#ifdef IPSEC
463 error = ipsec_init_policy(so, &inp->inp_sp);
464 if (error != 0) {
465 in_pcbdetach(inp);
466 return error;
467 }
468#endif /*IPSEC*/
469 return 0;
470}
471
472static int
473rip_detach(struct socket *so)
474{
475 struct inpcb *inp;
476
477 inp = sotoinpcb(so);
478 if (inp == 0)
479 panic("rip_detach");
480 if (so == ip_mrouter)
481 ip_mrouter_done();
482 ip_rsvp_force_done(so);
483 if (so == ip_rsvpd)
484 ip_rsvp_done();
485 in_pcbdetach(inp);
486 return 0;
487}
488
489static int
490rip_abort(struct socket *so)
491{
492 soisdisconnected(so);
493 return rip_detach(so);
494}
495
496static int
497rip_disconnect(struct socket *so)
498{
499 if ((so->so_state & SS_ISCONNECTED) == 0)
500 return ENOTCONN;
501 return rip_abort(so);
502}
503
504static int
505rip_bind(struct socket *so, struct sockaddr *nam, struct proc *p)
506{
507 struct inpcb *inp = sotoinpcb(so);
508 struct sockaddr_in *addr = (struct sockaddr_in *)nam;
509
510 if (nam->sa_len != sizeof(*addr))
511 return EINVAL;
512
513 if (TAILQ_EMPTY(&ifnet) || ((addr->sin_family != AF_INET) &&
514 (addr->sin_family != AF_IMPLINK)) ||
515 (addr->sin_addr.s_addr &&
516 ifa_ifwithaddr((struct sockaddr *)addr) == 0))
517 return EADDRNOTAVAIL;
518 inp->inp_laddr = addr->sin_addr;
519 return 0;
520}
521
522static int
523rip_connect(struct socket *so, struct sockaddr *nam, struct proc *p)
524{
525 struct inpcb *inp = sotoinpcb(so);
526 struct sockaddr_in *addr = (struct sockaddr_in *)nam;
527
528 if (nam->sa_len != sizeof(*addr))
529 return EINVAL;
530 if (TAILQ_EMPTY(&ifnet))
531 return EADDRNOTAVAIL;
532 if ((addr->sin_family != AF_INET) &&
533 (addr->sin_family != AF_IMPLINK))
534 return EAFNOSUPPORT;
535 inp->inp_faddr = addr->sin_addr;
536 soisconnected(so);
537 return 0;
538}
539
540static int
541rip_shutdown(struct socket *so)
542{
543 socantsendmore(so);
544 return 0;
545}
546
547static int
548rip_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam,
549 struct mbuf *control, struct proc *p)
550{
551 struct inpcb *inp = sotoinpcb(so);
552 register u_long dst;
553
554 if (so->so_state & SS_ISCONNECTED) {
555 if (nam) {
556 m_freem(m);
557 return EISCONN;
558 }
559 dst = inp->inp_faddr.s_addr;
560 } else {
561 if (nam == NULL) {
562 m_freem(m);
563 return ENOTCONN;
564 }
565 dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr;
566 }
567 return rip_output(m, so, dst);
568}
569
570static int
571rip_pcblist(SYSCTL_HANDLER_ARGS)
572{
573 int error, i, n, s;
574 struct inpcb *inp, **inp_list;
575 inp_gen_t gencnt;
576 struct xinpgen xig;
577
578 /*
579 * The process of preparing the TCB list is too time-consuming and
580 * resource-intensive to repeat twice on every request.
581 */
582 if (req->oldptr == 0) {
583 n = ripcbinfo.ipi_count;
584 req->oldidx = 2 * (sizeof xig)
585 + (n + n/8) * sizeof(struct xinpcb);
586 return 0;
587 }
588
589 if (req->newptr != 0)
590 return EPERM;
591
592 /*
593 * OK, now we're committed to doing something.
594 */
595 s = splnet();
596 gencnt = ripcbinfo.ipi_gencnt;
597 n = ripcbinfo.ipi_count;
598 splx(s);
599
600 xig.xig_len = sizeof xig;
601 xig.xig_count = n;
602 xig.xig_gen = gencnt;
603 xig.xig_sogen = so_gencnt;
604 error = SYSCTL_OUT(req, &xig, sizeof xig);
605 if (error)
606 return error;
607
608 inp_list = malloc(n * sizeof *inp_list, M_TEMP, M_WAITOK);
609 if (inp_list == 0)
610 return ENOMEM;
611
612 s = splnet();
613 for (inp = ripcbinfo.listhead->lh_first, i = 0; inp && i < n;
614 inp = inp->inp_list.le_next) {
615 if (inp->inp_gencnt <= gencnt)
616 inp_list[i++] = inp;
617 }
618 splx(s);
619 n = i;
620
621 error = 0;
622 for (i = 0; i < n; i++) {
623 inp = inp_list[i];
624 if (inp->inp_gencnt <= gencnt) {
625 struct xinpcb xi;
626 xi.xi_len = sizeof xi;
627 /* XXX should avoid extra copy */
628 bcopy(inp, &xi.xi_inp, sizeof *inp);
629 if (inp->inp_socket)
630 sotoxsocket(inp->inp_socket, &xi.xi_socket);
631 error = SYSCTL_OUT(req, &xi, sizeof xi);
632 }
633 }
634 if (!error) {
635 /*
636 * Give the user an updated idea of our state.
637 * If the generation differs from what we told
638 * her before, she knows that something happened
639 * while we were processing this request, and it
640 * might be necessary to retry.
641 */
642 s = splnet();
643 xig.xig_gen = ripcbinfo.ipi_gencnt;
644 xig.xig_sogen = so_gencnt;
645 xig.xig_count = ripcbinfo.ipi_count;
646 splx(s);
647 error = SYSCTL_OUT(req, &xig, sizeof xig);
648 }
649 free(inp_list, M_TEMP);
650 return error;
651}
652
653SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, 0, 0,
654 rip_pcblist, "S,xinpcb", "List of active raw IP sockets");
655
656struct pr_usrreqs rip_usrreqs = {
657 rip_abort, pru_accept_notsupp, rip_attach, rip_bind, rip_connect,
658 pru_connect2_notsupp, in_control, rip_detach, rip_disconnect,
659 pru_listen_notsupp, in_setpeeraddr, pru_rcvd_notsupp,
660 pru_rcvoob_notsupp, rip_send, pru_sense_null, rip_shutdown,
661 in_setsockaddr, sosend, soreceive, sopoll
662};