if_loop.c revision 191548 
1/*-
2 * Copyright (c) 1982, 1986, 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 * 4. Neither the name of the University nor the names of its contributors
14 *    may be used to endorse or promote products derived from this software
15 *    without specific prior written permission.
16 *
17 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27 * SUCH DAMAGE.
28 *
29 *	@(#)if_loop.c	8.2 (Berkeley) 1/9/95
30 * $FreeBSD: head/sys/net/if_loop.c 191548 2009-04-26 22:06:42Z zec $
31 */
32
33/*
34 * Loopback interface driver for protocol testing and timing.
35 */
36
37#include "opt_atalk.h"
38#include "opt_inet.h"
39#include "opt_inet6.h"
40#include "opt_ipx.h"
41#include "opt_route.h"
42#include "opt_mac.h"
43
44#include <sys/param.h>
45#include <sys/systm.h>
46#include <sys/kernel.h>
47#include <sys/mbuf.h>
48#include <sys/module.h>
49#include <machine/bus.h>
50#include <sys/rman.h>
51#include <sys/socket.h>
52#include <sys/sockio.h>
53#include <sys/sysctl.h>
54#include <sys/vimage.h>
55
56#include <net/if.h>
57#include <net/if_clone.h>
58#include <net/if_types.h>
59#include <net/netisr.h>
60#include <net/route.h>
61#include <net/bpf.h>
62#include <net/vnet.h>
63
64#ifdef	INET
65#include <netinet/in.h>
66#include <netinet/in_var.h>
67#endif
68
69#ifdef IPX
70#include <netipx/ipx.h>
71#include <netipx/ipx_if.h>
72#endif
73
74#ifdef INET6
75#ifndef INET
76#include <netinet/in.h>
77#endif
78#include <netinet6/in6_var.h>
79#include <netinet/ip6.h>
80#endif
81
82#ifdef NETATALK
83#include <netatalk/at.h>
84#include <netatalk/at_var.h>
85#endif
86
87#include <security/mac/mac_framework.h>
88
89#ifdef TINY_LOMTU
90#define	LOMTU	(1024+512)
91#elif defined(LARGE_LOMTU)
92#define LOMTU	131072
93#else
94#define LOMTU	16384
95#endif
96
97#define	LO_CSUM_FEATURES	(CSUM_IP | CSUM_TCP | CSUM_UDP | CSUM_SCTP)
98#define	LO_CSUM_SET		(CSUM_DATA_VALID | CSUM_PSEUDO_HDR | \
99				    CSUM_IP_CHECKED | CSUM_IP_VALID | \
100				    CSUM_SCTP_VALID)
101
102int		loioctl(struct ifnet *, u_long, caddr_t);
103static void	lortrequest(int, struct rtentry *, struct rt_addrinfo *);
104int		looutput(struct ifnet *ifp, struct mbuf *m,
105		    struct sockaddr *dst, struct route *ro);
106static int	lo_clone_create(struct if_clone *, int, caddr_t);
107static void	lo_clone_destroy(struct ifnet *);
108static int	vnet_loif_iattach(const void *);
109
110#ifdef VIMAGE_GLOBALS
111struct ifnet *loif;			/* Used externally */
112#endif
113
114#ifndef VIMAGE_GLOBALS
115static const vnet_modinfo_t vnet_loif_modinfo = {
116	.vmi_id		= VNET_MOD_LOIF,
117	.vmi_name	= "loif",
118	.vmi_iattach	= vnet_loif_iattach
119};
120#endif /* !VIMAGE_GLOBALS */
121
122IFC_SIMPLE_DECLARE(lo, 1);
123
124static void
125lo_clone_destroy(struct ifnet *ifp)
126{
127#ifdef INVARIANTS
128	INIT_VNET_NET(ifp->if_vnet);
129#endif
130
131	/* XXX: destroying lo0 will lead to panics. */
132	KASSERT(V_loif != ifp, ("%s: destroying lo0", __func__));
133
134	bpfdetach(ifp);
135	if_detach(ifp);
136	if_free(ifp);
137}
138
139static int
140lo_clone_create(struct if_clone *ifc, int unit, caddr_t params)
141{
142	INIT_VNET_NET(curvnet);
143	struct ifnet *ifp;
144
145	ifp = if_alloc(IFT_LOOP);
146	if (ifp == NULL)
147		return (ENOSPC);
148
149	if_initname(ifp, ifc->ifc_name, unit);
150	ifp->if_mtu = LOMTU;
151	ifp->if_flags = IFF_LOOPBACK | IFF_MULTICAST;
152	ifp->if_ioctl = loioctl;
153	ifp->if_output = looutput;
154	ifp->if_snd.ifq_maxlen = ifqmaxlen;
155	ifp->if_capabilities = ifp->if_capenable = IFCAP_HWCSUM;
156	ifp->if_hwassist = LO_CSUM_FEATURES;
157	if_attach(ifp);
158	bpfattach(ifp, DLT_NULL, sizeof(u_int32_t));
159	if (V_loif == NULL)
160		V_loif = ifp;
161
