xref: /netbsd-src/sys/netinet/raw_ip.c (revision deb6f0161a9109e7de9b519dc8dfb9478668dcdd)
1 /*	$NetBSD: raw_ip.c,v 1.178 2018/09/14 05:09:51 maxv Exp $	*/
2 
3 /*
4  * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
5  * All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions
9  * are met:
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice, this list of conditions and the following disclaimer in the
14  *    documentation and/or other materials provided with the distribution.
15  * 3. Neither the name of the project nor the names of its contributors
16  *    may be used to endorse or promote products derived from this software
17  *    without specific prior written permission.
18  *
19  * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
20  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22  * ARE DISCLAIMED.  IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
23  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29  * SUCH DAMAGE.
30  */
31 
32 /*
33  * Copyright (c) 1982, 1986, 1988, 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  * 3. 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 /*
64  * Raw interface to IP protocol.
65  */
66 
67 #include <sys/cdefs.h>
68 __KERNEL_RCSID(0, "$NetBSD: raw_ip.c,v 1.178 2018/09/14 05:09:51 maxv Exp $");
69 
70 #ifdef _KERNEL_OPT
71 #include "opt_inet.h"
72 #include "opt_ipsec.h"
73 #include "opt_mrouting.h"
74 #include "opt_net_mpsafe.h"
75 #endif
76 
77 #include <sys/param.h>
78 #include <sys/sysctl.h>
79 #include <sys/mbuf.h>
80 #include <sys/socket.h>
81 #include <sys/protosw.h>
82 #include <sys/socketvar.h>
83 #include <sys/errno.h>
84 #include <sys/systm.h>
85 #include <sys/proc.h>
86 #include <sys/kauth.h>
87 
88 #include <net/if.h>
89 
90 #include <netinet/in.h>
91 #include <netinet/in_systm.h>
92 #include <netinet/ip.h>
93 #include <netinet/ip_var.h>
94 #include <netinet/ip_private.h>
95 #include <netinet/ip_mroute.h>
96 #include <netinet/ip_icmp.h>
97 #include <netinet/in_pcb.h>
98 #include <netinet/in_proto.h>
99 #include <netinet/in_var.h>
100 
101 #ifdef IPSEC
102 #include <netipsec/ipsec.h>
103 #endif
104 
105 struct inpcbtable rawcbtable;
106 
107 int	 rip_pcbnotify(struct inpcbtable *, struct in_addr,
108     struct in_addr, int, int, void (*)(struct inpcb *, int));
109 static int	 rip_connect_pcb(struct inpcb *, struct sockaddr_in *);
110 static void	 rip_disconnect1(struct inpcb *);
111 
112 static void sysctl_net_inet_raw_setup(struct sysctllog **);
113 
114 /*
115  * Nominal space allocated to a raw ip socket.
116  */
117 #define	RIPSNDQ		8192
118 #define	RIPRCVQ		8192
119 
120 static u_long		rip_sendspace = RIPSNDQ;
121 static u_long		rip_recvspace = RIPRCVQ;
122 
123 /*
124  * Raw interface to IP protocol.
125  */
126 
127 /*
128  * Initialize raw connection block q.
129  */
130 void
131 rip_init(void)
132 {
133 
134 	sysctl_net_inet_raw_setup(NULL);
135 	in_pcbinit(&rawcbtable, 1, 1);
136 }
137 
138 static void
139 rip_sbappendaddr(struct inpcb *last, struct ip *ip, const struct sockaddr *sa,
140     int hlen, struct mbuf *n)
141 {
142 	struct mbuf *opts = NULL;
143 
144 	if (last->inp_flags & INP_NOHEADER)
145 		m_adj(n, hlen);
146 	if (last->inp_flags & INP_CONTROLOPTS ||
147 	    SOOPT_TIMESTAMP(last->inp_socket->so_options))
148 		ip_savecontrol(last, &opts, ip, n);
149 	if (sbappendaddr(&last->inp_socket->so_rcv, sa, n, opts) == 0) {
150 		soroverflow(last->inp_socket);
151 		m_freem(n);
152 		if (opts)
153 			m_freem(opts);
154 	} else {
155 		sorwakeup(last->inp_socket);
156 	}
157 }
158 
159 /*
160  * Setup generic address and protocol structures
161  * for raw_input routine, then pass them along with
162  * mbuf chain.
