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 * $FreeBSD$
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
42#include <sys/param.h>
43#include <sys/systm.h>
44#include <sys/kernel.h>
45#include <sys/mbuf.h>
46#include <sys/module.h>
47#include <machine/bus.h>
48#include <sys/rman.h>
49#include <sys/socket.h>
50#include <sys/sockio.h>
51#include <sys/sysctl.h>
52
53#include <net/if.h>
54#include <net/if_clone.h>
55#include <net/if_types.h>
56#include <net/netisr.h>
57#include <net/route.h>
58#include <net/bpf.h>
59#include <net/vnet.h>
60
61#ifdef	INET
62#include <netinet/in.h>
63#include <netinet/in_var.h>
64#endif
65
66#ifdef IPX
67#include <netipx/ipx.h>
68#include <netipx/ipx_if.h>
69#endif
70
71#ifdef INET6
72#ifndef INET
73#include <netinet/in.h>
74#endif
75#include <netinet6/in6_var.h>
76#include <netinet/ip6.h>
77#endif
78
79#ifdef NETATALK
80#include <netatalk/at.h>
81#include <netatalk/at_var.h>
82#endif
83
84#include <security/mac/mac_framework.h>
85
86#ifdef TINY_LOMTU
87#define	LOMTU	(1024+512)
88#elif defined(LARGE_LOMTU)
89#define LOMTU	131072
90#else
91#define LOMTU	16384
92#endif
93
94#define	LO_CSUM_FEATURES	(CSUM_IP | CSUM_TCP | CSUM_UDP | CSUM_SCTP)
95#define	LO_CSUM_FEATURES6	(CSUM_TCP_IPV6 | CSUM_UDP_IPV6 | CSUM_SCTP_IPV6)
96#define	LO_CSUM_SET		(CSUM_DATA_VALID | CSUM_DATA_VALID_IPV6 | \
97				    CSUM_PSEUDO_HDR | \
98				    CSUM_IP_CHECKED | CSUM_IP_VALID | \
99				    CSUM_SCTP_VALID)
100
101int		loioctl(struct ifnet *, u_long, caddr_t);
102static void	lortrequest(int, struct rtentry *, struct rt_addrinfo *);
103int		looutput(struct ifnet *ifp, struct mbuf *m,
104		    const struct sockaddr *dst, struct route *ro);
105static int	lo_clone_create(struct if_clone *, int, caddr_t);
106static void	lo_clone_destroy(struct ifnet *);
107
108VNET_DEFINE(struct ifnet *, loif);	/* Used externally */
109
110#ifdef VIMAGE
111static VNET_DEFINE(struct if_clone *, lo_cloner);
112#define	V_lo_cloner		VNET(lo_cloner)
113#endif
114
115static struct if_clone *lo_cloner;
116static const char loname[] = "lo";
117
118static void
119lo_clone_destroy(struct ifnet *ifp)
120{
121
122#ifndef VIMAGE
123	/* XXX: destroying lo0 will lead to panics. */
124	KASSERT(V_loif != ifp, ("%s: destroying lo0", __func__));
125#endif
126
127	bpfdetach(ifp);
128	if_detach(ifp);
129	if_free(ifp);
130}
131
132static int
133lo_clone_create(struct if_clone *ifc, int unit, caddr_t params)
134{
135	struct ifnet *ifp;
136
137	ifp = if_alloc(IFT_LOOP);
138	if (ifp == NULL)
139		return (ENOSPC);
140
141	if_initname(ifp, loname, unit);
142	ifp->if_mtu = LOMTU;
143	ifp->if_flags = IFF_LOOPBACK | IFF_MULTICAST;
144	ifp->if_ioctl = loioctl;
145	ifp->if_output = looutput;
146	ifp->if_snd.ifq_maxlen = ifqmaxlen;
147	ifp->if_capabilities = ifp->if_capenable =
148	    IFCAP_HWCSUM | IFCAP_HWCSUM_IPV6;
149	ifp->if_hwassist = LO_CSUM_FEATURES | LO_CSUM_FEATURES6;
150	if_attach(ifp);
151	bpfattach(ifp, DLT_NULL, sizeof(u_int32_t));
152	if (V_loif == NULL)
153		V_loif = ifp;
154
155	return (0);
156}
157
158static void
159vnet_loif_init(const void *unused __unused)
160{
161
162#ifdef VIMAGE
163	lo_cloner = if_clone_simple(loname, lo_clone_create, lo_clone_destroy,
164	    1);
165	V_lo_cloner = lo_cloner;
166#else
167	lo_cloner = if_clone_simple(loname, lo_clone_create, lo_clone_destroy,
168	    1);
169#endif
170}
171VNET_SYSINIT(vnet_loif_init, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY,
172    vnet_loif_init, NULL);
173
174#ifdef VIMAGE
175static void
176vnet_loif_uninit(const void *unused __unused)
177{
178
179	if_clone_detach(V_lo_cloner);
180	V_loif = NULL;
181}
182VNET_SYSUNINIT(vnet_loif_uninit, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY,
183    vnet_loif_uninit, NULL);
184#endif
185
186static int
187loop_modevent(module_t mod, int type, void *data)
