if_loop.c revision 1.96 1 /* $NetBSD: if_loop.c,v 1.96 2017/10/23 09:32:00 msaitoh 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 1.96 2017/10/23 09:32:00 msaitoh 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 | IFF_RUNNING;
186 ifp->if_extflags = IFEF_OUTPUT_MPSAFE;
187 ifp->if_ioctl = loioctl;
188 ifp->if_output = looutput;
189 #ifdef ALTQ
190 ifp->if_start = lostart;
191 #endif
192 ifp->if_type = IFT_LOOP;
193 ifp->if_hdrlen = 0;
194 ifp->if_addrlen = 0;
195 ifp->if_dlt = DLT_NULL;
196 IFQ_SET_READY(&ifp->if_snd);
197 if (unit == 0)
198 lo0ifp = ifp;
199 rv = if_attach(ifp);
200 if (rv != 0) {
201 if_free(ifp);
202 return rv;
203 }
204 if_alloc_sadl(ifp);
205 bpf_attach(ifp, DLT_NULL, sizeof(u_int));
206 #ifdef MBUFTRACE
207 ifp->if_mowner = malloc(sizeof(struct mowner), M_DEVBUF,
208 M_WAITOK | M_ZERO);
209 strlcpy(ifp->if_mowner->mo_name, ifp->if_xname,
210 sizeof(ifp->if_mowner->mo_name));
211 MOWNER_ATTACH(ifp->if_mowner);
212 #endif
213
214 return (0);
215 }
216
217 static int
218 loop_clone_destroy(struct ifnet *ifp)
219 {
220
221 if (ifp == lo0ifp)
222 return (EPERM);
223
224 #ifdef MBUFTRACE
225 MOWNER_DETACH(ifp->if_mowner);
226 free(ifp->if_mowner, M_DEVBUF);
227 #endif
228
229 bpf_detach(ifp);
230 if_detach(ifp);
231
232 if_free(ifp);
233
234 return (0);
235 }
236
237 int
238 looutput(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst,
239 const struct rtentry *rt)
240 {
241 pktqueue_t *pktq = NULL;
242 struct ifqueue *ifq = NULL;
243 int s, isr = -1;
244 int csum_flags;
245 int error = 0;
246 size_t pktlen;
247
248 MCLAIM(m, ifp->if_mowner);
249
250 KERNEL_LOCK(1, NULL);
251
252 if ((m->m_flags & M_PKTHDR) == 0)
253 panic("looutput: no header mbuf");
254 if (ifp->if_flags & IFF_LOOPBACK)
255 bpf_mtap_af(ifp, dst->sa_family, m);
256 m_set_rcvif(m, ifp);
257
258 if (rt && rt->rt_flags & (RTF_REJECT|RTF_BLACKHOLE)) {
259 m_freem(m);
260 error = (rt->rt_flags & RTF_BLACKHOLE ? 0 :
261 rt->rt_flags & RTF_HOST ? EHOSTUNREACH : ENETUNREACH);
262 goto out;
263 }
264
265 pktlen = m->m_pkthdr.len;
266 ifp->if_opackets++;
267 ifp->if_obytes += pktlen;
268
269 #ifdef ALTQ
270 /*
271 * ALTQ on the loopback interface is just for debugging. It's
272 * used only for loopback interfaces, not for a simplex interface.
273 */
274 if ((ALTQ_IS_ENABLED(&ifp->if_snd) || TBR_IS_ENABLED(&ifp->if_snd)) &&
275 ifp->if_start == lostart) {
276 /*
277 * If the queueing discipline needs packet classification,
278 * do it before prepending the link headers.
279 */
280 IFQ_CLASSIFY(&ifp->if_snd, m, dst->sa_family);
281
282 M_PREPEND(m, sizeof(uint32_t), M_DONTWAIT);
283 if (m == NULL) {
284 error = ENOBUFS;
285 goto out;
286 }
287 *(mtod(m, uint32_t *)) = dst->sa_family;
288
289 error = if_transmit_lock(ifp, m);
290 goto out;
291 }
292 #endif /* ALTQ */
293
294 m_tag_delete_nonpersistent(m);
295
296 #ifdef MPLS
297 if (rt != NULL && rt_gettag(rt) != NULL &&
298 rt_gettag(rt)->sa_family == AF_MPLS &&
299 (m->m_flags & (M_MCAST | M_BCAST)) == 0) {
300 union mpls_shim msh;
301 msh.s_addr = MPLS_GETSADDR(rt);
302 if (msh.shim.label != MPLS_LABEL_IMPLNULL) {
303 ifq = &mplsintrq;
304 isr = NETISR_MPLS;
305 }
306 }
307 if (isr != NETISR_MPLS)
308 #endif
309 switch (dst->sa_family) {
310
311 #ifdef INET
312 case AF_INET:
313 csum_flags = m->m_pkthdr.csum_flags;
314 KASSERT((csum_flags & ~(M_CSUM_IPv4|M_CSUM_UDPv4)) == 0);
315 if (csum_flags != 0 && IN_LOOPBACK_NEED_CHECKSUM(csum_flags)) {
316 ip_undefer_csum(m, 0, csum_flags);
317 }
318 m->m_pkthdr.csum_flags = 0;
319 pktq = ip_pktq;
320 break;
321 #endif
322 #ifdef INET6
323 case AF_INET6:
324 csum_flags = m->m_pkthdr.csum_flags;
325 KASSERT((csum_flags & ~M_CSUM_UDPv6) == 0);
326 if (csum_flags != 0 &&
