1/*-
2 * SPDX-License-Identifier: BSD-3-Clause
3 *
4 * Copyright (c) 1982, 1986, 1993
5 *	The Regents of the University of California.  All rights reserved.
6 *
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
9 * are met:
10 * 1. Redistributions of source code must retain the above copyright
11 *    notice, this list of conditions and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 *    notice, this list of conditions and the following disclaimer in the
14 *    documentation and/or other materials provided with the distribution.
15 * 3. Neither the name of the University nor the names of its contributors
16 *    may be used to endorse or promote products derived from this software
17 *    without specific prior written permission.
18 *
19 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 * SUCH DAMAGE.
30 *
31 *	@(#)if_loop.c	8.2 (Berkeley) 1/9/95
32 * $FreeBSD$
33 */
34
35/*
36 * Loopback interface driver for protocol testing and timing.
37 */
38
39#include "opt_inet.h"
40#include "opt_inet6.h"
41#include "opt_rss.h"
42
43#include <sys/param.h>
44#include <sys/systm.h>
45#include <sys/kernel.h>
46#include <sys/mbuf.h>
47#include <sys/module.h>
48#include <machine/bus.h>
49#include <sys/rman.h>
50#include <sys/socket.h>
51#include <sys/sockio.h>
52#include <sys/sysctl.h>
53
54#include <net/if.h>
55#include <net/if_var.h>
56#include <net/if_clone.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/vnet.h>
62
63#ifdef	INET
64#include <netinet/in.h>
65#include <netinet/in_var.h>
66#endif
67
68#ifdef INET6
69#ifndef INET
70#include <netinet/in.h>
71#endif
72#include <netinet6/in6_var.h>
73#include <netinet/ip6.h>
74#endif
75
76#include <security/mac/mac_framework.h>
77
78#ifdef TINY_LOMTU
79#define	LOMTU	(1024+512)
80#elif defined(LARGE_LOMTU)
81#define LOMTU	131072
82#else
83#define LOMTU	16384
84#endif
85
86#define	LO_CSUM_FEATURES	(CSUM_IP | CSUM_TCP | CSUM_UDP | CSUM_SCTP)
87#define	LO_CSUM_FEATURES6	(CSUM_TCP_IPV6 | CSUM_UDP_IPV6 | CSUM_SCTP_IPV6)
88#define	LO_CSUM_SET		(CSUM_DATA_VALID | CSUM_DATA_VALID_IPV6 | \
89				    CSUM_PSEUDO_HDR | \
90				    CSUM_IP_CHECKED | CSUM_IP_VALID | \
91				    CSUM_SCTP_VALID)
92
93int		loioctl(struct ifnet *, u_long, caddr_t);
94int		looutput(struct ifnet *ifp, struct mbuf *m,
95		    const struct sockaddr *dst, struct route *ro);
96static int	lo_clone_create(struct if_clone *, int, caddr_t);
97static void	lo_clone_destroy(struct ifnet *);
98
99VNET_DEFINE(struct ifnet *, loif);	/* Used externally */
100
101#ifdef VIMAGE
102VNET_DEFINE_STATIC(struct if_clone *, lo_cloner);
103#define	V_lo_cloner		VNET(lo_cloner)
104#endif
105
106static struct if_clone *lo_cloner;
107static const char loname[] = "lo";
108
109static void
110lo_clone_destroy(struct ifnet *ifp)
111{
112
113#ifndef VIMAGE
114	/* XXX: destroying lo0 will lead to panics. */
115	KASSERT(V_loif != ifp, ("%s: destroying lo0", __func__));
116#endif
117
118	bpfdetach(ifp);
119	if_detach(ifp);
120	if_free(ifp);
121}
122
123static int
124lo_clone_create(struct if_clone *ifc, int unit, caddr_t params)
125{
126	struct ifnet *ifp;
127
128	ifp = if_alloc(IFT_LOOP);
129	if (ifp == NULL)
130		return (ENOSPC);
131
132	if_initname(ifp, loname, unit);
133	ifp->if_mtu = LOMTU;
134	ifp->if_flags = IFF_LOOPBACK | IFF_MULTICAST;
135	ifp->if_ioctl = loioctl;
136	ifp->if_output = looutput;
137	ifp->if_snd.ifq_maxlen = ifqmaxlen;
138	ifp->if_capabilities = ifp->if_capenable =
139	    IFCAP_HWCSUM | IFCAP_HWCSUM_IPV6 | IFCAP_LINKSTATE;
140	ifp->if_hwassist = LO_CSUM_FEATURES | LO_CSUM_FEATURES6;
