if_loop.c revision 1.101 1 /* $NetBSD: if_loop.c,v 1.101 2017/12/19 03:32:35 ozaki-r Exp $ */
2
3 /*
4 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
5 * 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 project 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 PROJECT 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 PROJECT 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
32 /*
33 * Copyright (c) 1982, 1986, 1993
34 * The Regents of the University of California. All rights reserved.
35 *
36 * Redistribution and use in source and binary forms, with or without
37 * modification, are permitted provided that the following conditions
38 * are met:
39 * 1. Redistributions of source code must retain the above copyright
40 * notice, this list of conditions and the following disclaimer.
41 * 2. Redistributions in binary form must reproduce the above copyright
42 * notice, this list of conditions and the following disclaimer in the
43 * documentation and/or other materials provided with the distribution.
44 * 3. Neither the name of the University nor the names of its contributors
45 * may be used to endorse or promote products derived from this software
46 * without specific prior written permission.
47 *
48 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
49 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
50 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
51 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
52 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
53 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
54 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
55 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
56 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
57 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
58 * SUCH DAMAGE.
59 *
60 * @(#)if_loop.c 8.2 (Berkeley) 1/9/95
61 */
62
63 /*
64 * Loopback interface driver for protocol testing and timing.
65 */
66
67 #include <sys/cdefs.h>
68 __KERNEL_RCSID(0, "$NetBSD: if_loop.c,v 1.101 2017/12/19 03:32:35 ozaki-r Exp $");
69
70 #ifdef _KERNEL_OPT
71 #include "opt_inet.h"
72 #include "opt_atalk.h"
73 #include "opt_mbuftrace.h"
74 #include "opt_mpls.h"
75 #include "opt_net_mpsafe.h"
76 #endif
77
78 #include <sys/param.h>
79 #include <sys/systm.h>
80 #include <sys/kernel.h>
81 #include <sys/mbuf.h>
82 #include <sys/socket.h>
83 #include <sys/errno.h>
84 #include <sys/ioctl.h>
85 #include <sys/time.h>
86 #include <sys/device.h>
87 #include <sys/module.h>
88
89 #include <sys/cpu.h>
90
91 #include <net/if.h>
92 #include <net/if_types.h>
93 #include <net/netisr.h>
94 #include <net/route.h>
95
96 #ifdef INET
97 #include <netinet/in.h>
98 #include <netinet/in_systm.h>
99 #include <netinet/in_var.h>
100 #include <netinet/in_offload.h>
101 #include <netinet/ip.h>
102 #endif
103
104 #ifdef INET6
105 #ifndef INET
106 #include <netinet/in.h>
107 #endif
108 #include <netinet6/in6_var.h>
109 #include <netinet6/in6_offload.h>
110 #include <netinet/ip6.h>
111 #endif
112
113 #ifdef MPLS
114 #include <netmpls/mpls.h>
115 #include <netmpls/mpls_var.h>
116 #endif
117
118 #ifdef NETATALK
119 #include <netatalk/at.h>
120 #include <netatalk/at_var.h>
121 #endif
122
123 #include <net/bpf.h>
124
125 #if defined(LARGE_LOMTU)
126 #define LOMTU (131072 + MHLEN + MLEN)
127 #define LOMTU_MAX LOMTU
128 #else
129 #define LOMTU (32768 + MHLEN + MLEN)
