1/*-
2 * Copyright (c) 2009-2010 Ana Kukec <anchie@FreeBSD.org>
3 * All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 * 1. Redistributions of source code must retain the above copyright
9 *    notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 *    notice, this list of conditions and the following disclaimer in the
12 *    documentation and/or other materials provided with the distribution.
13 *
14 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
15 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
16 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
17 * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
18 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
19 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
20 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
21 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
22 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
23 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
24 * SUCH DAMAGE.
25 */
26
27#include <sys/cdefs.h>
28__FBSDID("$FreeBSD: releng/10.3/sys/netinet6/send.c 254889 2013-08-25 21:54:41Z markj $");
29
30#include "opt_kdtrace.h"
31
32#include <sys/param.h>
33#include <sys/kernel.h>
34#include <sys/mbuf.h>
35#include <sys/module.h>
36#include <sys/priv.h>
37#include <sys/protosw.h>
38#include <sys/sdt.h>
39#include <sys/systm.h>
40#include <sys/socket.h>
41#include <sys/sockstate.h>
42#include <sys/sockbuf.h>
43#include <sys/socketvar.h>
44#include <sys/types.h>
45
46#include <net/route.h>
47#include <net/if.h>
48#include <net/if_var.h>
49#include <net/vnet.h>
50
51#include <netinet/in.h>
52#include <netinet/in_kdtrace.h>
53#include <netinet/ip_var.h>
54#include <netinet/ip6.h>
55#include <netinet/icmp6.h>
56
57#include <netinet6/in6_var.h>
58#include <netinet6/nd6.h>
59#include <netinet6/scope6_var.h>
60#include <netinet6/send.h>
61
62static MALLOC_DEFINE(M_SEND, "send", "Secure Neighbour Discovery");
63
64/*
65 * The socket used to communicate with the SeND daemon.
66 */
67static VNET_DEFINE(struct socket *, send_so);
68#define	V_send_so	VNET(send_so)
69
70u_long	send_sendspace	= 8 * (1024 + sizeof(struct sockaddr_send));
71u_long	send_recvspace	= 9216;
72
73struct mtx	send_mtx;
74#define SEND_LOCK_INIT()	mtx_init(&send_mtx, "send_mtx", NULL, MTX_DEF)
75#define SEND_LOCK()		mtx_lock(&send_mtx)
76#define SEND_UNLOCK()		mtx_unlock(&send_mtx)
77#define SEND_LOCK_DESTROY()     mtx_destroy(&send_mtx)
78
79static int
80send_attach(struct socket *so, int proto, struct thread *td)
81{
82	int error;
83
84	SEND_LOCK();
85	if (V_send_so != NULL) {
86		SEND_UNLOCK();
87		return (EEXIST);
88	}
89
90	error = priv_check(td, PRIV_NETINET_RAW);
91	if (error) {
92		SEND_UNLOCK();
93		return(error);
94	}
95
96	if (proto != IPPROTO_SEND) {
97		SEND_UNLOCK();
98		return (EPROTONOSUPPORT);
99	}
100	error = soreserve(so, send_sendspace, send_recvspace);
101	if (error) {
102		SEND_UNLOCK();
103		return(error);
104	}
105
106	V_send_so = so;
107	SEND_UNLOCK();
108
109	return (0);
110}
111
112static int
113send_output(struct mbuf *m, struct ifnet *ifp, int direction)
114{
115	struct ip6_hdr *ip6;
116	struct sockaddr_in6 dst;
117	struct icmp6_hdr *icmp6;
118	int icmp6len;
119
120	/*
121	 * Receive incoming (SeND-protected) or outgoing traffic
122	 * (SeND-validated) from the SeND user space application.