163  */
164 void
165 rip_input(struct mbuf *m, int off, int proto)
166 {
167 	struct ip *ip = mtod(m, struct ip *);
168 	struct inpcb_hdr *inph;
169 	struct inpcb *inp;
170 	struct inpcb *last = NULL;
171 	struct mbuf *n;
172 	struct sockaddr_in ripsrc;
173 	int hlen;
174 
175 	sockaddr_in_init(&ripsrc, &ip->ip_src, 0);
176 
177 	/*
178 	 * XXX Compatibility: programs using raw IP expect ip_len
179 	 * XXX to have the header length subtracted, and in host order.
180 	 * XXX ip_off is also expected to be host order.
181 	 */
182 	hlen = ip->ip_hl << 2;
183 	ip->ip_len = ntohs(ip->ip_len) - hlen;
184 	NTOHS(ip->ip_off);
185 
186 	TAILQ_FOREACH(inph, &rawcbtable.inpt_queue, inph_queue) {
187 		inp = (struct inpcb *)inph;
188 		if (inp->inp_af != AF_INET)
189 			continue;
190 		if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto)
191 			continue;
192 		if (!in_nullhost(inp->inp_laddr) &&
193 		    !in_hosteq(inp->inp_laddr, ip->ip_dst))
194 			continue;
195 		if (!in_nullhost(inp->inp_faddr) &&
196 		    !in_hosteq(inp->inp_faddr, ip->ip_src))
197 			continue;
198 
199 		if (last == NULL) {
200 			;
201 		}
202 #if defined(IPSEC)
203 		else if (ipsec_used && ipsec_in_reject(m, last)) {
204 			/* do not inject data into pcb */
205 		}
206 #endif
207 		else if ((n = m_copypacket(m, M_DONTWAIT)) != NULL) {
208 			rip_sbappendaddr(last, ip, sintosa(&ripsrc), hlen, n);
209 		}
210 
211 		last = inp;
212 	}
213 
214 #if defined(IPSEC)
215 	if (ipsec_used && last != NULL && ipsec_in_reject(m, last)) {
216 		m_freem(m);
217 		IP_STATDEC(IP_STAT_DELIVERED);
218 		/* do not inject data into pcb */
219 	} else
220 #endif
221 	if (last != NULL) {
222 		rip_sbappendaddr(last, ip, sintosa(&ripsrc), hlen, m);
223 	} else if (inetsw[ip_protox[ip->ip_p]].pr_input == rip_input) {
224 		uint64_t *ips;
225 
226 		icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_PROTOCOL,
227 		    0, 0);
228 		ips = IP_STAT_GETREF();
229 		ips[IP_STAT_NOPROTO]++;
230 		ips[IP_STAT_DELIVERED]--;
231 		IP_STAT_PUTREF();
232 	} else {
233 		m_freem(m);
234 	}
235 
236 	return;
237 }
238 
239 int
240 rip_pcbnotify(struct inpcbtable *table,
241     struct in_addr faddr, struct in_addr laddr, int proto, int errno,
242     void (*notify)(struct inpcb *, int))
243 {
244 	struct inpcb_hdr *inph, *ninph;
245 	int nmatch;
246 
247 	nmatch = 0;
248 	TAILQ_FOREACH_SAFE(inph, &table->inpt_queue, inph_queue, ninph) {
249 		struct inpcb *inp = (struct inpcb *)inph;
250 		if (inp->inp_af != AF_INET)
251 			continue;
252 		if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto)
253 			continue;
254 		if (in_hosteq(inp->inp_faddr, faddr) &&
255 		    in_hosteq(inp->inp_laddr, laddr)) {
256 			(*notify)(inp, errno);
257 			nmatch++;
258 		}
259 	}
260 
261 	return nmatch;
262 }
263 
264 void *
265 rip_ctlinput(int cmd, const struct sockaddr *sa, void *v)
266 {
267 	struct ip *ip = v;
268 	void (*notify)(struct inpcb *, int) = in_rtchange;
269 	int errno;
270 
271 	if (sa->sa_family != AF_INET ||
272 	    sa->sa_len != sizeof(struct sockaddr_in))
273 		return NULL;
274 	if ((unsigned)cmd >= PRC_NCMDS)
275 		return NULL;
276 	errno = inetctlerrmap[cmd];
277 	if (PRC_IS_REDIRECT(cmd))
278 		notify = in_rtchange, ip = 0;
279 	else if (cmd == PRC_HOSTDEAD)
280 		ip = 0;
281 	else if (errno == 0)
282 		return NULL;
283 	if (ip) {
284 		rip_pcbnotify(&rawcbtable, satocsin(sa)->sin_addr,
285 		    ip->ip_src, ip->ip_p, errno, notify);
286 
287 		/* XXX mapped address case */
288 	} else
289 		in_pcbnotifyall(&rawcbtable, satocsin(sa)->sin_addr, errno,
290 		    notify);
291 	return NULL;
292 }
293 
294 /*
295  * Generate IP header and pass packet to ip_output.