188{
189
190	switch (type) {
191	case MOD_LOAD:
192		break;
193
194	case MOD_UNLOAD:
195		printf("loop module unload - not possible for this module type\n");
196		return (EINVAL);
197
198	default:
199		return (EOPNOTSUPP);
200	}
201	return (0);
202}
203
204static moduledata_t loop_mod = {
205	"if_lo",
206	loop_modevent,
207	0
208};
209
210DECLARE_MODULE(if_lo, loop_mod, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY);
211
212int
213looutput(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst,
214    struct route *ro)
215{
216	u_int32_t af;
217	struct rtentry *rt = NULL;
218#ifdef MAC
219	int error;
220#endif
221
222	M_ASSERTPKTHDR(m); /* check if we have the packet header */
223
224	if (ro != NULL)
225		rt = ro->ro_rt;
226#ifdef MAC
227	error = mac_ifnet_check_transmit(ifp, m);
228	if (error) {
229		m_freem(m);
230		return (error);
231	}
232#endif
233
234	if (rt && rt->rt_flags & (RTF_REJECT|RTF_BLACKHOLE)) {
235		m_freem(m);
236		return (rt->rt_flags & RTF_BLACKHOLE ? 0 :
237		        rt->rt_flags & RTF_HOST ? EHOSTUNREACH : ENETUNREACH);
238	}
239
240	ifp->if_opackets++;
241	ifp->if_obytes += m->m_pkthdr.len;
242
243	/* BPF writes need to be handled specially. */
244	if (dst->sa_family == AF_UNSPEC || dst->sa_family == pseudo_AF_HDRCMPLT)
245		bcopy(dst->sa_data, &af, sizeof(af));
246	else
247		af = dst->sa_family;
248
249#if 1	/* XXX */
250	switch (af) {
251	case AF_INET:
252		if (ifp->if_capenable & IFCAP_RXCSUM) {
253			m->m_pkthdr.csum_data = 0xffff;
254			m->m_pkthdr.csum_flags = LO_CSUM_SET;
255		}
256		m->m_pkthdr.csum_flags &= ~LO_CSUM_FEATURES;
257		break;
258	case AF_INET6:
259#if 0
260		/*
261		 * XXX-BZ for now always claim the checksum is good despite
262		 * any interface flags.   This is a workaround for 9.1-R and
263		 * a proper solution ought to be sought later.
264		 */
265		if (ifp->if_capenable & IFCAP_RXCSUM_IPV6) {
266			m->m_pkthdr.csum_data = 0xffff;
267			m->m_pkthdr.csum_flags = LO_CSUM_SET;
268		}
269#else
270		m->m_pkthdr.csum_data = 0xffff;
271		m->m_pkthdr.csum_flags = LO_CSUM_SET;
272#endif
273		m->m_pkthdr.csum_flags &= ~LO_CSUM_FEATURES6;
274		break;
275	case AF_IPX:
276	case AF_APPLETALK:
277		break;
278	default:
279		printf("looutput: af=%d unexpected\n", af);
280		m_freem(m);
281		return (EAFNOSUPPORT);
282	}
283#endif
284	return (if_simloop(ifp, m, af, 0));
285}
286
287/*
288 * if_simloop()
289 *
290 * This function is to support software emulation of hardware loopback,
291 * i.e., for interfaces with the IFF_SIMPLEX attribute. Since they can't
292 * hear their own broadcasts, we create a copy of the packet that we
293 * would normally receive via a hardware loopback.
294 *
295 * This function expects the packet to include the media header of length hlen.
296 */
297int
298if_simloop(struct ifnet *ifp, struct mbuf *m, int af, int hlen)
299{
300	int isr;
301
302	M_ASSERTPKTHDR(m);
303	m_tag_delete_nonpersistent(m);
304	m->m_pkthdr.rcvif = ifp;
305
306#ifdef MAC
307	mac_ifnet_create_mbuf(ifp, m);
308#endif
309
310	/*
311	 * Let BPF see incoming packet in the following manner:
312	 *  - Emulated packet loopback for a simplex interface
313	 *    (net/if_ethersubr.c)
314	 *	-> passes it to ifp's BPF
315	 *  - IPv4/v6 multicast packet loopback (netinet(6)/ip(6)_output.c)
316	 *	-> not passes it to any BPF
317	 *  - Normal packet loopback from myself to myself (net/if_loop.c)
318	 *	-> passes to lo0's BPF (even in case of IPv6, where ifp!=lo0)
319	 */
320	if (hlen > 0) {
321		if (bpf_peers_present(ifp->if_bpf)) {
322			bpf_mtap(ifp->if_bpf, m);
323		}
324	} else {
325		if (bpf_peers_present(V_loif->if_bpf)) {
326			if ((m->m_flags & M_MCAST) == 0 || V_loif == ifp) {
327				/* XXX beware sizeof(af) != 4 */
328				u_int32_t af1 = af;
329
330				/*
331				 * We need to prepend the address family.
