if_loop.c revision 1.92 1 /* $NetBSD: if_loop.c,v 1.92 2016/08/11 13:57:02 kre 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 1.92 2016/08/11 13:57:02 kre 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 struct if_clone loop_cloner =
141 IF_CLONE_INITIALIZER("lo", loop_clone_create, loop_clone_destroy);
142
143 void
144 loopattach(int n)
145 {
146
147 /*
148 * Nothing to do here, initialization is handled by the
149 * module initialization code in loopnit() below).
150 */
151 }
152
153 void
154 loopinit(void)
155 {
156
157 if (lo0ifp != NULL) /* can happen in rump kernel */
158 return;
159
160 (void)loop_clone_create(&loop_cloner, 0); /* lo0 always exists */
161 if_clone_attach(&loop_cloner);
162 }
163
164 static int
165 loopdetach(void)
166 {
167 /* no detach for now; we don't allow lo0 to be deleted */
168 return EBUSY;
169 }
170
171 static int
172 loop_clone_create(struct if_clone *ifc, int unit)
173 {
174 struct ifnet *ifp;
175
176 ifp = if_alloc(IFT_LOOP);
177
178 if_initname(ifp, ifc->ifc_name, unit);
179
180 ifp->if_mtu = LOMTU;
181 ifp->if_flags = IFF_LOOPBACK | IFF_MULTICAST | IFF_RUNNING;
182 ifp->if_extflags = IFEF_OUTPUT_MPSAFE;
183 ifp->if_ioctl = loioctl;
184 ifp->if_output = looutput;
185 #ifdef ALTQ
186 ifp->if_start = lostart;
187 #endif
188 ifp->if_type = IFT_LOOP;
189 ifp->if_hdrlen = 0;
190 ifp->if_addrlen = 0;
191 ifp->if_dlt = DLT_NULL;
192 IFQ_SET_READY(&ifp->if_snd);
193 if (unit == 0)
194 lo0ifp = ifp;
195 if_attach(ifp);
196 if_alloc_sadl(ifp);
197 bpf_attach(ifp, DLT_NULL, sizeof(u_int));
198 #ifdef MBUFTRACE
199 ifp->if_mowner = malloc(sizeof(struct mowner), M_DEVBUF,
200 M_WAITOK | M_ZERO);
201 strlcpy(ifp->if_mowner->mo_name, ifp->if_xname,
202 sizeof(ifp->if_mowner->mo_name));
203 MOWNER_ATTACH(ifp->if_mowner);
204 #endif
205
206 return (0);
207 }
208
209 static int
210 loop_clone_destroy(struct ifnet *ifp)
211 {
212
213 if (ifp == lo0ifp)
214 return (EPERM);
215
216 #ifdef MBUFTRACE
217 MOWNER_DETACH(ifp->if_mowner);
218 free(ifp->if_mowner, M_DEVBUF);
219 #endif
220
221 bpf_detach(ifp);
222 if_detach(ifp);
223
224 if_free(ifp);
225
226 return (0);
227 }
228
229 int
230 looutput(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst,
231 const struct rtentry *rt)
232 {
233 pktqueue_t *pktq = NULL;
234 struct ifqueue *ifq = NULL;
235 int s, isr = -1;
236 int csum_flags;
237 int error = 0;
238 size_t pktlen;
239
240 MCLAIM(m, ifp->if_mowner);
241
242 KERNEL_LOCK(1, NULL);
243
244 if ((m->m_flags & M_PKTHDR) == 0)
245 panic("looutput: no header mbuf");
246 if (ifp->if_flags & IFF_LOOPBACK)
247 bpf_mtap_af(ifp, dst->sa_family, m);
248 m_set_rcvif(m, ifp);
249
250 if (rt && rt->rt_flags & (RTF_REJECT|RTF_BLACKHOLE)) {
251 m_freem(m);
252 error = (rt->rt_flags & RTF_BLACKHOLE ? 0 :
253 rt->rt_flags & RTF_HOST ? EHOSTUNREACH : ENETUNREACH);
254 goto out;
255 }
256
257 pktlen = m->m_pkthdr.len;
258 ifp->if_opackets++;
259 ifp->if_obytes += pktlen;
260
261 #ifdef ALTQ
262 /*
263 * ALTQ on the loopback interface is just for debugging. It's
264 * used only for loopback interfaces, not for a simplex interface.