141	if_attach(ifp);
142	bpfattach(ifp, DLT_NULL, sizeof(u_int32_t));
143	if (V_loif == NULL)
144		V_loif = ifp;
145
146	return (0);
147}
148
149static void
150vnet_loif_init(const void *unused __unused)
151{
152
153#ifdef VIMAGE
154	lo_cloner = if_clone_simple(loname, lo_clone_create, lo_clone_destroy,
155	    1);
156	V_lo_cloner = lo_cloner;
157#else
158	lo_cloner = if_clone_simple(loname, lo_clone_create, lo_clone_destroy,
159	    1);
160#endif
161}
162VNET_SYSINIT(vnet_loif_init, SI_SUB_PSEUDO, SI_ORDER_ANY,
163    vnet_loif_init, NULL);
164
165#ifdef VIMAGE
166static void
167vnet_loif_uninit(const void *unused __unused)
168{
169
170	if_clone_detach(V_lo_cloner);
171	V_loif = NULL;
172}
173VNET_SYSUNINIT(vnet_loif_uninit, SI_SUB_INIT_IF, SI_ORDER_SECOND,
174    vnet_loif_uninit, NULL);
175#endif
176
177static int
178loop_modevent(module_t mod, int type, void *data)
179{
180
181	switch (type) {
182	case MOD_LOAD:
183		break;
184
185	case MOD_UNLOAD:
186		printf("loop module unload - not possible for this module type\n");
187		return (EINVAL);
188
189	default:
190		return (EOPNOTSUPP);
191	}
192	return (0);
193}
194
195static moduledata_t loop_mod = {
196	"if_lo",
197	loop_modevent,
198	0
199};
200
201DECLARE_MODULE(if_lo, loop_mod, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY);
202
203int
204looutput(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst,
205    struct route *ro)
206{
207	u_int32_t af;
208#ifdef MAC
209	int error;
210#endif
211
212	M_ASSERTPKTHDR(m); /* check if we have the packet header */
213
214#ifdef MAC
215	error = mac_ifnet_check_transmit(ifp, m);
216	if (error) {
217		m_freem(m);
218		return (error);
219	}
220#endif
221
222	if (ro != NULL && ro->ro_flags & (RT_REJECT|RT_BLACKHOLE)) {
223		m_freem(m);
224		return (ro->ro_flags & RT_BLACKHOLE ? 0 : EHOSTUNREACH);
225	}
226
227	if_inc_counter(ifp, IFCOUNTER_OPACKETS, 1);
228	if_inc_counter(ifp, IFCOUNTER_OBYTES, m->m_pkthdr.len);
229
230#ifdef RSS
231	M_HASHTYPE_CLEAR(m);
232#endif
233
234	/* BPF writes need to be handled specially. */
235	if (dst->sa_family == AF_UNSPEC || dst->sa_family == pseudo_AF_HDRCMPLT)
236		bcopy(dst->sa_data, &af, sizeof(af));
237	else
238		af = dst->sa_family;
239
240#if 1	/* XXX */
241	switch (af) {
242	case AF_INET:
243		if (ifp->if_capenable & IFCAP_RXCSUM) {
244			m->m_pkthdr.csum_data = 0xffff;
245			m->m_pkthdr.csum_flags = LO_CSUM_SET;
246		}
247		m->m_pkthdr.csum_flags &= ~LO_CSUM_FEATURES;
248		break;
249	case AF_INET6:
250#if 0
251		/*
252		 * XXX-BZ for now always claim the checksum is good despite
253		 * any interface flags.   This is a workaround for 9.1-R and
254		 * a proper solution ought to be sought later.
255		 */
256		if (ifp->if_capenable & IFCAP_RXCSUM_IPV6) {
257			m->m_pkthdr.csum_data = 0xffff;
258			m->m_pkthdr.csum_flags = LO_CSUM_SET;
259		}
260#else
261		m->m_pkthdr.csum_data = 0xffff;
262		m->m_pkthdr.csum_flags = LO_CSUM_SET;
263#endif
264		m->m_pkthdr.csum_flags &= ~LO_CSUM_FEATURES6;
265		break;
266	default:
267		printf("looutput: af=%d unexpected\n", af);
268		m_freem(m);
269		return (EAFNOSUPPORT);
270	}
271#endif
272	return (if_simloop(ifp, m, af, 0));
273}
274
275/*
276 * if_simloop()
277 *
278 * This function is to support software emulation of hardware loopback,
279 * i.e., for interfaces with the IFF_SIMPLEX attribute. Since they can't
280 * hear their own broadcasts, we create a copy of the packet that we
281 * would normally receive via a hardware loopback.
282 *
283 * This function expects the packet to include the media header of length hlen.