332				 */
333				bpf_mtap2(V_loif->if_bpf, &af1, sizeof(af1), m);
334			}
335		}
336	}
337
338	/* Strip away media header */
339	if (hlen > 0) {
340		m_adj(m, hlen);
341#ifndef __NO_STRICT_ALIGNMENT
342		/*
343		 * Some archs do not like unaligned data, so
344		 * we move data down in the first mbuf.
345		 */
346		if (mtod(m, vm_offset_t) & 3) {
347			KASSERT(hlen >= 3, ("if_simloop: hlen too small"));
348			bcopy(m->m_data,
349			    (char *)(mtod(m, vm_offset_t)
350				- (mtod(m, vm_offset_t) & 3)),
351			    m->m_len);
352			m->m_data -= (mtod(m,vm_offset_t) & 3);
353		}
354#endif
355	}
356
357	/* Deliver to upper layer protocol */
358	switch (af) {
359#ifdef INET
360	case AF_INET:
361		isr = NETISR_IP;
362		break;
363#endif
364#ifdef INET6
365	case AF_INET6:
366		m->m_flags |= M_LOOP;
367		isr = NETISR_IPV6;
368		break;
369#endif
370#ifdef IPX
371	case AF_IPX:
372		isr = NETISR_IPX;
373		break;
374#endif
375#ifdef NETATALK
376	case AF_APPLETALK:
377		isr = NETISR_ATALK2;
378		break;
379#endif
380	default:
381		printf("if_simloop: can't handle af=%d\n", af);
382		m_freem(m);
383		return (EAFNOSUPPORT);
384	}
385	ifp->if_ipackets++;
386	ifp->if_ibytes += m->m_pkthdr.len;
387	netisr_queue(isr, m);	/* mbuf is free'd on failure. */
388	return (0);
389}
390
391/* ARGSUSED */
392static void
393lortrequest(int cmd, struct rtentry *rt, struct rt_addrinfo *info)
394{
395
396	RT_LOCK_ASSERT(rt);
397	rt->rt_mtu = rt->rt_ifp->if_mtu;
398}
399
400/*
401 * Process an ioctl request.
402 */
403/* ARGSUSED */
404int
405loioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
406{
407	struct ifaddr *ifa;
408	struct ifreq *ifr = (struct ifreq *)data;
409	int error = 0, mask;
410
411	switch (cmd) {
412	case SIOCSIFADDR:
413		ifp->if_flags |= IFF_UP;
414		ifp->if_drv_flags |= IFF_DRV_RUNNING;
415		ifa = (struct ifaddr *)data;
416		ifa->ifa_rtrequest = lortrequest;
417		/*
418		 * Everything else is done at a higher level.
419		 */
420		break;
421
422	case SIOCADDMULTI:
423	case SIOCDELMULTI:
424		if (ifr == 0) {
425			error = EAFNOSUPPORT;		/* XXX */
426			break;
427		}
428		switch (ifr->ifr_addr.sa_family) {
429
430#ifdef INET
431		case AF_INET:
432			break;
433#endif
434#ifdef INET6
435		case AF_INET6:
436			break;
437#endif
438
439		default:
440			error = EAFNOSUPPORT;
441			break;
442		}
443		break;
444
445	case SIOCSIFMTU:
446		ifp->if_mtu = ifr->ifr_mtu;
447		break;
448
449	case SIOCSIFFLAGS:
450		break;
451
452	case SIOCSIFCAP:
453		mask = ifp->if_capenable ^ ifr->ifr_reqcap;
454		if ((mask & IFCAP_RXCSUM) != 0)
455			ifp->if_capenable ^= IFCAP_RXCSUM;
456		if ((mask & IFCAP_TXCSUM) != 0)
457			ifp->if_capenable ^= IFCAP_TXCSUM;
458		if ((mask & IFCAP_RXCSUM_IPV6) != 0) {
459#if 0
460			ifp->if_capenable ^= IFCAP_RXCSUM_IPV6;
461#else
462			error = EOPNOTSUPP;
463			break;
464#endif
465		}
466		if ((mask & IFCAP_TXCSUM_IPV6) != 0) {
467#if 0
468			ifp->if_capenable ^= IFCAP_TXCSUM_IPV6;
469#else
470			error = EOPNOTSUPP;
471			break;
472#endif
473		}
474		ifp->if_hwassist = 0;
475		if (ifp->if_capenable & IFCAP_TXCSUM)
476			ifp->if_hwassist = LO_CSUM_FEATURES;
477#if 0
478		if (ifp->if_capenable & IFCAP_TXCSUM_IPV6)
479			ifp->if_hwassist |= LO_CSUM_FEATURES6;
480#endif
481		break;
482
483	default:
484		error = EINVAL;
485	}
486	return (error);
487}
488