if_loop.c revision 195727 
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 195727 2009-07-16 21:13:04Z rwatson $
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#include <sys/vimage.h>
53
54#include <net/if.h>
55#include <net/if_clone.h>
56#include <net/if_types.h>
57#include <net/netisr.h>
58#include <net/route.h>
59#include <net/bpf.h>
60#include <net/vnet.h>
61
62#ifdef	INET
63#include <netinet/in.h>
64#include <netinet/in_var.h>
65#endif
66
67#ifdef IPX
68#include <netipx/ipx.h>
69#include <netipx/ipx_if.h>
70#endif
71
72#ifdef INET6
73#ifndef INET
74#include <netinet/in.h>
75#endif
76#include <netinet6/in6_var.h>
77#include <netinet/ip6.h>
78#endif
79
80#ifdef NETATALK
81#include <netatalk/at.h>
82#include <netatalk/at_var.h>
83#endif
84
85#include <security/mac/mac_framework.h>
86
87#ifdef TINY_LOMTU
88#define	LOMTU	(1024+512)
89#elif defined(LARGE_LOMTU)
90#define LOMTU	131072
91#else
92#define LOMTU	16384
93#endif
94
95#define	LO_CSUM_FEATURES	(CSUM_IP | CSUM_TCP | CSUM_UDP | CSUM_SCTP)
96#define	LO_CSUM_SET		(CSUM_DATA_VALID | CSUM_PSEUDO_HDR | \
97				    CSUM_IP_CHECKED | CSUM_IP_VALID | \
98				    CSUM_SCTP_VALID)
99
100int		loioctl(struct ifnet *, u_long, caddr_t);
101static void	lortrequest(int, struct rtentry *, struct rt_addrinfo *);
102int		looutput(struct ifnet *ifp, struct mbuf *m,
103		    struct sockaddr *dst, struct route *ro);
104static int	lo_clone_create(struct if_clone *, int, caddr_t);
105static void	lo_clone_destroy(struct ifnet *);
106static int	vnet_loif_iattach(const void *);
107#ifdef VIMAGE
108static int	vnet_loif_idetach(const void *);
109#endif
110
111VNET_DEFINE(struct ifnet *, loif);	/* Used externally */
112
113#ifdef VIMAGE
114static VNET_DEFINE(struct ifc_simple_data *, lo_cloner_data);
115static VNET_DEFINE(struct if_clone *, lo_cloner);
116#define	V_lo_cloner_data	VNET(lo_cloner_data)
117#define	V_lo_cloner		VNET(lo_cloner)
118
119MALLOC_DEFINE(M_LO_CLONER, "lo_cloner", "lo_cloner");
120#endif
121
122#ifdef VIMAGE
123static const vnet_modinfo_t vnet_loif_modinfo = {
124	.vmi_id		= VNET_MOD_LOIF,
125	.vmi_dependson	= VNET_MOD_IF_CLONE,
126	.vmi_name	= "loif",
127	.vmi_iattach	= vnet_loif_iattach,
128	.vmi_idetach	= vnet_loif_idetach
129};
130#endif
131
132IFC_SIMPLE_DECLARE(lo, 1);
133
134static void
135lo_clone_destroy(struct ifnet *ifp)
136{
137
138#ifndef VIMAGE
139	/* XXX: destroying lo0 will lead to panics. */
140	KASSERT(V_loif != ifp, ("%s: destroying lo0", __func__));
141#endif
142
143	bpfdetach(ifp);
144	if_detach(ifp);
145	if_free(ifp);
146}
147
148static int
149lo_clone_create(struct if_clone *ifc, int unit, caddr_t params)
150{
151	struct ifnet *ifp;
152
153	ifp = if_alloc(IFT_LOOP);
154	if (ifp == NULL)
155		return (ENOSPC);
156
157	if_initname(ifp, ifc->ifc_name, unit);
158	ifp->if_mtu = LOMTU;
159	ifp->if_flags = IFF_LOOPBACK | IFF_MULTICAST;
160	ifp->if_ioctl = loioctl;
161	ifp->if_output = looutput;
162	ifp->if_snd.ifq_maxlen = ifqmaxlen;
163	ifp->if_capabilities = ifp->if_capenable = IFCAP_HWCSUM;
164	ifp->if_hwassist = LO_CSUM_FEATURES;
165	if_attach(ifp);
166	bpfattach(ifp, DLT_NULL, sizeof(u_int32_t));
167	if (V_loif == NULL)
168		V_loif = ifp;
169
170	return (0);
171}
172
173static int
174vnet_loif_iattach(const void *unused __unused)
175{
176
177#ifdef VIMAGE
178	V_lo_cloner = malloc(sizeof(*V_lo_cloner), M_LO_CLONER,
179	    M_WAITOK | M_ZERO);
180	V_lo_cloner_data = malloc(sizeof(*V_lo_cloner_data), M_LO_CLONER,
181	    M_WAITOK | M_ZERO);
182	bcopy(&lo_cloner, V_lo_cloner, sizeof(*V_lo_cloner));
183	bcopy(lo_cloner.ifc_data, V_lo_cloner_data, sizeof(*V_lo_cloner_data));
184	V_lo_cloner->ifc_data = V_lo_cloner_data;
185	if_clone_attach(V_lo_cloner);
186#else
187	if_clone_attach(&lo_cloner);
188#endif
