if_loop.c revision 85074 
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 * 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 *	@(#)if_loop.c	8.2 (Berkeley) 1/9/95
34 * $FreeBSD: head/sys/net/if_loop.c 85074 2001-10-17 18:07:05Z ru $
35 */
36
37/*
38 * Loopback interface driver for protocol testing and timing.
39 */
40
41#include "opt_atalk.h"
42#include "opt_inet.h"
43#include "opt_inet6.h"
44#include "opt_ipx.h"
45
46#include <sys/param.h>
47#include <sys/systm.h>
48#include <sys/kernel.h>
49#include <sys/malloc.h>
50#include <sys/mbuf.h>
51#include <sys/module.h>
52#include <sys/socket.h>
53#include <sys/sockio.h>
54#include <sys/sysctl.h>
55
56#include <net/if.h>
57#include <net/if_types.h>
58#include <net/netisr.h>
59#include <net/route.h>
60#include <net/bpf.h>
61#include <net/bpfdesc.h>
62
63#ifdef	INET
64#include <netinet/in.h>
65#include <netinet/in_var.h>
66#endif
67
68#ifdef IPX
69#include <netipx/ipx.h>
70#include <netipx/ipx_if.h>
71#endif
72
73#ifdef INET6
74#ifndef INET
75#include <netinet/in.h>
76#endif
77#include <netinet6/in6_var.h>
78#include <netinet/ip6.h>
79#endif
80
81#ifdef NS
82#include <netns/ns.h>
83#include <netns/ns_if.h>
84#endif
85
86#ifdef NETATALK
87#include <netatalk/at.h>
88#include <netatalk/at_var.h>
89#endif
90
91int loioctl __P((struct ifnet *, u_long, caddr_t));
92static void lortrequest __P((int, struct rtentry *, struct rt_addrinfo *));
93
94int looutput __P((struct ifnet *ifp,
95		struct mbuf *m, struct sockaddr *dst, struct rtentry *rt));
96
97#ifdef TINY_LOMTU
98#define	LOMTU	(1024+512)
99#elif defined(LARGE_LOMTU)
100#define LOMTU	131072
101#else
102#define LOMTU	16384
103#endif
104
105static int nloop = 1;
106
107struct ifnet *loif;			/* Used externally */
108
109static MALLOC_DEFINE(M_LO, "lo", "Loopback Interface");
110
111struct lo_softc {
112	struct	ifnet sc_if;		/* network-visible interface */
113        LIST_ENTRY(lo_softc) sc_next;
114};
115static LIST_HEAD(lo_list, lo_softc) lo_list;
116
117static void
118locreate(int unit)
119{
120	struct lo_softc *sc;
121
122	MALLOC(sc, struct lo_softc *, sizeof(*sc), M_LO, M_WAITOK | M_ZERO);
123
124	sc->sc_if.if_name = "lo";
125	sc->sc_if.if_unit = unit;
126	sc->sc_if.if_mtu = LOMTU;
127	sc->sc_if.if_flags = IFF_LOOPBACK | IFF_MULTICAST;
128	sc->sc_if.if_ioctl = loioctl;
129	sc->sc_if.if_output = looutput;
130	sc->sc_if.if_type = IFT_LOOP;
131	sc->sc_if.if_snd.ifq_maxlen = ifqmaxlen;
132	if_attach(&sc->sc_if);
133	bpfattach(&sc->sc_if, DLT_NULL, sizeof(u_int));
134	LIST_INSERT_HEAD(&lo_list, sc, sc_next);
135	if (loif == NULL)
136		loif = &sc->sc_if;
137}
138
139static void
140lodestroy(struct lo_softc *sc)
141{
142	bpfdetach(&sc->sc_if);
143	if_detach(&sc->sc_if);
144	LIST_REMOVE(sc, sc_next);
145	FREE(sc, M_LO);
146}
147
148
149static int
150sysctl_net_nloop(SYSCTL_HANDLER_ARGS)
151{
152	int newnloop;
153	int error;
154
155	newnloop = nloop;
156
157	error = sysctl_handle_opaque(oidp, &newnloop, sizeof newnloop, req);
158	if (error || !req->newptr)
159		return (error);
160
161	if (newnloop < 1)
162		return (EINVAL);
163	while (newnloop > nloop) {
164		locreate(nloop);
165		nloop++;
166	}
167	while (newnloop < nloop) {
168		lodestroy(LIST_FIRST(&lo_list));
169		nloop--;
170	}
171	return (0);
172}
173SYSCTL_PROC(_net, OID_AUTO, nloop, CTLTYPE_INT | CTLFLAG_RW,
174	    0, 0, sysctl_net_nloop, "I", "");
175
176static int
177loop_modevent(module_t mod, int type, void *data)
178{
179	int i;
180
181	switch (type) {
182	case MOD_LOAD:
183		TUNABLE_INT_FETCH("net.nloop", &nloop);
184		if (nloop < 1)			/* sanity check */
185			nloop = 1;
186		for (i = 0; i < nloop; i++)
187			locreate(i);
188		break;
189	case MOD_UNLOAD:
190		printf("loop module unload - not possible for this module type\n");
191		return EINVAL;
192	}
193	return 0;
194}
195
196static moduledata_t loop_mod = {
197	"loop",
198	loop_modevent,
199	0
200};
201
202DECLARE_MODULE(loop, loop_mod, SI_SUB_PSEUDO, SI_ORDER_ANY);
203
204int
205looutput(ifp, m, dst, rt)
206	struct ifnet *ifp;
207	register struct mbuf *m;
208	struct sockaddr *dst;
209	register struct rtentry *rt;
210{
211	if ((m->m_flags & M_PKTHDR) == 0)
212		panic("looutput no HDR");
213
214	if (rt && rt->rt_flags & (RTF_REJECT|RTF_BLACKHOLE)) {
215		m_freem(m);
216		return (rt->rt_flags & RTF_BLACKHOLE ? 0 :
217		        rt->rt_flags & RTF_HOST ? EHOSTUNREACH : ENETUNREACH);
218	}
219	/*
220	 * KAME requires that the packet to be contiguous on the
221	 * mbuf.  We need to make that sure.
222	 * this kind of code should be avoided.
223	 * XXX: fails to join if interface MTU > MCLBYTES.  jumbogram?
224	 */
225	if (m && m->m_next != NULL && m->m_pkthdr.len < MCLBYTES) {
226		struct mbuf *n;
227
228		MGETHDR(n, M_DONTWAIT, MT_HEADER);
229		if (!n)
230			goto contiguousfail;
231		MCLGET(n, M_DONTWAIT);
232		if (! (n->m_flags & M_EXT)) {
233			m_freem(n);
234			goto contiguousfail;
235		}
236
237		m_copydata(m, 0, m->m_pkthdr.len, mtod(n, caddr_t));
238		n->m_pkthdr = m->m_pkthdr;
239		n->m_len = m->m_pkthdr.len;
240		n->m_pkthdr.aux = m->m_pkthdr.aux;
241		m->m_pkthdr.aux = (struct mbuf *)NULL;
242		m_freem(m);
243		m = n;
244	}
245	if (0) {
246contiguousfail:
247		printf("looutput: mbuf allocation failed\n");
248	}
249
250	ifp->if_opackets++;
251	ifp->if_obytes += m->m_pkthdr.len;
252#if 1	/* XXX */
253	switch (dst->sa_family) {
254	case AF_INET:
255	case AF_INET6:
256	case AF_IPX:
257	case AF_NS:
258	case AF_APPLETALK:
259		break;
260	default:
261		printf("looutput: af=%d unexpected\n", dst->sa_family);
262		m_freem(m);
263		return (EAFNOSUPPORT);
264	}
265#endif
266	return(if_simloop(ifp, m, dst->sa_family, 0));
267}
268
269/*
270 * if_simloop()
271 *
272 * This function is to support software emulation of hardware loopback,
273 * i.e., for interfaces with the IFF_SIMPLEX attribute. Since they can't
274 * hear their own broadcasts, we create a copy of the packet that we
275 * would normally receive via a hardware loopback.
276 *
277 * This function expects the packet to include the media header of length hlen.
278 */
279
280int
281if_simloop(ifp, m, af, hlen)
282	struct ifnet *ifp;
283	struct mbuf *m;
284	int af;
285	int hlen;
286{
287	int isr;
288	struct ifqueue *inq = 0;
289
290	KASSERT((m->m_flags & M_PKTHDR) != 0, ("if_simloop: no HDR"));
291	m->m_pkthdr.rcvif = ifp;
292
293	/* BPF write needs to be handled specially */
294	if (af == AF_UNSPEC) {
295		KASSERT(m->m_len >= sizeof(int), ("if_simloop: m_len"));
296		af = *(mtod(m, int *));
297		m->m_len -= sizeof(int);
298		m->m_pkthdr.len -= sizeof(int);
299		m->m_data += sizeof(int);
300	}
301
302	/* Let BPF see incoming packet */
303	if (ifp->if_bpf) {
304		struct mbuf m0, *n = m;
305
306		if (ifp->if_bpf->bif_dlt == DLT_NULL) {
307			/*
308			 * We need to prepend the address family as
309			 * a four byte field.  Cons up a dummy header
310			 * to pacify bpf.  This is safe because bpf
311			 * will only read from the mbuf (i.e., it won't
312			 * try to free it or keep a pointer a to it).