130 #define LOMTU_MAX (65536 + MHLEN + MLEN)
131 #endif
132
133 #ifdef ALTQ
134 static void lostart(struct ifnet *);
135 #endif
136
137 static int loop_clone_create(struct if_clone *, int);
138 static int loop_clone_destroy(struct ifnet *);
139
140 static void loop_rtrequest(int, struct rtentry *, const struct rt_addrinfo *);
141
142 static struct if_clone loop_cloner =
143 IF_CLONE_INITIALIZER("lo", loop_clone_create, loop_clone_destroy);
144
145 void
146 loopattach(int n)
147 {
148
149 #ifndef _MODULE
150 loop_clone_create(&loop_cloner, 0); /* lo0 always exists */
151 #endif
152 }
153
154 void
155 loopinit(void)
156 {
157
158 if (lo0ifp != NULL) /* can happen in rump kernel */
159 return;
160
161 #ifdef _MODULE
162 loop_clone_create(&loop_cloner, 0); /* lo0 always exists */
163 #endif
164 if_clone_attach(&loop_cloner);
165 }
166
167 static int
168 loopdetach(void)
169 {
170 /* no detach for now; we don't allow lo0 to be deleted */
171 return EBUSY;
172 }
173
174 static int
175 loop_clone_create(struct if_clone *ifc, int unit)
176 {
177 struct ifnet *ifp;
178 int rv;
179
180 ifp = if_alloc(IFT_LOOP);
181
182 if_initname(ifp, ifc->ifc_name, unit);
183
184 ifp->if_mtu = LOMTU;
185 ifp->if_flags = IFF_LOOPBACK | IFF_MULTICAST;
186 #ifdef NET_MPSAFE
187 ifp->if_extflags = IFEF_MPSAFE;
188 #endif
189 ifp->if_ioctl = loioctl;
190 ifp->if_output = looutput;
191 #ifdef ALTQ
192 ifp->if_start = lostart;
193 #endif
194 ifp->if_type = IFT_LOOP;
195 ifp->if_hdrlen = 0;
196 ifp->if_addrlen = 0;
197 ifp->if_dlt = DLT_NULL;
198 IFQ_SET_READY(&ifp->if_snd);
199 if (unit == 0)
200 lo0ifp = ifp;
201 rv = if_attach(ifp);
202 if (rv != 0) {
203 if_free(ifp);
204 return rv;
205 }
206 if_alloc_sadl(ifp);
207 bpf_attach(ifp, DLT_NULL, sizeof(u_int));
208 #ifdef MBUFTRACE
209 ifp->if_mowner = malloc(sizeof(struct mowner), M_DEVBUF,
210 M_WAITOK | M_ZERO);
211 strlcpy(ifp->if_mowner->mo_name, ifp->if_xname,
212 sizeof(ifp->if_mowner->mo_name));
213 MOWNER_ATTACH(ifp->if_mowner);
214 #endif
215
216 ifp->if_flags |= IFF_RUNNING;
217
218 return (0);
219 }
220
221 static int
222 loop_clone_destroy(struct ifnet *ifp)
223 {
224
225 if (ifp == lo0ifp)
226 return (EPERM);
227
228 ifp->if_flags &= ~IFF_RUNNING;
229
230 #ifdef MBUFTRACE
231 MOWNER_DETACH(ifp->if_mowner);
232 free(ifp->if_mowner, M_DEVBUF);
233 #endif
234
235 bpf_detach(ifp);
236 if_detach(ifp);
237
238 if_free(ifp);
239
240 return (0);
241 }
242
243 int
244 looutput(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst,
245 const struct rtentry *rt)
246 {
247 pktqueue_t *pktq = NULL;
248 struct ifqueue *ifq = NULL;
249 int s, isr = -1;
250 int csum_flags;
251 int error = 0;
252 size_t pktlen;
253
254 MCLAIM(m, ifp->if_mowner);
255
256 KERNEL_LOCK_UNLESS_NET_MPSAFE();
257
258 if ((m->m_flags & M_PKTHDR) == 0)
259 panic("looutput: no header mbuf");
260 if (ifp->if_flags & IFF_LOOPBACK)
261 bpf_mtap_af(ifp, dst->sa_family, m);
262 m_set_rcvif(m, ifp);
263
264 if (rt && rt->rt_flags & (RTF_REJECT|RTF_BLACKHOLE)) {
265 m_freem(m);
266 error = (rt->rt_flags & RTF_BLACKHOLE ? 0 :
267 rt->rt_flags & RTF_HOST ? EHOSTUNREACH : ENETUNREACH);
268 goto out;
269 }
270
271 pktlen = m->m_pkthdr.len;
272 ifp->if_opackets++;
273 ifp->if_obytes += pktlen;
274
275 #ifdef ALTQ
276 /*
277 * ALTQ on the loopback interface is just for debugging. It's
278 * used only for loopback interfaces, not for a simplex interface.