123	 */
124
125	switch (direction) {
126	case SND_IN:
127		if (m->m_len < (sizeof(struct ip6_hdr) +
128		    sizeof(struct icmp6_hdr))) {
129			m = m_pullup(m, sizeof(struct ip6_hdr) +
130			    sizeof(struct icmp6_hdr));
131			if (!m)
132				return (ENOBUFS);
133		}
134
135		/* Before passing off the mbuf record the proper interface. */
136		m->m_pkthdr.rcvif = ifp;
137
138		if (m->m_flags & M_PKTHDR)
139			icmp6len = m->m_pkthdr.len - sizeof(struct ip6_hdr);
140		else
141			panic("Doh! not the first mbuf.");
142
143		ip6 = mtod(m, struct ip6_hdr *);
144		icmp6 = (struct icmp6_hdr *)(ip6 + 1);
145
146		/*
147		 * Output the packet as icmp6.c:icpm6_input() would do.
148		 * The mbuf is always consumed, so we do not have to
149		 * care about that.
150		 */
151		switch (icmp6->icmp6_type) {
152		case ND_NEIGHBOR_SOLICIT:
153			nd6_ns_input(m, sizeof(struct ip6_hdr), icmp6len);
154			break;
155		case ND_NEIGHBOR_ADVERT:
156			nd6_na_input(m, sizeof(struct ip6_hdr), icmp6len);
157			break;
158		case ND_REDIRECT:
159			icmp6_redirect_input(m, sizeof(struct ip6_hdr));
160			break;
161		case ND_ROUTER_SOLICIT:
162			nd6_rs_input(m, sizeof(struct ip6_hdr), icmp6len);
163			break;
164		case ND_ROUTER_ADVERT:
165			nd6_ra_input(m, sizeof(struct ip6_hdr), icmp6len);
166			break;
167		default:
168			return (ENOSYS);
169		}
170		return (0);
171
172	case SND_OUT:
173		if (m->m_len < sizeof(struct ip6_hdr)) {
174			m = m_pullup(m, sizeof(struct ip6_hdr));
175			if (!m)
176				return (ENOBUFS);
177		}
178		ip6 = mtod(m, struct ip6_hdr *);
179		if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst))
180			m->m_flags |= M_MCAST;
181
182		bzero(&dst, sizeof(dst));
183		dst.sin6_family = AF_INET6;
184		dst.sin6_len = sizeof(dst);
185		dst.sin6_addr = ip6->ip6_dst;
186
187		m_clrprotoflags(m);	/* Avoid confusing lower layers. */
188
189		IP_PROBE(send, NULL, NULL, ip6, ifp, NULL, ip6);
190
191		/*
192		 * Output the packet as nd6.c:nd6_output_lle() would do.
193		 * The mbuf is always consumed, so we do not have to care
194		 * about that.
195		 * XXX-BZ as we added data, what about fragmenting,
196		 * if now needed?
197		 */
198		int error;
199		error = ((*ifp->if_output)(ifp, m, (struct sockaddr *)&dst,
200		    NULL));
201		if (error)
202			error = ENOENT;
203		return (error);
204
205	default:
206		panic("%s: direction %d neither SND_IN nor SND_OUT.",
207		     __func__, direction);
208	}
209}
210
211/*
212 * Receive a SeND message from user space to be either send out by the kernel
213 * or, with SeND ICMPv6 options removed, to be further processed by the icmp6
214 * input path.
215 */
216static int
217send_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam,
218    struct mbuf *control, struct thread *td)
219{
220	struct sockaddr_send *sendsrc;
221	struct ifnet *ifp;
222	int error;
223
224	KASSERT(V_send_so == so, ("%s: socket %p not send socket %p",
225		__func__, so, V_send_so));
226
227	sendsrc = (struct sockaddr_send *)nam;
228	ifp = ifnet_byindex_ref(sendsrc->send_ifidx);
229	if (ifp == NULL) {
230		error = ENETUNREACH;
231		goto err;
232	}
233
234	error = send_output(m, ifp, sendsrc->send_direction);
235	if_rele(ifp);
236	m = NULL;
237
238err:
239	if (m != NULL)
240		m_freem(m);
241	return (error);
242}
243
244static void
245send_close(struct socket *so)
246{
247
248	SEND_LOCK();
249	if (V_send_so)
250		V_send_so = NULL;
251	SEND_UNLOCK();
252}
253
254/*
255 * Send a SeND message to user space, that was either received and has to be
256 * validated or was about to be send out and has to be handled by the SEND
257 * daemon adding SeND ICMPv6 options.
