1 /* $NetBSD: if_loop.c,v 1.119 2025/09/21 15:11:52 christos 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 1.119 2025/09/21 15:11:52 christos 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/route.h> 94 95 #ifdef INET 96 #include <netinet/in.h> 97 #include <netinet/in_systm.h> 98 #include <netinet/in_var.h> 99 #include <netinet/in_offload.h> 100 #include <netinet/ip.h> 101 #include <netinet/ip_var.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; 178 179 ifp = if_alloc(IFT_LOOP); 180 181 if_initname(ifp, ifc->ifc_name, unit); 182 183 ifp->if_mtu = LOMTU; 184 ifp->if_flags = IFF_LOOPBACK | IFF_MULTICAST; 185 #ifdef NET_MPSAFE 186 ifp->if_extflags = IFEF_MPSAFE; 187 #endif 188 ifp->if_ioctl = loioctl; 189 ifp->if_output = looutput; 190 #ifdef ALTQ 191 ifp->if_start = lostart; 192 #endif 193 ifp->if_type = IFT_LOOP; 194 ifp->if_hdrlen = 0; 195 ifp->if_addrlen = 0; 196 ifp->if_dlt = DLT_NULL; 197 IFQ_SET_READY(&ifp->if_snd); 198 if (unit == 0) 199 lo0ifp = ifp; 200 if_initialize(ifp); 201 ifp->if_link_state = LINK_STATE_UP; 202 if_alloc_sadl(ifp); 203 bpf_attach(ifp, DLT_NULL, sizeof(u_int)); 204 #ifdef MBUFTRACE 205 ifp->if_mowner = malloc(sizeof(struct mowner), M_DEVBUF, 206 M_WAITOK | M_ZERO); 207 strlcpy(ifp->if_mowner->mo_name, ifp->if_xname, 208 sizeof(ifp->if_mowner->mo_name)); 209 MOWNER_ATTACH(ifp->if_mowner); 210 #endif 211 212 ifp->if_flags |= IFF_RUNNING; 213 if_register(ifp); 214 215 return (0); 216 } 217 218 static int 219 loop_clone_destroy(struct ifnet *ifp) 220 { 221 222 if (ifp == lo0ifp) 223 return (EPERM); 224 225 ifp->if_flags &= ~IFF_RUNNING; 226 227 #ifdef MBUFTRACE 228 MOWNER_DETACH(ifp->if_mowner); 229 free(ifp->if_mowner, M_DEVBUF); 230 #endif 231 232 bpf_detach(ifp); 233 if_detach(ifp); 234 235 if_free(ifp); 236 237 return (0); 238 } 239 240 int 241 looutput(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst, 242 const struct rtentry *rt) 243 { 244 pktqueue_t *pktq = NULL; 245 int s; 246 int csum_flags; 247 int error = 0; 248 size_t pktlen; 249 250 MCLAIM(m, ifp->if_mowner); 251 252 KERNEL_LOCK_UNLESS_NET_MPSAFE(); 253 254 if ((m->m_flags & M_PKTHDR) == 0) 255 panic("looutput: no header mbuf"); 256 if (ifp->if_flags & IFF_LOOPBACK) 257 bpf_mtap_af(ifp, dst->sa_family, m, BPF_D_OUT); 258 m_set_rcvif(m, ifp); 259 260 if (rt && rt->rt_flags & (RTF_REJECT|RTF_BLACKHOLE)) { 261 m_freem(m); 262 error = (rt->rt_flags & RTF_BLACKHOLE ? 0 : 263 rt->rt_flags & RTF_HOST ? EHOSTUNREACH : ENETUNREACH); 264 goto out; 265 } 266 267 pktlen = m->m_pkthdr.len; 268 269 if_statadd2(ifp, if_opackets, 1, if_obytes, pktlen); 270 271 #ifdef ALTQ 272 /* 273 * ALTQ on the loopback interface is just for debugging. It's 274 * used only for loopback interfaces, not for a simplex interface. 275 */ 276 if ((ALTQ_IS_ENABLED(&ifp->if_snd) || TBR_IS_ENABLED(&ifp->if_snd)) && 277 ifp->if_start == lostart) { 278 /* 279 * If the queueing discipline needs packet classification, 280 * do it before prepending the link headers. 