if_loop.c revision 191148 
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 191148 2009-04-16 20:30:28Z kmacy $
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	INIT_VNET_NET(curvnet);
178
179	switch (type) {
180	case MOD_LOAD:
181#ifndef VIMAGE_GLOBALS
182		vnet_mod_register(&vnet_loif_modinfo);
183#else
184		vnet_loif_iattach(NULL);
185#endif
186		break;
187
188	case MOD_UNLOAD:
189		printf("loop module unload - not possible for this module type\n");
190		return (EINVAL);
191
192	default:
193		return (EOPNOTSUPP);
194	}
195	return (0);
196}
197
198static moduledata_t loop_mod = {
199	"loop",
200	loop_modevent,
201	0
202};
203
204DECLARE_MODULE(loop, loop_mod, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY);
205
206int
207looutput(struct ifnet *ifp, struct mbuf *m, struct sockaddr *dst,
208    struct route *ro)
209{
210	u_int32_t af;
211	struct rtentry *rt = NULL;
212#ifdef MAC
213	int error;
214#endif
215
216	M_ASSERTPKTHDR(m); /* check if we have the packet header */
217
218	if (ro != NULL)
219		rt = ro->ro_rt;
220#ifdef MAC
221	error = mac_ifnet_check_transmit(ifp, m);
222	if (error) {
223		m_freem(m);
224		return (error);
225	}
226#endif
227
228	if (rt && rt->rt_flags & (RTF_REJECT|RTF_BLACKHOLE)) {
229		m_freem(m);
230		return (rt->rt_flags & RTF_BLACKHOLE ? 0 :
231		        rt->rt_flags & RTF_HOST ? EHOSTUNREACH : ENETUNREACH);
232	}
233
234	ifp->if_opackets++;
235	ifp->if_obytes += m->m_pkthdr.len;
236
237	/* BPF writes need to be handled specially. */
238	if (dst->sa_family == AF_UNSPEC) {
239		bcopy(dst->sa_data, &af, sizeof(af));
240		dst->sa_family = af;
241	}
242
243#if 1	/* XXX */
244	switch (dst->sa_family) {
245	case AF_INET:
246		if (ifp->if_capenable & IFCAP_RXCSUM) {
247			m->m_pkthdr.csum_data = 0xffff;
248			m->m_pkthdr.csum_flags = LO_CSUM_SET;
249		}
250		m->m_pkthdr.csum_flags &= ~LO_CSUM_FEATURES;
251	case AF_INET6:
252	case AF_IPX:
253	case AF_APPLETALK:
254		break;
255	default:
256		printf("looutput: af=%d unexpected\n", dst->sa_family);
257		m_freem(m);
258		return (EAFNOSUPPORT);
259	}
260#endif
261	return (if_simloop(ifp, m, dst->sa_family, 0));
262}
263
264/*
265 * if_simloop()
266 *
267 * This function is to support software emulation of hardware loopback,
268 * i.e., for interfaces with the IFF_SIMPLEX attribute. Since they can't
269 * hear their own broadcasts, we create a copy of the packet that we
270 * would normally receive via a hardware loopback.
271 *
272 * This function expects the packet to include the media header of length hlen.
273 */
274int
275if_simloop(struct ifnet *ifp, struct mbuf *m, int af, int hlen)
276{
277	INIT_VNET_NET(ifp->if_vnet);
278	int isr;
279
280	M_ASSERTPKTHDR(m);
281	m_tag_delete_nonpersistent(m);
282	m->m_pkthdr.rcvif = ifp;
283
284#ifdef MAC
285	mac_ifnet_create_mbuf(ifp, m);
286#endif
287
288	/*
289	 * Let BPF see incoming packet in the following manner:
290	 *  - Emulated packet loopback for a simplex interface
291	 *    (net/if_ethersubr.c)
292	 *	-> passes it to ifp's BPF
293	 *  - IPv4/v6 multicast packet loopback (netinet(6)/ip(6)_output.c)
294	 *	-> not passes it to any BPF
295	 *  - Normal packet loopback from myself to myself (net/if_loop.c)
296	 *	-> passes to lo0's BPF (even in case of IPv6, where ifp!=lo0)
297	 */
298	if (hlen > 0) {
299		if (bpf_peers_present(ifp->if_bpf)) {
300			bpf_mtap(ifp->if_bpf, m);
301		}
302	} else {
303		if (bpf_peers_present(V_loif->if_bpf)) {
304			if ((m->m_flags & M_MCAST) == 0 || V_loif == ifp) {
305				/* XXX beware sizeof(af) != 4 */
306				u_int32_t af1 = af;
307
308				/*
309				 * We need to prepend the address family.