189	return (0);
190}
191
192#ifdef VIMAGE
193static int
194vnet_loif_idetach(const void *unused __unused)
195{
196
197	if_clone_detach(V_lo_cloner);
198	free(V_lo_cloner, M_LO_CLONER);
199	free(V_lo_cloner_data, M_LO_CLONER);
200	V_loif = NULL;
201
202	return (0);
203}
204#endif
205
206static int
207loop_modevent(module_t mod, int type, void *data)
208{
209
210	switch (type) {
211	case MOD_LOAD:
212#ifdef VIMAGE
213		vnet_mod_register(&vnet_loif_modinfo);
214#else
215		vnet_loif_iattach(NULL);
216#endif
217		break;
218
219	case MOD_UNLOAD:
220		printf("loop module unload - not possible for this module type\n");
221		return (EINVAL);
222
223	default:
224		return (EOPNOTSUPP);
225	}
226	return (0);
227}
228
229static moduledata_t loop_mod = {
230	"loop",
231	loop_modevent,
232	0
233};
234
235DECLARE_MODULE(loop, loop_mod, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY);
236
237int
238looutput(struct ifnet *ifp, struct mbuf *m, struct sockaddr *dst,
239    struct route *ro)
240{
241	u_int32_t af;
242	struct rtentry *rt = NULL;
243#ifdef MAC
244	int error;
245#endif
246
247	M_ASSERTPKTHDR(m); /* check if we have the packet header */
248
249	if (ro != NULL)
250		rt = ro->ro_rt;
251#ifdef MAC
252	error = mac_ifnet_check_transmit(ifp, m);
253	if (error) {
254		m_freem(m);
255		return (error);
256	}
257#endif
258
259	if (rt && rt->rt_flags & (RTF_REJECT|RTF_BLACKHOLE)) {
260		m_freem(m);
261		return (rt->rt_flags & RTF_BLACKHOLE ? 0 :
262		        rt->rt_flags & RTF_HOST ? EHOSTUNREACH : ENETUNREACH);
263	}
264
265	ifp->if_opackets++;
266	ifp->if_obytes += m->m_pkthdr.len;
267
268	/* BPF writes need to be handled specially. */
269	if (dst->sa_family == AF_UNSPEC) {
270		bcopy(dst->sa_data, &af, sizeof(af));
271		dst->sa_family = af;
272	}
273
274#if 1	/* XXX */
275	switch (dst->sa_family) {
276	case AF_INET:
277		if (ifp->if_capenable & IFCAP_RXCSUM) {
278			m->m_pkthdr.csum_data = 0xffff;
279			m->m_pkthdr.csum_flags = LO_CSUM_SET;
280		}
281		m->m_pkthdr.csum_flags &= ~LO_CSUM_FEATURES;
282	case AF_INET6:
283	case AF_IPX:
284	case AF_APPLETALK:
285		break;
286	default:
287		printf("looutput: af=%d unexpected\n", dst->sa_family);
288		m_freem(m);
289		return (EAFNOSUPPORT);
290	}
291#endif
292	return (if_simloop(ifp, m, dst->sa_family, 0));
293}
294
295/*
296 * if_simloop()
297 *
298 * This function is to support software emulation of hardware loopback,
299 * i.e., for interfaces with the IFF_SIMPLEX attribute. Since they can't
300 * hear their own broadcasts, we create a copy of the packet that we
301 * would normally receive via a hardware loopback.
302 *
303 * This function expects the packet to include the media header of length hlen.
304 */
305int
306if_simloop(struct ifnet *ifp, struct mbuf *m, int af, int hlen)
307{
308	int isr;
309
310	M_ASSERTPKTHDR(m);
311	m_tag_delete_nonpersistent(m);
312	m->m_pkthdr.rcvif = ifp;
313
314#ifdef MAC
315	mac_ifnet_create_mbuf(ifp, m);
316#endif
317
318	/*
319	 * Let BPF see incoming packet in the following manner:
320	 *  - Emulated packet loopback for a simplex interface
321	 *    (net/if_ethersubr.c)
322	 *	-> passes it to ifp's BPF
323	 *  - IPv4/v6 multicast packet loopback (netinet(6)/ip(6)_output.c)
324	 *	-> not passes it to any BPF
325	 *  - Normal packet loopback from myself to myself (net/if_loop.c)
326	 *	-> passes to lo0's BPF (even in case of IPv6, where ifp!=lo0)
327	 */
328	if (hlen > 0) {
329		if (bpf_peers_present(ifp->if_bpf)) {
330			bpf_mtap(ifp->if_bpf, m);
331		}
332	} else {
333		if (bpf_peers_present(V_loif->if_bpf)) {
334			if ((m->m_flags & M_MCAST) == 0 || V_loif == ifp) {
335				/* XXX beware sizeof(af) != 4 */
336				u_int32_t af1 = af;
337
338				/*
339				 * We need to prepend the address family.