327 IN6_LOOPBACK_NEED_CHECKSUM(csum_flags)) {
328 ip6_undefer_csum(m, 0, csum_flags);
329 }
330 m->m_pkthdr.csum_flags = 0;
331 m->m_flags |= M_LOOP;
332 pktq = ip6_pktq;
333 break;
334 #endif
335 #ifdef NETATALK
336 case AF_APPLETALK:
337 ifq = &atintrq2;
338 isr = NETISR_ATALK;
339 break;
340 #endif
341 default:
342 printf("%s: can't handle af%d\n", ifp->if_xname,
343 dst->sa_family);
344 m_freem(m);
345 error = EAFNOSUPPORT;
346 goto out;
347 }
348
349 s = splnet();
350 if (__predict_true(pktq)) {
351 error = 0;
352
353 if (__predict_true(pktq_enqueue(pktq, m, 0))) {
354 ifp->if_ipackets++;
355 ifp->if_ibytes += pktlen;
356 } else {
357 m_freem(m);
358 error = ENOBUFS;
359 }
360 splx(s);
361 goto out;
362 }
363 if (IF_QFULL(ifq)) {
364 IF_DROP(ifq);
365 m_freem(m);
366 splx(s);
367 error = ENOBUFS;
368 goto out;
369 }
370 ifp->if_ipackets++;
371 ifp->if_ibytes += m->m_pkthdr.len;
372 IF_ENQUEUE(ifq, m);
373 schednetisr(isr);
374 splx(s);
375 out:
376 KERNEL_UNLOCK_ONE(NULL);
377 return error;
378 }
379
380 #ifdef ALTQ
381 static void
382 lostart(struct ifnet *ifp)
383 {
384 for (;;) {
385 pktqueue_t *pktq = NULL;
386 struct ifqueue *ifq = NULL;
387 struct mbuf *m;
388 size_t pktlen;
389 uint32_t af;
390 int s, isr = 0;
391
392 IFQ_DEQUEUE(&ifp->if_snd, m);
393 if (m == NULL)
394 return;
395
396 af = *(mtod(m, uint32_t *));
397 m_adj(m, sizeof(uint32_t));
398
399 switch (af) {
400 #ifdef INET
401 case AF_INET:
402 pktq = ip_pktq;
403 break;
404 #endif
405 #ifdef INET6
406 case AF_INET6:
407 m->m_flags |= M_LOOP;
408 pktq = ip6_pktq;
409 break;
410 #endif
411 #ifdef NETATALK
412 case AF_APPLETALK:
413 ifq = &atintrq2;
414 isr = NETISR_ATALK;
415 break;
416 #endif
417 default:
418 printf("%s: can't handle af%d\n", ifp->if_xname, af);
419 m_freem(m);
420 return;
421 }
422 pktlen = m->m_pkthdr.len;
423
424 s = splnet();
425 if (__predict_true(pktq)) {
426 if (__predict_false(pktq_enqueue(pktq, m, 0))) {
427 m_freem(m);
428 splx(s);
429 return;
430 }
431 ifp->if_ipackets++;
432 ifp->if_ibytes += pktlen;
433 splx(s);
434 continue;
435 }
436 if (IF_QFULL(ifq)) {
437 IF_DROP(ifq);
438 splx(s);
439 m_freem(m);
440 return;
441 }
442 IF_ENQUEUE(ifq, m);
443 schednetisr(isr);
444 ifp->if_ipackets++;
445 ifp->if_ibytes += pktlen;
446 splx(s);
447 }
448 }
449 #endif /* ALTQ */
450
451 /* ARGSUSED */
452 static void
453 loop_rtrequest(int cmd, struct rtentry *rt,
454 const struct rt_addrinfo *info)
455 {
456
457 if (rt)
458 rt->rt_rmx.rmx_mtu = lo0ifp->if_mtu;
459 }
460
461 /*
462 * Process an ioctl request.
463 */
464 /* ARGSUSED */
465 int
466 loioctl(struct ifnet *ifp, u_long cmd, void *data)
467 {
468 struct ifaddr *ifa;
469 struct ifreq *ifr = data;
470 int error = 0;
471
472 switch (cmd) {
473
474 case SIOCINITIFADDR:
475 ifp->if_flags |= IFF_UP;
476 ifa = (struct ifaddr *)data;
477 if (ifa != NULL)
478 ifa->ifa_rtrequest = loop_rtrequest;
479 /*
480 * Everything else is done at a higher level.
481 */
482 break;
483
484 case SIOCSIFMTU:
485 if ((unsigned)ifr->ifr_mtu > LOMTU_MAX)
486 error = EINVAL;
487 else if ((error = ifioctl_common(ifp, cmd, data)) == ENETRESET){
488 error = 0;
489 }
490 break;
491
492 case SIOCADDMULTI:
493 case SIOCDELMULTI:
494 if (ifr == NULL) {
495 error = EAFNOSUPPORT; /* XXX */
496 break;
497 }
498 switch (ifreq_getaddr(cmd, ifr)->sa_family) {
499
500 #ifdef INET
501 case AF_INET:
502 break;
503 #endif
504 #ifdef INET6
505 case AF_INET6:
506 break;
507 #endif
508
509 default:
510 error = EAFNOSUPPORT;
511 break;
512 }
513 break;
514
515 default:
516 error = ifioctl_common(ifp, cmd, data);
517 }
518 return (error);
519 }
520
521 /*
522 * Module infrastructure
523 */
524 #include "if_module.h"
525
526 IF_MODULE(MODULE_CLASS_DRIVER, loop, "")
527