313			 */
314			m0.m_next = m;
315			m0.m_len = 4;
316			m0.m_data = (char *)&af;
317			n = &m0;
318		}
319		bpf_mtap(ifp, n);
320	}
321
322	/* Strip away media header */
323	if (hlen > 0) {
324		m_adj(m, hlen);
325#if defined(__alpha__) || defined(__ia64__)
326		/* The alpha doesn't like unaligned data.
327		 * We move data down in the first mbuf */
328		if (mtod(m, vm_offset_t) & 3) {
329			KASSERT(hlen >= 3, ("if_simloop: hlen too small"));
330			bcopy(m->m_data,
331			    (char *)(mtod(m, vm_offset_t)
332				- (mtod(m, vm_offset_t) & 3)),
333			    m->m_len);
334			mtod(m,vm_offset_t) -= (mtod(m, vm_offset_t) & 3);
335		}
336#endif
337	}
338
339	/* Deliver to upper layer protocol */
340	switch (af) {
341#ifdef INET
342	case AF_INET:
343		inq = &ipintrq;
344		isr = NETISR_IP;
345		break;
346#endif
347#ifdef INET6
348	case AF_INET6:
349		m->m_flags |= M_LOOP;
350		inq = &ip6intrq;
351		isr = NETISR_IPV6;
352		break;
353#endif
354#ifdef IPX
355	case AF_IPX:
356		inq = &ipxintrq;
357		isr = NETISR_IPX;
358		break;
359#endif
360#ifdef NS
361	case AF_NS:
362		inq = &nsintrq;
363		isr = NETISR_NS;
364		break;
365#endif
366#ifdef NETATALK
367	case AF_APPLETALK:
368	        inq = &atintrq2;
369		isr = NETISR_ATALK;
370		break;
371#endif
372	default:
373		printf("if_simloop: can't handle af=%d\n", af);
374		m_freem(m);
375		return (EAFNOSUPPORT);
376	}
377	ifp->if_ipackets++;
378	ifp->if_ibytes += m->m_pkthdr.len;
379	(void) IF_HANDOFF(inq, m, NULL);
380	schednetisr(isr);
381	return (0);
382}
383
384/* ARGSUSED */
385static void
386lortrequest(cmd, rt, info)
387	int cmd;
388	struct rtentry *rt;
389	struct rt_addrinfo *info;
390{
391	if (rt) {
392		rt->rt_rmx.rmx_mtu = rt->rt_ifp->if_mtu; /* for ISO */
393		/*
394		 * For optimal performance, the send and receive buffers
395		 * should be at least twice the MTU plus a little more for
396		 * overhead.
397		 */
398		rt->rt_rmx.rmx_recvpipe =
399			rt->rt_rmx.rmx_sendpipe = 3 * LOMTU;
400	}
401}
402
403/*
404 * Process an ioctl request.
405 */
406/* ARGSUSED */
407int
408loioctl(ifp, cmd, data)
409	register struct ifnet *ifp;
410	u_long cmd;
411	caddr_t data;
412{
413	register struct ifaddr *ifa;
414	register struct ifreq *ifr = (struct ifreq *)data;
415	register int error = 0;
416
417	switch (cmd) {
418
419	case SIOCSIFADDR:
420		ifp->if_flags |= IFF_UP | IFF_RUNNING;
421		ifa = (struct ifaddr *)data;
422		ifa->ifa_rtrequest = lortrequest;
423		/*
424		 * Everything else is done at a higher level.
425		 */
426		break;
427
428	case SIOCADDMULTI:
429	case SIOCDELMULTI:
430		if (ifr == 0) {
431			error = EAFNOSUPPORT;		/* XXX */
432			break;
433		}
434		switch (ifr->ifr_addr.sa_family) {
435
436#ifdef INET
437		case AF_INET:
438			break;
439#endif
440#ifdef INET6
441		case AF_INET6:
442			break;
443#endif
444
445		default:
446			error = EAFNOSUPPORT;
447			break;
448		}
449		break;
450
451	case SIOCSIFMTU:
452		ifp->if_mtu = ifr->ifr_mtu;
453		break;
454
455	case SIOCSIFFLAGS:
456		break;
457
458	default:
459		error = EINVAL;
460	}
461	return (error);
462}
463