258 */
259static int
260send_input(struct mbuf *m, struct ifnet *ifp, int direction, int msglen __unused)
261{
262	struct ip6_hdr *ip6;
263	struct sockaddr_send sendsrc;
264
265	SEND_LOCK();
266	if (V_send_so == NULL) {
267		SEND_UNLOCK();
268		return (-1);
269	}
270
271	/*
272	 * Make sure to clear any possible internally embedded scope before
273	 * passing the packet to user space for SeND cryptographic signature
274	 * validation to succeed.
275	 */
276	ip6 = mtod(m, struct ip6_hdr *);
277	in6_clearscope(&ip6->ip6_src);
278	in6_clearscope(&ip6->ip6_dst);
279
280	bzero(&sendsrc, sizeof(sendsrc));
281	sendsrc.send_len = sizeof(sendsrc);
282	sendsrc.send_family = AF_INET6;
283	sendsrc.send_direction = direction;
284	sendsrc.send_ifidx = ifp->if_index;
285
286	/*
287	 * Send incoming or outgoing traffic to user space either to be
288	 * protected (outgoing) or validated (incoming) according to rfc3971.
289	 */
290	SOCKBUF_LOCK(&V_send_so->so_rcv);
291	if (sbappendaddr_locked(&V_send_so->so_rcv,
292	    (struct sockaddr *)&sendsrc, m, NULL) == 0) {
293		SOCKBUF_UNLOCK(&V_send_so->so_rcv);
294		/* XXX stats. */
295		m_freem(m);
296	} else {
297		sorwakeup_locked(V_send_so);
298	}
299
300	SEND_UNLOCK();
301	return (0);
302}
303
304struct pr_usrreqs send_usrreqs = {
305	.pru_attach =		send_attach,
306	.pru_send =		send_send,
307	.pru_detach =		send_close
308};
309struct protosw send_protosw = {
310	.pr_type =		SOCK_RAW,
311	.pr_flags =		PR_ATOMIC|PR_ADDR,
312	.pr_protocol =		IPPROTO_SEND,
313	.pr_usrreqs =		&send_usrreqs
314};
315
316static int
317send_modevent(module_t mod, int type, void *unused)
318{
319#ifdef __notyet__
320	VNET_ITERATOR_DECL(vnet_iter);
321#endif
322	int error;
323
324	switch (type) {
325	case MOD_LOAD:
326		SEND_LOCK_INIT();
327
328		error = pf_proto_register(PF_INET6, &send_protosw);
329		if (error != 0) {
330			printf("%s:%d: MOD_LOAD pf_proto_register(): %d\n",
331			   __func__, __LINE__, error);
332			SEND_LOCK_DESTROY();
333			break;
334		}
335		send_sendso_input_hook = send_input;
336		break;
337	case MOD_UNLOAD:
338		/* Do not allow unloading w/o locking. */
339		return (EBUSY);
340#ifdef __notyet__
341		VNET_LIST_RLOCK_NOSLEEP();
342		SEND_LOCK();
343		VNET_FOREACH(vnet_iter) {
344			CURVNET_SET(vnet_iter);
345			if (V_send_so != NULL) {
346				CURVNET_RESTORE();
347				SEND_UNLOCK();
348				VNET_LIST_RUNLOCK_NOSLEEP();
349				return (EBUSY);
350			}
351			CURVNET_RESTORE();
352		}
353		SEND_UNLOCK();
354		VNET_LIST_RUNLOCK_NOSLEEP();
355		error = pf_proto_unregister(PF_INET6, IPPROTO_SEND, SOCK_RAW);
356		if (error == 0)
357			SEND_LOCK_DESTROY();
358		send_sendso_input_hook = NULL;
359		break;
360#endif
361	default:
362		error = 0;
363		break;
364	}
365
366	return (error);
367}
368
369static moduledata_t sendmod = {
370	"send",
371	send_modevent,
372	0
373};
374
375DECLARE_MODULE(send, sendmod, SI_SUB_PROTO_DOMAIN, SI_ORDER_ANY);
376