284 */
285int
286if_simloop(struct ifnet *ifp, struct mbuf *m, int af, int hlen)
287{
288	int isr;
289
290	M_ASSERTPKTHDR(m);
291	m_tag_delete_nonpersistent(m);
292	m->m_pkthdr.rcvif = ifp;
293
294#ifdef MAC
295	mac_ifnet_create_mbuf(ifp, m);
296#endif
297
298	/*
299	 * Let BPF see incoming packet in the following manner:
300	 *  - Emulated packet loopback for a simplex interface
301	 *    (net/if_ethersubr.c)
302	 *	-> passes it to ifp's BPF
303	 *  - IPv4/v6 multicast packet loopback (netinet(6)/ip(6)_output.c)
304	 *	-> not passes it to any BPF
305	 *  - Normal packet loopback from myself to myself (net/if_loop.c)
306	 *	-> passes to lo0's BPF (even in case of IPv6, where ifp!=lo0)
307	 */
308	if (hlen > 0) {
309		if (bpf_peers_present(ifp->if_bpf)) {
310			bpf_mtap(ifp->if_bpf, m);
311		}
312	} else {
313		if (bpf_peers_present(V_loif->if_bpf)) {
314			if ((m->m_flags & M_MCAST) == 0 || V_loif == ifp) {
315				/* XXX beware sizeof(af) != 4 */
316				u_int32_t af1 = af;
317
318				/*
319				 * We need to prepend the address family.
320				 */
321				bpf_mtap2(V_loif->if_bpf, &af1, sizeof(af1), m);
322			}
323		}
324	}
325
326	/* Strip away media header */
327	if (hlen > 0) {
328		m_adj(m, hlen);
329#ifndef __NO_STRICT_ALIGNMENT
330		/*
331		 * Some archs do not like unaligned data, so
332		 * we move data down in the first mbuf.
333		 */
334		if (mtod(m, vm_offset_t) & 3) {
335			KASSERT(hlen >= 3, ("if_simloop: hlen too small"));
336			bcopy(m->m_data,
337			    (char *)(mtod(m, vm_offset_t)
338				- (mtod(m, vm_offset_t) & 3)),
339			    m->m_len);
340			m->m_data -= (mtod(m,vm_offset_t) & 3);
341		}
342#endif
343	}
344
345	/* Deliver to upper layer protocol */
346	switch (af) {
347#ifdef INET
348	case AF_INET:
349		isr = NETISR_IP;
350		break;
351#endif
352#ifdef INET6
353	case AF_INET6:
354		m->m_flags |= M_LOOP;
355		isr = NETISR_IPV6;
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	if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1);
364	if_inc_counter(ifp, IFCOUNTER_IBYTES, m->m_pkthdr.len);
365	netisr_queue(isr, m);	/* mbuf is free'd on failure. */
366	return (0);
367}
368
369/*
370 * Process an ioctl request.
371 */
372/* ARGSUSED */
373int
374loioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
375{
376	struct ifreq *ifr = (struct ifreq *)data;
377	int error = 0, mask;
378
379	switch (cmd) {
380	case SIOCSIFADDR:
381		ifp->if_flags |= IFF_UP;
382		ifp->if_drv_flags |= IFF_DRV_RUNNING;
383		if_link_state_change(ifp, LINK_STATE_UP);
384		/*
385		 * Everything else is done at a higher level.
386		 */
387		break;
388
389	case SIOCADDMULTI:
390	case SIOCDELMULTI:
391		if (ifr == NULL) {
392			error = EAFNOSUPPORT;		/* XXX */
393			break;
394		}
395		switch (ifr->ifr_addr.sa_family) {
396#ifdef INET
397		case AF_INET:
398			break;
399#endif
400#ifdef INET6
401		case AF_INET6:
402			break;
403#endif
404
405		default:
406			error = EAFNOSUPPORT;
407			break;
408		}
409		break;
410
411	case SIOCSIFMTU:
412		ifp->if_mtu = ifr->ifr_mtu;
413		break;
414
415	case SIOCSIFFLAGS:
416		if_link_state_change(ifp, (ifp->if_flags & IFF_UP) ?
417		    LINK_STATE_UP: LINK_STATE_DOWN);
418		break;
419
420	case SIOCSIFCAP:
421		mask = ifp->if_capenable ^ ifr->ifr_reqcap;
422		if ((mask & IFCAP_RXCSUM) != 0)
423			ifp->if_capenable ^= IFCAP_RXCSUM;
424		if ((mask & IFCAP_TXCSUM) != 0)
425			ifp->if_capenable ^= IFCAP_TXCSUM;
426		if ((mask & IFCAP_RXCSUM_IPV6) != 0) {
427#if 0
428			ifp->if_capenable ^= IFCAP_RXCSUM_IPV6;
429#else
430			error = EOPNOTSUPP;
431			break;
432#endif
433		}
434		if ((mask & IFCAP_TXCSUM_IPV6) != 0) {
435#if 0
436			ifp->if_capenable ^= IFCAP_TXCSUM_IPV6;
437#else
438			error = EOPNOTSUPP;
439			break;
440#endif
441		}
442		ifp->if_hwassist = 0;
443		if (ifp->if_capenable & IFCAP_TXCSUM)
444			ifp->if_hwassist = LO_CSUM_FEATURES;
445#if 0
446		if (ifp->if_capenable & IFCAP_TXCSUM_IPV6)
447			ifp->if_hwassist |= LO_CSUM_FEATURES6;
448#endif
449		break;
450
451	default:
452		error = EINVAL;
453	}
454	return (error);
455}
456