296  * Tack on options user may have setup with control call.
297  */
298 int
299 rip_output(struct mbuf *m, struct inpcb *inp, struct mbuf *control,
300     struct lwp *l)
301 {
302 	struct ip *ip;
303 	struct mbuf *opts;
304 	struct ip_pktopts pktopts;
305 	kauth_cred_t cred;
306 	int error, flags;
307 
308 	flags = (inp->inp_socket->so_options & SO_DONTROUTE) |
309 	    IP_ALLOWBROADCAST | IP_RETURNMTU;
310 
311 	if (l == NULL)
312 		cred = NULL;
313 	else
314 		cred = l->l_cred;
315 
316 	/* Setup IP outgoing packet options */
317 	memset(&pktopts, 0, sizeof(pktopts));
318 	error = ip_setpktopts(control, &pktopts, &flags, inp, cred);
319 	if (control != NULL)
320 		m_freem(control);
321 	if (error != 0)
322 		goto release;
323 
324 	/*
325 	 * If the user handed us a complete IP packet, use it.
326 	 * Otherwise, allocate an mbuf for a header and fill it in.
327 	 */
328 	if ((inp->inp_flags & INP_HDRINCL) == 0) {
329 		if ((m->m_pkthdr.len + sizeof(struct ip)) > IP_MAXPACKET) {
330 			error = EMSGSIZE;
331 			goto release;
332 		}
333 		M_PREPEND(m, sizeof(struct ip), M_DONTWAIT);
334 		if (!m) {
335 			error = ENOBUFS;
336 			goto release;
337 		}
338 		ip = mtod(m, struct ip *);
339 		ip->ip_tos = 0;
340 		ip->ip_off = htons(0);
341 		ip->ip_p = inp->inp_ip.ip_p;
342 		ip->ip_len = htons(m->m_pkthdr.len);
343 		ip->ip_src = pktopts.ippo_laddr.sin_addr;
344 		ip->ip_dst = inp->inp_faddr;
345 		ip->ip_ttl = MAXTTL;
346 		opts = inp->inp_options;
347 	} else {
348 		if (m->m_pkthdr.len > IP_MAXPACKET) {
349 			error = EMSGSIZE;
350 			goto release;
351 		}
352 		if (m->m_pkthdr.len < sizeof(struct ip)) {
353 			error = EINVAL;
354 			goto release;
355 		}
356 		ip = mtod(m, struct ip *);
357 
358 		/*
359 		 * If the mbuf is read-only, we need to allocate
360 		 * a new mbuf for the header, since we need to
361 		 * modify the header.
362 		 */
363 		if (M_READONLY(m)) {
364 			int hlen = ip->ip_hl << 2;
365 
366 			m = m_copyup(m, hlen, (max_linkhdr + 3) & ~3);
367 			if (m == NULL) {
368 				error = ENOMEM;
369 				goto release;
370 			}
371 			ip = mtod(m, struct ip *);
372 		}
373 
374 		/* XXX userland passes ip_len and ip_off in host order */
375 		if (m->m_pkthdr.len != ip->ip_len) {
376 			error = EINVAL;
377 			goto release;
378 		}
379 		HTONS(ip->ip_len);
380 		HTONS(ip->ip_off);
381 
382 		if (ip->ip_id != 0 || m->m_pkthdr.len < IP_MINFRAGSIZE)
383 			flags |= IP_NOIPNEWID;
384 		opts = NULL;
385 
386 		/* Prevent ip_output from overwriting header fields. */
387 		flags |= IP_RAWOUTPUT;
388 
389 		IP_STATINC(IP_STAT_RAWOUT);
390 	}
391 
392 	/*
393 	 * IP output.  Note: if IP_RETURNMTU flag is set, the MTU size
394 	 * will be stored in inp_errormtu.