265 */
266 if ((ALTQ_IS_ENABLED(&ifp->if_snd) || TBR_IS_ENABLED(&ifp->if_snd)) &&
267 ifp->if_start == lostart) {
268 /*
269 * If the queueing discipline needs packet classification,
270 * do it before prepending the link headers.
271 */
272 IFQ_CLASSIFY(&ifp->if_snd, m, dst->sa_family);
273
274 M_PREPEND(m, sizeof(uint32_t), M_DONTWAIT);
275 if (m == NULL) {
276 error = ENOBUFS;
277 goto out;
278 }
279 *(mtod(m, uint32_t *)) = dst->sa_family;
280
281 error = if_transmit_lock(ifp, m);
282 goto out;
283 }
284 #endif /* ALTQ */
285
286 m_tag_delete_nonpersistent(m);
287
288 #ifdef MPLS
289 if (rt != NULL && rt_gettag(rt) != NULL &&
290 rt_gettag(rt)->sa_family == AF_MPLS &&
291 (m->m_flags & (M_MCAST | M_BCAST)) == 0) {
292 union mpls_shim msh;
293 msh.s_addr = MPLS_GETSADDR(rt);
294 if (msh.shim.label != MPLS_LABEL_IMPLNULL) {
295 ifq = &mplsintrq;
296 isr = NETISR_MPLS;
297 }
298 }
299 if (isr != NETISR_MPLS)
300 #endif
301 switch (dst->sa_family) {
302
303 #ifdef INET
304 case AF_INET:
305 csum_flags = m->m_pkthdr.csum_flags;
306 KASSERT((csum_flags & ~(M_CSUM_IPv4|M_CSUM_UDPv4)) == 0);
307 if (csum_flags != 0 && IN_LOOPBACK_NEED_CHECKSUM(csum_flags)) {
308 ip_undefer_csum(m, 0, csum_flags);
309 }
310 m->m_pkthdr.csum_flags = 0;
311 pktq = ip_pktq;
312 break;
313 #endif
314 #ifdef INET6
315 case AF_INET6:
316 csum_flags = m->m_pkthdr.csum_flags;
317 KASSERT((csum_flags & ~M_CSUM_UDPv6) == 0);
318 if (csum_flags != 0 &&
319 IN6_LOOPBACK_NEED_CHECKSUM(csum_flags)) {
320 ip6_undefer_csum(m, 0, csum_flags);
321 }
322 m->m_pkthdr.csum_flags = 0;
323 m->m_flags |= M_LOOP;
324 pktq = ip6_pktq;
325 break;
326 #endif
327 #ifdef NETATALK
328 case AF_APPLETALK:
329 ifq = &atintrq2;
330 isr = NETISR_ATALK;
331 break;
332 #endif
333 default:
334 printf("%s: can't handle af%d\n", ifp->if_xname,
335 dst->sa_family);
336 m_freem(m);
337 error = EAFNOSUPPORT;
338 goto out;
339 }
340
341 s = splnet();
342 if (__predict_true(pktq)) {
343 error = 0;
344
345 if (__predict_true(pktq_enqueue(pktq, m, 0))) {
346 ifp->if_ipackets++;
347 ifp->if_ibytes += pktlen;
348 } else {
349 m_freem(m);
350 error = ENOBUFS;
351 }
352 splx(s);
353 goto out;
354 }
355 if (IF_QFULL(ifq)) {
356 IF_DROP(ifq);
357 m_freem(m);
358 splx(s);
359 error = ENOBUFS;
360 goto out;
361 }
362 IF_ENQUEUE(ifq, m);
363 schednetisr(isr);
364 ifp->if_ipackets++;
365 ifp->if_ibytes += m->m_pkthdr.len;
366 splx(s);
367 out:
368 KERNEL_UNLOCK_ONE(NULL);
369 return error;
370 }
371
372 #ifdef ALTQ
373 static void
374 lostart(struct ifnet *ifp)
375 {
376 for (;;) {
377 pktqueue_t *pktq = NULL;