279 */
280 if ((ALTQ_IS_ENABLED(&ifp->if_snd) || TBR_IS_ENABLED(&ifp->if_snd)) &&
281 ifp->if_start == lostart) {
282 /*
283 * If the queueing discipline needs packet classification,
284 * do it before prepending the link headers.
285 */
286 IFQ_CLASSIFY(&ifp->if_snd, m, dst->sa_family);
287
288 M_PREPEND(m, sizeof(uint32_t), M_DONTWAIT);
289 if (m == NULL) {
290 error = ENOBUFS;
291 goto out;
292 }
293 *(mtod(m, uint32_t *)) = dst->sa_family;
294
295 error = if_transmit_lock(ifp, m);
296 goto out;
297 }
298 #endif /* ALTQ */
299
300 m_tag_delete_nonpersistent(m);
301
302 #ifdef MPLS
303 if (rt != NULL && rt_gettag(rt) != NULL &&
304 rt_gettag(rt)->sa_family == AF_MPLS &&
305 (m->m_flags & (M_MCAST | M_BCAST)) == 0) {
306 union mpls_shim msh;
307 msh.s_addr = MPLS_GETSADDR(rt);
308 if (msh.shim.label != MPLS_LABEL_IMPLNULL) {
309 ifq = &mplsintrq;
310 isr = NETISR_MPLS;
311 }
312 }
313 if (isr != NETISR_MPLS)
314 #endif
315 switch (dst->sa_family) {
316
317 #ifdef INET
318 case AF_INET:
319 csum_flags = m->m_pkthdr.csum_flags;
320 KASSERT((csum_flags & ~(M_CSUM_IPv4|M_CSUM_UDPv4)) == 0);
321 if (csum_flags != 0 && IN_LOOPBACK_NEED_CHECKSUM(csum_flags)) {
322 ip_undefer_csum(m, 0, csum_flags);
323 }
324 m->m_pkthdr.csum_flags = 0;
325 pktq = ip_pktq;
326 break;
327 #endif
328 #ifdef INET6
329 case AF_INET6:
330 csum_flags = m->m_pkthdr.csum_flags;
331 KASSERT((csum_flags & ~M_CSUM_UDPv6) == 0);
332 if (csum_flags != 0 &&
333 IN6_LOOPBACK_NEED_CHECKSUM(csum_flags)) {
334 ip6_undefer_csum(m, 0, csum_flags);
335 }
336 m->m_pkthdr.csum_flags = 0;
337 m->m_flags |= M_LOOP;
338 pktq = ip6_pktq;
339 break;
340 #endif
341 #ifdef NETATALK
342 case AF_APPLETALK:
343 ifq = &atintrq2;
344 isr = NETISR_ATALK;
345 break;
346 #endif
347 default:
348 printf("%s: can't handle af%d\n", ifp->if_xname,
349 dst->sa_family);
350 m_freem(m);
351 error = EAFNOSUPPORT;
352 goto out;
353 }
354
355 s = splnet();
356 if (__predict_true(pktq)) {
357 error = 0;
358
359 if (__predict_true(pktq_enqueue(pktq, m, 0))) {
360 ifp->if_ipackets++;
361 ifp->if_ibytes += pktlen;
362 } else {
363 m_freem(m);
364 error = ENOBUFS;
365 }
366 splx(s);
367 goto out;
368 }
369 if (IF_QFULL(ifq)) {
370 IF_DROP(ifq);
371 m_freem(m);
372 splx(s);
373 error = ENOBUFS;
374 goto out;
375 }
376 ifp->if_ipackets++;
377 ifp->if_ibytes += m->m_pkthdr.len;
378 IF_ENQUEUE(ifq, m);
379 schednetisr(isr);
380 splx(s);
381 out:
382 KERNEL_UNLOCK_UNLESS_NET_MPSAFE();
383 return error;
384 }
385
386 #ifdef ALTQ
387 static void
388 lostart(struct ifnet *ifp)
389 {
390 for (;;) {
391 pktqueue_t *pktq = NULL;