395 	 */
396 	return ip_output(m, opts, &inp->inp_route, flags, pktopts.ippo_imo,
397 	    inp);
398 
399  release:
400 	if (m != NULL)
401 		m_freem(m);
402 	return error;
403 }
404 
405 /*
406  * Raw IP socket option processing.
407  */
408 int
409 rip_ctloutput(int op, struct socket *so, struct sockopt *sopt)
410 {
411 	struct inpcb *inp = sotoinpcb(so);
412 	int error = 0;
413 	int optval;
414 
415 	if (sopt->sopt_level == SOL_SOCKET && sopt->sopt_name == SO_NOHEADER) {
416 		if (op == PRCO_GETOPT) {
417 			optval = (inp->inp_flags & INP_NOHEADER) ? 1 : 0;
418 			error = sockopt_set(sopt, &optval, sizeof(optval));
419 		} else if (op == PRCO_SETOPT) {
420 			error = sockopt_getint(sopt, &optval);
421 			if (error)
422 				goto out;
423 			if (optval) {
424 				inp->inp_flags &= ~INP_HDRINCL;
425 				inp->inp_flags |= INP_NOHEADER;
426 			} else
427 				inp->inp_flags &= ~INP_NOHEADER;
428 		}
429 		goto out;
430 	} else if (sopt->sopt_level != IPPROTO_IP)
431 		return ip_ctloutput(op, so, sopt);
432 
433 	switch (op) {
434 
435 	case PRCO_SETOPT:
436 		switch (sopt->sopt_name) {
437 		case IP_HDRINCL:
438 			error = sockopt_getint(sopt, &optval);
439 			if (error)
440 				break;
441 			if (optval)
442 				inp->inp_flags |= INP_HDRINCL;
443 			else
444 				inp->inp_flags &= ~INP_HDRINCL;
445 			break;
446 
447 #ifdef MROUTING
448 		case MRT_INIT:
449 		case MRT_DONE:
450 		case MRT_ADD_VIF:
451 		case MRT_DEL_VIF:
452 		case MRT_ADD_MFC:
453 		case MRT_DEL_MFC:
454 		case MRT_ASSERT:
455 		case MRT_API_CONFIG:
456 		case MRT_ADD_BW_UPCALL:
457 		case MRT_DEL_BW_UPCALL:
458 			error = ip_mrouter_set(so, sopt);
459 			break;
460 #endif
461 
462 		default:
463 			error = ip_ctloutput(op, so, sopt);
464 			break;
465 		}
466 		break;
467 
468 	case PRCO_GETOPT:
469 		switch (sopt->sopt_name) {
470 		case IP_HDRINCL:
471 			optval = inp->inp_flags & INP_HDRINCL;
472 			error = sockopt_set(sopt, &optval, sizeof(optval));
473 			break;
474 
475 #ifdef MROUTING
476 		case MRT_VERSION:
477 		case MRT_ASSERT:
478 		case MRT_API_SUPPORT:
479 		case MRT_API_CONFIG:
480 			error = ip_mrouter_get(so, sopt);
481 			break;
482 #endif
483 
484 		default:
485 			error = ip_ctloutput(op, so, sopt);
486 			break;
487 		}
488 		break;
489 	}
490  out:
491 	return error;
492 }
493 
494 int
495 rip_connect_pcb(struct inpcb *inp, struct sockaddr_in *addr)
496 {
497 
498 	if (IFNET_READER_EMPTY())
499 		return (EADDRNOTAVAIL);
500 	if (addr->sin_family != AF_INET)
501 		return (EAFNOSUPPORT);
502 	inp->inp_faddr = addr->sin_addr;
503 	return (0);
504 }
505 
506 static void
507 rip_disconnect1(struct inpcb *inp)
508 {
509 
510 	inp->inp_faddr = zeroin_addr;
511 }
512 
513 static int
514 rip_attach(struct socket *so, int proto)
515 {
516 	struct inpcb *inp;
517 	int error;
518 
519 	KASSERT(sotoinpcb(so) == NULL);
520 	sosetlock(so);
521 
522 	if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) {
523 		error = soreserve(so, rip_sendspace, rip_recvspace);
524 		if (error) {
525 			return error;
526 		}
527 	}
528 