378 struct ifqueue *ifq = NULL;
379 struct mbuf *m;
380 size_t pktlen;
381 uint32_t af;
382 int s, isr = 0;
383
384 IFQ_DEQUEUE(&ifp->if_snd, m);
385 if (m == NULL)
386 return;
387
388 af = *(mtod(m, uint32_t *));
389 m_adj(m, sizeof(uint32_t));
390
391 switch (af) {
392 #ifdef INET
393 case AF_INET:
394 pktq = ip_pktq;
395 break;
396 #endif
397 #ifdef INET6
398 case AF_INET6:
399 m->m_flags |= M_LOOP;
400 pktq = ip6_pktq;
401 break;
402 #endif
403 #ifdef NETATALK
404 case AF_APPLETALK:
405 ifq = &atintrq2;
406 isr = NETISR_ATALK;
407 break;
408 #endif
409 default:
410 printf("%s: can't handle af%d\n", ifp->if_xname, af);
411 m_freem(m);
412 return;
413 }
414 pktlen = m->m_pkthdr.len;
415
416 s = splnet();
417 if (__predict_true(pktq)) {
418 if (__predict_false(pktq_enqueue(pktq, m, 0))) {
419 m_freem(m);
420 splx(s);
421 return;
422 }
423 ifp->if_ipackets++;
424 ifp->if_ibytes += pktlen;
425 splx(s);
426 continue;
427 }
428 if (IF_QFULL(ifq)) {
429 IF_DROP(ifq);
430 splx(s);
431 m_freem(m);
432 return;
433 }
434 IF_ENQUEUE(ifq, m);
435 schednetisr(isr);
436 ifp->if_ipackets++;
437 ifp->if_ibytes += pktlen;
438 splx(s);
439 }
440 }
441 #endif /* ALTQ */
442
443 /* ARGSUSED */
444 void
445 lortrequest(int cmd, struct rtentry *rt,
446 const struct rt_addrinfo *info)
447 {
448
449 if (rt)
450 rt->rt_rmx.rmx_mtu = lo0ifp->if_mtu;
451 }
452
453 /*
454 * Process an ioctl request.
455 */
456 /* ARGSUSED */
457 int
458 loioctl(struct ifnet *ifp, u_long cmd, void *data)
459 {
460 struct ifaddr *ifa;
461 struct ifreq *ifr = data;
462 int error = 0;
463
464 switch (cmd) {
465
466 case SIOCINITIFADDR:
467 ifp->if_flags |= IFF_UP;
468 ifa = (struct ifaddr *)data;
469 if (ifa != NULL)
470 ifa->ifa_rtrequest = lortrequest;
471 /*
472 * Everything else is done at a higher level.
473 */
474 break;
475
476 case SIOCSIFMTU:
477 if ((unsigned)ifr->ifr_mtu > LOMTU_MAX)
478 error = EINVAL;
479 else if ((error = ifioctl_common(ifp, cmd, data)) == ENETRESET){
480 error = 0;
481 }
482 break;
483
484 case SIOCADDMULTI:
485 case SIOCDELMULTI:
486 if (ifr == NULL) {
487 error = EAFNOSUPPORT; /* XXX */
488 break;
489 }
490 switch (ifreq_getaddr(cmd, ifr)->sa_family) {
491
492 #ifdef INET
493 case AF_INET:
494 break;
495 #endif
496 #ifdef INET6
497 case AF_INET6:
498 break;
499 #endif
500
501 default:
502 error = EAFNOSUPPORT;
503 break;
504 }
505 break;
506
507 default:
508 error = ifioctl_common(ifp, cmd, data);
509 }
510 return (error);
511 }
512
513 /*
514 * Module infrastructure
515 */
516 #include "if_module.h"
517
518 IF_MODULE(MODULE_CLASS_DRIVER, loop, "")
519