310				 */
311				bpf_mtap2(V_loif->if_bpf, &af1, sizeof(af1), m);
312			}
313		}
314	}
315
316	/* Strip away media header */
317	if (hlen > 0) {
318		m_adj(m, hlen);
319#ifndef __NO_STRICT_ALIGNMENT
320		/*
321		 * Some archs do not like unaligned data, so
322		 * we move data down in the first mbuf.
323		 */
324		if (mtod(m, vm_offset_t) & 3) {
325			KASSERT(hlen >= 3, ("if_simloop: hlen too small"));
326			bcopy(m->m_data,
327			    (char *)(mtod(m, vm_offset_t)
328				- (mtod(m, vm_offset_t) & 3)),
329			    m->m_len);
330			m->m_data -= (mtod(m,vm_offset_t) & 3);
331		}
332#endif
333	}
334
335	/* Deliver to upper layer protocol */
336	switch (af) {
337#ifdef INET
338	case AF_INET:
339		isr = NETISR_IP;
340		break;
341#endif
342#ifdef INET6
343	case AF_INET6:
344		m->m_flags |= M_LOOP;
345		isr = NETISR_IPV6;
346		break;
347#endif
348#ifdef IPX
349	case AF_IPX:
350		isr = NETISR_IPX;
351		break;
352#endif
353#ifdef NETATALK
354	case AF_APPLETALK:
355		isr = NETISR_ATALK2;
356		break;
357#endif
358	default:
359		printf("if_simloop: can't handle af=%d\n", af);
360		m_freem(m);
361		return (EAFNOSUPPORT);
362	}
363	ifp->if_ipackets++;
364	ifp->if_ibytes += m->m_pkthdr.len;
365	netisr_queue(isr, m);	/* mbuf is free'd on failure. */
366	return (0);
367}
368
369/* ARGSUSED */
370static void
371lortrequest(int cmd, struct rtentry *rt, struct rt_addrinfo *info)
372{
373
374	RT_LOCK_ASSERT(rt);
375	rt->rt_rmx.rmx_mtu = rt->rt_ifp->if_mtu;
376}
377
378/*
379 * Process an ioctl request.
380 */
381/* ARGSUSED */
382int
383loioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
384{
385	struct ifaddr *ifa;
386	struct ifreq *ifr = (struct ifreq *)data;
387	int error = 0, mask;
388
389	switch (cmd) {
390	case SIOCSIFADDR:
391		ifp->if_flags |= IFF_UP;
392		ifp->if_drv_flags |= IFF_DRV_RUNNING;
393		ifa = (struct ifaddr *)data;
394		ifa->ifa_rtrequest = lortrequest;
395		/*
396		 * Everything else is done at a higher level.
397		 */
398		break;
399
400	case SIOCADDMULTI:
401	case SIOCDELMULTI:
402		if (ifr == 0) {
403			error = EAFNOSUPPORT;		/* XXX */
404			break;
405		}
406		switch (ifr->ifr_addr.sa_family) {
407
408#ifdef INET
409		case AF_INET:
410			break;
411#endif
412#ifdef INET6
413		case AF_INET6:
414			break;
415#endif
416
417		default:
418			error = EAFNOSUPPORT;
419			break;
420		}
421		break;
422
423	case SIOCSIFMTU:
424		ifp->if_mtu = ifr->ifr_mtu;
425		break;
426
427	case SIOCSIFFLAGS:
428		break;
429
430	case SIOCSIFCAP:
431		mask = ifp->if_capenable ^ ifr->ifr_reqcap;
432		if ((mask & IFCAP_RXCSUM) != 0)
433			ifp->if_capenable ^= IFCAP_RXCSUM;
434		if ((mask & IFCAP_TXCSUM) != 0)
435			ifp->if_capenable ^= IFCAP_TXCSUM;
436		if (ifp->if_capenable & IFCAP_TXCSUM)
437			ifp->if_hwassist = LO_CSUM_FEATURES;
438		else
439			ifp->if_hwassist = 0;
440		break;
441
442	default:
443		error = EINVAL;
444	}
445	return (error);
446}
447