162	return (0);
163}
164
165static int vnet_loif_iattach(const void *unused __unused)
166{
167	INIT_VNET_NET(curvnet);
168
169	V_loif = NULL;
170	if_clone_attach(&lo_cloner);
171	return (0);
172}
173
174static int
175loop_modevent(module_t mod, int type, void *data)
176{
177
178	switch (type) {
179	case MOD_LOAD:
180#ifndef VIMAGE_GLOBALS
181		vnet_mod_register(&vnet_loif_modinfo);
182#else
183		vnet_loif_iattach(NULL);
184#endif
185		break;
186
187	case MOD_UNLOAD:
188		printf("loop module unload - not possible for this module type\n");
189		return (EINVAL);
190
191	default:
192		return (EOPNOTSUPP);
193	}
194	return (0);
195}
196
197static moduledata_t loop_mod = {
198	"loop",
199	loop_modevent,
200	0
201};
202
203DECLARE_MODULE(loop, loop_mod, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY);
204
205int
206looutput(struct ifnet *ifp, struct mbuf *m, struct sockaddr *dst,
207    struct route *ro)
208{
209	u_int32_t af;
210	struct rtentry *rt = NULL;
211#ifdef MAC
212	int error;
213#endif
214
215	M_ASSERTPKTHDR(m); /* check if we have the packet header */
216
217	if (ro != NULL)
218		rt = ro->ro_rt;
219#ifdef MAC
220	error = mac_ifnet_check_transmit(ifp, m);
221	if (error) {
222		m_freem(m);
223		return (error);
224	}
225#endif
226
227	if (rt && rt->rt_flags & (RTF_REJECT|RTF_BLACKHOLE)) {
228		m_freem(m);
229		return (rt->rt_flags & RTF_BLACKHOLE ? 0 :
230		        rt->rt_flags & RTF_HOST ? EHOSTUNREACH : ENETUNREACH);
231	}
232
233	ifp->if_opackets++;
234	ifp->if_obytes += m->m_pkthdr.len;
235
236	/* BPF writes need to be handled specially. */
237	if (dst->sa_family == AF_UNSPEC) {
238		bcopy(dst->sa_data, &af, sizeof(af));
239		dst->sa_family = af;
240	}
241
242#if 1	/* XXX */
243	switch (dst->sa_family) {
244	case AF_INET:
245		if (ifp->if_capenable & IFCAP_RXCSUM) {
246			m->m_pkthdr.csum_data = 0xffff;
247			m->m_pkthdr.csum_flags = LO_CSUM_SET;
248		}
249		m->m_pkthdr.csum_flags &= ~LO_CSUM_FEATURES;
250	case AF_INET6:
251	case AF_IPX:
252	case AF_APPLETALK:
253		break;
254	default:
255		printf("looutput: af=%d unexpected\n", dst->sa_family);
256		m_freem(m);
257		return (EAFNOSUPPORT);
258	}
259#endif
260	return (if_simloop(ifp, m, dst->sa_family, 0));
261}
262
263/*
264 * if_simloop()
265 *
266 * This function is to support software emulation of hardware loopback,
267 * i.e., for interfaces with the IFF_SIMPLEX attribute. Since they can't
268 * hear their own broadcasts, we create a copy of the packet that we
269 * would normally receive via a hardware loopback.
270 *
271 * This function expects the packet to include the media header of length hlen.
272 */
273int
274if_simloop(struct ifnet *ifp, struct mbuf *m, int af, int hlen)
275{
276	INIT_VNET_NET(ifp->if_vnet);
277	int isr;
278
279	M_ASSERTPKTHDR(m);
280	m_tag_delete_nonpersistent(m);
281	m->m_pkthdr.rcvif = ifp;
282
283#ifdef MAC
284	mac_ifnet_create_mbuf(ifp, m);
285#endif
286
287	/*
288	 * Let BPF see incoming packet in the following manner:
289	 *  - Emulated packet loopback for a simplex interface
290	 *    (net/if_ethersubr.c)
291	 *	-> passes it to ifp's BPF
292	 *  - IPv4/v6 multicast packet loopback (netinet(6)/ip(6)_output.c)
293	 *	-> not passes it to any BPF
294	 *  - Normal packet loopback from myself to myself (net/if_loop.c)
295	 *	-> passes to lo0's BPF (even in case of IPv6, where ifp!=lo0)
296	 */
297	if (hlen > 0) {
298		if (bpf_peers_present(ifp->if_bpf)) {
299			bpf_mtap(ifp->if_bpf, m);
300		}
301	} else {
302		if (bpf_peers_present(V_loif->if_bpf)) {
303			if ((m->m_flags & M_MCAST) == 0 || V_loif == ifp) {
304				/* XXX beware sizeof(af) != 4 */
305				u_int32_t af1 = af;
306
307				/*
308				 * We need to prepend the address family.