529 	error = in_pcballoc(so, &rawcbtable);
530 	if (error) {
531 		return error;
532 	}
533 	inp = sotoinpcb(so);
534 	inp->inp_ip.ip_p = proto;
535 	KASSERT(solocked(so));
536 
537 	return 0;
538 }
539 
540 static void
541 rip_detach(struct socket *so)
542 {
543 	struct inpcb *inp;
544 
545 	KASSERT(solocked(so));
546 	inp = sotoinpcb(so);
547 	KASSERT(inp != NULL);
548 
549 #ifdef MROUTING
550 	extern struct socket *ip_mrouter;
551 	if (so == ip_mrouter) {
552 		ip_mrouter_done();
553 	}
554 #endif
555 	in_pcbdetach(inp);
556 }
557 
558 static int
559 rip_accept(struct socket *so, struct sockaddr *nam)
560 {
561 	KASSERT(solocked(so));
562 
563 	panic("rip_accept");
564 
565 	return EOPNOTSUPP;
566 }
567 
568 static int
569 rip_bind(struct socket *so, struct sockaddr *nam, struct lwp *l)
570 {
571 	struct inpcb *inp = sotoinpcb(so);
572 	struct sockaddr_in *addr = (struct sockaddr_in *)nam;
573 	int error = 0;
574 	int s, ss;
575 	struct ifaddr *ifa;
576 
577 	KASSERT(solocked(so));
578 	KASSERT(inp != NULL);
579 	KASSERT(nam != NULL);
580 
581 	if (addr->sin_len != sizeof(*addr))
582 		return EINVAL;
583 
584 	s = splsoftnet();
585 	if (IFNET_READER_EMPTY()) {
586 		error = EADDRNOTAVAIL;
587 		goto release;
588 	}
589 	if (addr->sin_family != AF_INET) {
590 		error = EAFNOSUPPORT;
591 		goto release;
592 	}
593 	ss = pserialize_read_enter();
594 	if ((ifa = ifa_ifwithaddr(sintosa(addr))) == NULL &&
595 	    !in_nullhost(addr->sin_addr))
596 	{
597 		pserialize_read_exit(ss);
598 		error = EADDRNOTAVAIL;
599 		goto release;
600 	}
601         if (ifa && (ifatoia(ifa))->ia4_flags & IN6_IFF_DUPLICATED) {
602 		pserialize_read_exit(ss);
603 		error = EADDRNOTAVAIL;
604 		goto release;
605 	}
606 	pserialize_read_exit(ss);
607 
608 	inp->inp_laddr = addr->sin_addr;
609 
610 release:
611 	splx(s);
612 	return error;
613 }
614 
615 static int
616 rip_listen(struct socket *so, struct lwp *l)
617 {
618 	KASSERT(solocked(so));
619 
620 	return EOPNOTSUPP;
621 }
622 
623 static int
624 rip_connect(struct socket *so, struct sockaddr *nam, struct lwp *l)
625 {
626 	struct inpcb *inp = sotoinpcb(so);
627 	int error = 0;
628 	int s;
629 
630 	KASSERT(solocked(so));
631 	KASSERT(inp != NULL);
632 	KASSERT(nam != NULL);
633 
634 	s = splsoftnet();
635 	error = rip_connect_pcb(inp, (struct sockaddr_in *)nam);
636 	if (! error)
637 		soisconnected(so);
638 	splx(s);
639 
640 	return error;
641 }
642 
643 static int
644 rip_connect2(struct socket *so, struct socket *so2)
645 {
646 	KASSERT(solocked(so));
647 
648 	return EOPNOTSUPP;
649 }
650 
651 static int
652 rip_disconnect(struct socket *so)
653 {
654 	struct inpcb *inp = sotoinpcb(so);
655 	int s;
656 
657 	KASSERT(solocked(so));
658 	KASSERT(inp != NULL);
659 
660 	s = splsoftnet();
661 	soisdisconnected(so);
662 	rip_disconnect1(inp);
663 	splx(s);
664 
665 	return 0;
666 }
667 
668 static int
669 rip_shutdown(struct socket *so)
670 {
671 	int s;
672 
673 	KASSERT(solocked(so));
674 
675 	/*
676 	 * Mark the connection as being incapable of further input.