392 struct ifqueue *ifq = NULL;
393 struct mbuf *m;
394 size_t pktlen;
395 uint32_t af;
396 int s, isr = 0;
397
398 IFQ_DEQUEUE(&ifp->if_snd, m);
399 if (m == NULL)
400 return;
401
402 af = *(mtod(m, uint32_t *));
403 m_adj(m, sizeof(uint32_t));
404
405 switch (af) {
406 #ifdef INET
407 case AF_INET:
408 pktq = ip_pktq;
409 break;
410 #endif
411 #ifdef INET6
412 case AF_INET6:
413 m->m_flags |= M_LOOP;
414 pktq = ip6_pktq;
415 break;
416 #endif
417 #ifdef NETATALK
418 case AF_APPLETALK:
419 ifq = &atintrq2;
420 isr = NETISR_ATALK;
421 break;
422 #endif
423 default:
424 printf("%s: can't handle af%d\n", ifp->if_xname, af);
425 m_freem(m);
426 return;
427 }
428 pktlen = m->m_pkthdr.len;
429
430 s = splnet();
431 if (__predict_true(pktq)) {
432 if (__predict_false(pktq_enqueue(pktq, m, 0))) {
433 m_freem(m);
434 splx(s);
435 return;
436 }
437 ifp->if_ipackets++;
438 ifp->if_ibytes += pktlen;
439 splx(s);
440 continue;
441 }
442 if (IF_QFULL(ifq)) {
443 IF_DROP(ifq);
444 splx(s);
445 m_freem(m);
446 return;
447 }
448 IF_ENQUEUE(ifq, m);
449 schednetisr(isr);
450 ifp->if_ipackets++;
451 ifp->if_ibytes += pktlen;
452 splx(s);
453 }
454 }
455 #endif /* ALTQ */
456
457 /* ARGSUSED */
458 static void
459 loop_rtrequest(int cmd, struct rtentry *rt,
460 const struct rt_addrinfo *info)
461 {
462
463 if (rt)
464 rt->rt_rmx.rmx_mtu = lo0ifp->if_mtu;
465 }
466
467 /*
468 * Process an ioctl request.
469 */
470 /* ARGSUSED */
471 int
472 loioctl(struct ifnet *ifp, u_long cmd, void *data)
473 {
474 struct ifaddr *ifa;
475 struct ifreq *ifr = data;
476 int error = 0;
477
478 switch (cmd) {
479
480 case SIOCINITIFADDR:
481 ifp->if_flags |= IFF_UP;
482 ifa = (struct ifaddr *)data;
483 if (ifa != NULL)
484 ifa->ifa_rtrequest = loop_rtrequest;
485 /*
486 * Everything else is done at a higher level.
487 */
488 break;
489
490 case SIOCSIFMTU:
491 if ((unsigned)ifr->ifr_mtu > LOMTU_MAX)
492 error = EINVAL;
493 else if ((error = ifioctl_common(ifp, cmd, data)) == ENETRESET){
494 error = 0;
495 }
496 break;
497
498 case SIOCADDMULTI:
499 case SIOCDELMULTI:
500 if (ifr == NULL) {
501 error = EAFNOSUPPORT; /* XXX */
502 break;
503 }
504 switch (ifreq_getaddr(cmd, ifr)->sa_family) {
505
506 #ifdef INET
507 case AF_INET:
508 break;
509 #endif
510 #ifdef INET6
511 case AF_INET6:
512 break;
513 #endif
514
515 default:
516 error = EAFNOSUPPORT;
517 break;
518 }
519 break;
520
521 default:
522 error = ifioctl_common(ifp, cmd, data);
523 }
524 return (error);
525 }
526
527 /*
528 * Module infrastructure
529 */
530 #include "if_module.h"
531
532 IF_MODULE(MODULE_CLASS_DRIVER, loop, "")
533