309				 */
310				bpf_mtap2(V_loif->if_bpf, &af1, sizeof(af1), m);
311			}
312		}
313	}
314
315	/* Strip away media header */
316	if (hlen > 0) {
317		m_adj(m, hlen);
318#ifndef __NO_STRICT_ALIGNMENT
319		/*
320		 * Some archs do not like unaligned data, so
321		 * we move data down in the first mbuf.
322		 */
323		if (mtod(m, vm_offset_t) & 3) {
324			KASSERT(hlen >= 3, ("if_simloop: hlen too small"));
325			bcopy(m->m_data,
326			    (char *)(mtod(m, vm_offset_t)
327				- (mtod(m, vm_offset_t) & 3)),
328			    m->m_len);
329			m->m_data -= (mtod(m,vm_offset_t) & 3);
330		}
331#endif
332	}
333
334	/* Deliver to upper layer protocol */
335	switch (af) {
336#ifdef INET
337	case AF_INET:
338		isr = NETISR_IP;
339		break;
340#endif
341#ifdef INET6
342	case AF_INET6:
343		m->m_flags |= M_LOOP;
344		isr = NETISR_IPV6;
345		break;
346#endif
347#ifdef IPX
348	case AF_IPX:
349		isr = NETISR_IPX;
350		break;
351#endif
352#ifdef NETATALK
353	case AF_APPLETALK:
354		isr = NETISR_ATALK2;
355		break;
356#endif
357	default:
358		printf("if_simloop: can't handle af=%d\n", af);
359		m_freem(m);
360		return (EAFNOSUPPORT);
361	}
362	ifp->if_ipackets++;
363	ifp->if_ibytes += m->m_pkthdr.len;
364	netisr_queue(isr, m);	/* mbuf is free'd on failure. */
365	return (0);
366}
367
368/* ARGSUSED */
369static void
370lortrequest(int cmd, struct rtentry *rt, struct rt_addrinfo *info)
371{
372
373	RT_LOCK_ASSERT(rt);
374	rt->rt_rmx.rmx_mtu = rt->rt_ifp->if_mtu;
375}
376
377/*
378 * Process an ioctl request.
379 */
380/* ARGSUSED */
381int
382loioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
383{
384	struct ifaddr *ifa;
385	struct ifreq *ifr = (struct ifreq *)data;
386	int error = 0, mask;
387
388	switch (cmd) {
389	case SIOCSIFADDR:
390		ifp->if_flags |= IFF_UP;
391		ifp->if_drv_flags |= IFF_DRV_RUNNING;
392		ifa = (struct ifaddr *)data;
393		ifa->ifa_rtrequest = lortrequest;
394		/*
395		 * Everything else is done at a higher level.
396		 */
397		break;
398
399	case SIOCADDMULTI:
400	case SIOCDELMULTI:
401		if (ifr == 0) {
402			error = EAFNOSUPPORT;		/* XXX */
403			break;
404		}
405		switch (ifr->ifr_addr.sa_family) {
406
407#ifdef INET
408		case AF_INET:
409			break;
410#endif
411#ifdef INET6
412		case AF_INET6:
413			break;
414#endif
415
416		default:
417			error = EAFNOSUPPORT;
418			break;
419		}
420		break;
421
422	case SIOCSIFMTU:
423		ifp->if_mtu = ifr->ifr_mtu;
424		break;
425
426	case SIOCSIFFLAGS:
427		break;
428
429	case SIOCSIFCAP:
430		mask = ifp->if_capenable ^ ifr->ifr_reqcap;
431		if ((mask & IFCAP_RXCSUM) != 0)
432			ifp->if_capenable ^= IFCAP_RXCSUM;
433		if ((mask & IFCAP_TXCSUM) != 0)
434			ifp->if_capenable ^= IFCAP_TXCSUM;
435		if (ifp->if_capenable & IFCAP_TXCSUM)
436			ifp->if_hwassist = LO_CSUM_FEATURES;
437		else
438			ifp->if_hwassist = 0;
439		break;
440
441	default:
442		error = EINVAL;
443	}
444	return (error);
445}
446