677 	 */
678 	s = splsoftnet();
679 	socantsendmore(so);
680 	splx(s);
681 
682 	return 0;
683 }
684 
685 static int
686 rip_abort(struct socket *so)
687 {
688 	KASSERT(solocked(so));
689 
690 	panic("rip_abort");
691 
692 	return EOPNOTSUPP;
693 }
694 
695 static int
696 rip_ioctl(struct socket *so, u_long cmd, void *nam, struct ifnet *ifp)
697 {
698 	return in_control(so, cmd, nam, ifp);
699 }
700 
701 static int
702 rip_stat(struct socket *so, struct stat *ub)
703 {
704 	KASSERT(solocked(so));
705 
706 	/* stat: don't bother with a blocksize. */
707 	return 0;
708 }
709 
710 static int
711 rip_peeraddr(struct socket *so, struct sockaddr *nam)
712 {
713 	int s;
714 
715 	KASSERT(solocked(so));
716 	KASSERT(sotoinpcb(so) != NULL);
717 	KASSERT(nam != NULL);
718 
719 	s = splsoftnet();
720 	in_setpeeraddr(sotoinpcb(so), (struct sockaddr_in *)nam);
721 	splx(s);
722 
723 	return 0;
724 }
725 
726 static int
727 rip_sockaddr(struct socket *so, struct sockaddr *nam)
728 {
729 	int s;
730 
731 	KASSERT(solocked(so));
732 	KASSERT(sotoinpcb(so) != NULL);
733 	KASSERT(nam != NULL);
734 
735 	s = splsoftnet();
736 	in_setsockaddr(sotoinpcb(so), (struct sockaddr_in *)nam);
737 	splx(s);
738 
739 	return 0;
740 }
741 
742 static int
743 rip_rcvd(struct socket *so, int flags, struct lwp *l)
744 {
745 	KASSERT(solocked(so));
746 
747 	return EOPNOTSUPP;
748 }
749 
750 static int
751 rip_recvoob(struct socket *so, struct mbuf *m, int flags)
752 {
753 	KASSERT(solocked(so));
754 
755 	return EOPNOTSUPP;
756 }
757 
758 static int
759 rip_send(struct socket *so, struct mbuf *m, struct sockaddr *nam,
760     struct mbuf *control, struct lwp *l)
761 {
762 	struct inpcb *inp = sotoinpcb(so);
763 	int error = 0;
764 	int s;
765 
766 	KASSERT(solocked(so));
767 	KASSERT(inp != NULL);
768 	KASSERT(m != NULL);
769 
770 	/*
771 	 * Ship a packet out.  The appropriate raw output
772 	 * routine handles any massaging necessary.