340				 */
341				bpf_mtap2(V_loif->if_bpf, &af1, sizeof(af1), m);
342			}
343		}
344	}
345
346	/* Strip away media header */
347	if (hlen > 0) {
348		m_adj(m, hlen);
349#ifndef __NO_STRICT_ALIGNMENT
350		/*
351		 * Some archs do not like unaligned data, so
352		 * we move data down in the first mbuf.
353		 */
354		if (mtod(m, vm_offset_t) & 3) {
355			KASSERT(hlen >= 3, ("if_simloop: hlen too small"));
356			bcopy(m->m_data,
357			    (char *)(mtod(m, vm_offset_t)
358				- (mtod(m, vm_offset_t) & 3)),
359			    m->m_len);
360			m->m_data -= (mtod(m,vm_offset_t) & 3);
361		}
362#endif
363	}
364
365	/* Deliver to upper layer protocol */
366	switch (af) {
367#ifdef INET
368	case AF_INET:
369		isr = NETISR_IP;
370		break;
371#endif
372#ifdef INET6
373	case AF_INET6:
374		m->m_flags |= M_LOOP;
375		isr = NETISR_IPV6;
376		break;
377#endif
378#ifdef IPX
379	case AF_IPX:
380		isr = NETISR_IPX;
381		break;
382#endif
383#ifdef NETATALK
384	case AF_APPLETALK:
385		isr = NETISR_ATALK2;
386		break;
387#endif
388	default:
389		printf("if_simloop: can't handle af=%d\n", af);
390		m_freem(m);
391		return (EAFNOSUPPORT);
392	}
393	ifp->if_ipackets++;
394	ifp->if_ibytes += m->m_pkthdr.len;
395	netisr_queue(isr, m);	/* mbuf is free'd on failure. */
396	return (0);
397}
398
399/* ARGSUSED */
400static void
401lortrequest(int cmd, struct rtentry *rt, struct rt_addrinfo *info)
402{
403
404	RT_LOCK_ASSERT(rt);
405	rt->rt_rmx.rmx_mtu = rt->rt_ifp->if_mtu;
406}
407
408/*
409 * Process an ioctl request.
410 */
411/* ARGSUSED */
412int
413loioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
414{
415	struct ifaddr *ifa;
416	struct ifreq *ifr = (struct ifreq *)data;
417	int error = 0, mask;
418
419	switch (cmd) {
420	case SIOCSIFADDR:
421		ifp->if_flags |= IFF_UP;
422		ifp->if_drv_flags |= IFF_DRV_RUNNING;
423		ifa = (struct ifaddr *)data;
424		ifa->ifa_rtrequest = lortrequest;
425		/*
426		 * Everything else is done at a higher level.
427		 */
428		break;
429
430	case SIOCADDMULTI:
431	case SIOCDELMULTI:
432		if (ifr == 0) {
433			error = EAFNOSUPPORT;		/* XXX */
434			break;
435		}
436		switch (ifr->ifr_addr.sa_family) {
437
438#ifdef INET
439		case AF_INET:
440			break;
441#endif
442#ifdef INET6
443		case AF_INET6:
444			break;
445#endif
446
447		default:
448			error = EAFNOSUPPORT;
449			break;
450		}
451		break;
452
453	case SIOCSIFMTU:
454		ifp->if_mtu = ifr->ifr_mtu;
455		break;
456
457	case SIOCSIFFLAGS:
458		break;
459
460	case SIOCSIFCAP:
461		mask = ifp->if_capenable ^ ifr->ifr_reqcap;
462		if ((mask & IFCAP_RXCSUM) != 0)
463			ifp->if_capenable ^= IFCAP_RXCSUM;
464		if ((mask & IFCAP_TXCSUM) != 0)
465			ifp->if_capenable ^= IFCAP_TXCSUM;
466		if (ifp->if_capenable & IFCAP_TXCSUM)
467			ifp->if_hwassist = LO_CSUM_FEATURES;
468		else
469			ifp->if_hwassist = 0;
470		break;
471
472	default:
473		error = EINVAL;
474	}
475	return (error);
476}
477