773 	 */
774 	s = splsoftnet();
775 	if (nam) {
776 		if ((so->so_state & SS_ISCONNECTED) != 0) {
777 			error = EISCONN;
778 			goto die;
779 		}
780 		error = rip_connect_pcb(inp, (struct sockaddr_in *)nam);
781 		if (error)
782 			goto die;
783 	} else {
784 		if ((so->so_state & SS_ISCONNECTED) == 0) {
785 			error = ENOTCONN;
786 			goto die;
787 		}
788 	}
789 	error = rip_output(m, inp, control, l);
790 	m = NULL;
791 	control = NULL;
792 	if (nam)
793 		rip_disconnect1(inp);
794  die:
795 	if (m != NULL)
796 		m_freem(m);
797 	if (control != NULL)
798 		m_freem(control);
799 
800 	splx(s);
801 	return error;
802 }
803 
804 static int
805 rip_sendoob(struct socket *so, struct mbuf *m, struct mbuf *control)
806 {
807 	KASSERT(solocked(so));
808 
809 	m_freem(m);
810 	m_freem(control);
811 
812 	return EOPNOTSUPP;
813 }
814 
815 static int
816 rip_purgeif(struct socket *so, struct ifnet *ifp)
817 {
818 	int s;
819 
820 	s = splsoftnet();
821 	mutex_enter(softnet_lock);
822 	in_pcbpurgeif0(&rawcbtable, ifp);
823 #ifdef NET_MPSAFE
824 	mutex_exit(softnet_lock);
825 #endif
826 	in_purgeif(ifp);
827 #ifdef NET_MPSAFE
828 	mutex_enter(softnet_lock);
829 #endif
830 	in_pcbpurgeif(&rawcbtable, ifp);
831 	mutex_exit(softnet_lock);
832 	splx(s);
833 
834 	return 0;
835 }
836 
837 PR_WRAP_USRREQS(rip)
838 #define	rip_attach	rip_attach_wrapper
839 #define	rip_detach	rip_detach_wrapper
840 #define	rip_accept	rip_accept_wrapper
841 #define	rip_bind	rip_bind_wrapper
842 #define	rip_listen	rip_listen_wrapper
843 #define	rip_connect	rip_connect_wrapper
844 #define	rip_connect2	rip_connect2_wrapper
845 #define	rip_disconnect	rip_disconnect_wrapper
846 #define	rip_shutdown	rip_shutdown_wrapper
847 #define	rip_abort	rip_abort_wrapper
848 #define	rip_ioctl	rip_ioctl_wrapper
849 #define	rip_stat	rip_stat_wrapper
850 #define	rip_peeraddr	rip_peeraddr_wrapper
851 #define	rip_sockaddr	rip_sockaddr_wrapper
852 #define	rip_rcvd	rip_rcvd_wrapper
853 #define	rip_recvoob	rip_recvoob_wrapper
854 #define	rip_send	rip_send_wrapper
855 #define	rip_sendoob	rip_sendoob_wrapper
856 #define	rip_purgeif	rip_purgeif_wrapper
857 
858 const struct pr_usrreqs rip_usrreqs = {
859 	.pr_attach	= rip_attach,
860 	.pr_detach	= rip_detach,
861 	.pr_accept	= rip_accept,
862 	.pr_bind	= rip_bind,
863 	.pr_listen	= rip_listen,
864 	.pr_connect	= rip_connect,
865 	.pr_connect2	= rip_connect2,
866 	.pr_disconnect	= rip_disconnect,
867 	.pr_shutdown	= rip_shutdown,
868 	.pr_abort	= rip_abort,
869 	.pr_ioctl	= rip_ioctl,
870 	.pr_stat	= rip_stat,
871 	.pr_peeraddr	= rip_peeraddr,
872 	.pr_sockaddr	= rip_sockaddr,
873 	.pr_rcvd	= rip_rcvd,
874 	.pr_recvoob	= rip_recvoob,
875 	.pr_send	= rip_send,
876 	.pr_sendoob	= rip_sendoob,
877 	.pr_purgeif	= rip_purgeif,
878 };
879 
880 static void
881 sysctl_net_inet_raw_setup(struct sysctllog **clog)
882 {
883 
884 	sysctl_createv(clog, 0, NULL, NULL,
885 		       CTLFLAG_PERMANENT,
886 		       CTLTYPE_NODE, "inet", NULL,
887 		       NULL, 0, NULL, 0,
888 		       CTL_NET, PF_INET, CTL_EOL);
889 	sysctl_createv(clog, 0, NULL, NULL,
890 		       CTLFLAG_PERMANENT,
891 		       CTLTYPE_NODE, "raw",
892 		       SYSCTL_DESCR("Raw IPv4 settings"),
893 		       NULL, 0, NULL, 0,
894 		       CTL_NET, PF_INET, IPPROTO_RAW, CTL_EOL);
895 
896 	sysctl_createv(clog, 0, NULL, NULL,
897 		       CTLFLAG_PERMANENT,
898 		       CTLTYPE_STRUCT, "pcblist",
899 		       SYSCTL_DESCR("Raw IPv4 control block list"),
900 		       sysctl_inpcblist, 0, &rawcbtable, 0,
901 		       CTL_NET, PF_INET, IPPROTO_RAW,
902 		       CTL_CREATE, CTL_EOL);
903 }
904