281 */ 282 IFQ_CLASSIFY(&ifp->if_snd, m, dst->sa_family); 283 284 M_PREPEND(m, sizeof(uint32_t), M_DONTWAIT); 285 if (m == NULL) { 286 if_statinc(ifp, if_oerrors); 287 error = ENOBUFS; 288 goto out; 289 } 290 *(mtod(m, uint32_t *)) = dst->sa_family; 291 292 error = if_transmit_lock(ifp, m); 293 goto out; 294 } 295 #endif /* ALTQ */ 296 297 m_tag_delete_chain(m); 298 299 #ifdef MPLS 300 bool is_mpls = false; 301 if (rt != NULL && rt_gettag(rt) != NULL && 302 rt_gettag(rt)->sa_family == AF_MPLS && 303 (m->m_flags & (M_MCAST | M_BCAST)) == 0) { 304 union mpls_shim msh; 305 msh.s_addr = MPLS_GETSADDR(rt); 306 if (msh.shim.label != MPLS_LABEL_IMPLNULL) { 307 is_mpls = true; 308 pktq = mpls_pktq; 309 } 310 } 311 if (!is_mpls) 312 #endif 313 switch (dst->sa_family) { 314 315 #ifdef INET 316 case AF_INET: 317 csum_flags = m->m_pkthdr.csum_flags; 318 KASSERT((csum_flags & ~(M_CSUM_IPv4|M_CSUM_UDPv4)) == 0); 319 if (csum_flags != 0 && IN_LOOPBACK_NEED_CHECKSUM(csum_flags)) { 320 in_undefer_cksum(m, 0, csum_flags); 321 m->m_pkthdr.csum_flags = 0; 322 } else { 323 /* 324 * Do nothing. Pass M_CSUM_IPv4 and M_CSUM_UDPv4 as 325 * they are to tell those are calculated and good. 326 */ 327 } 328 pktq = ip_pktq; 329 break; 330 #endif 331 #ifdef INET6 332 case AF_INET6: 333 csum_flags = m->m_pkthdr.csum_flags; 334 KASSERT((csum_flags & ~M_CSUM_UDPv6) == 0); 335 if (csum_flags != 0 && 336 IN6_LOOPBACK_NEED_CHECKSUM(csum_flags)) { 337 in6_undefer_cksum(m, 0, csum_flags); 338 m->m_pkthdr.csum_flags = 0; 339 } else { 340 /* 341 * Do nothing. Pass M_CSUM_UDPv6 as 342 * they are to tell those are calculated and good. 343 */ 344 } 345 m->m_flags |= M_LOOP; 346 pktq = ip6_pktq; 347 break; 348 #endif 349 #ifdef NETATALK 350 case AF_APPLETALK: 351 pktq = at_pktq2; 352 break; 353 #endif 354 default: 355 rt_unhandled(__func__, ifp, dst); 356 m_freem(m); 357 error = EAFNOSUPPORT; 358 goto out; 359 } 360 361 KASSERT(pktq != NULL); 362 363 error = 0; 364 s = splnet(); 365 if (__predict_true(pktq_enqueue(pktq, m, 0))) { 366 if_statadd2(ifp, if_ipackets, 1, if_ibytes, pktlen); 367 } else { 368 m_freem(m); 369 if_statinc(ifp, if_oerrors); 370 error = ENOBUFS; 371 } 372 splx(s); 373 out: 374 KERNEL_UNLOCK_UNLESS_NET_MPSAFE(); 375 return error; 376 } 377 378 #ifdef ALTQ 379 static void 380 lostart(struct ifnet *ifp) 381 { 382 for (;;) { 383 pktqueue_t *pktq = NULL; 384 struct mbuf *m; 385 struct sockaddr sa; 386 size_t pktlen; 387 uint32_t af; 388 int s; 389 390 IFQ_DEQUEUE(&ifp->if_snd, m); 391 if (m == NULL) 392 return; 393 394 af = *(mtod(m, uint32_t *)); 395 m_adj(m, sizeof(uint32_t)); 396 397 switch (af) { 398 #ifdef INET 399 case AF_INET: 400 pktq = ip_pktq; 401 break; 402 #endif 403 #ifdef INET6 404 case AF_INET6: 405 m->m_flags |= M_LOOP; 406 pktq = ip6_pktq; 407 break; 408 #endif 409 #ifdef NETATALK 410 case AF_APPLETALK: 411 pktq = at_pktq2; 412 break; 413 #endif 414 default: 415 memset(&sa, 0, sizeof(sa)); 416 sa.sa_family = af; 417 rt_unhandled(__func__, ifp, &sa); 418 m_freem(m); 419 return; 420 } 421 pktlen = m->m_pkthdr.len; 422 423 KASSERT(pktq != NULL); 424 425 s = splnet(); 426 if (__predict_false(pktq_enqueue(pktq, m, 0))) { 427 m_freem(m); 428 splx(s); 429 return; 430 } 431 if_statadd2(ifp, if_ipackets, 1, if_ibytes, pktlen); 432 splx(s); 433 } 434 } 435 #endif /* ALTQ */ 436 437 /* ARGSUSED */ 438 static void 439 loop_rtrequest(int cmd, struct rtentry *rt, 440 const struct rt_addrinfo *info) 441 { 442 443 if (rt) 444 rt->rt_rmx.rmx_mtu = lo0ifp->if_mtu; 445 } 446 447 /* 448 * Process an ioctl request. 449 */ 450 /* ARGSUSED */ 451 int 452 loioctl(struct ifnet *ifp, u_long cmd, void *data) 453 { 454 struct ifaddr *ifa; 455 struct ifreq *ifr = data; 456 int error = 0; 457 458 switch (cmd) { 459 460 case SIOCINITIFADDR: 461 ifp->if_flags |= IFF_UP; 462 ifa = (struct ifaddr *)data; 463 if (ifa != NULL) 464 ifa->ifa_rtrequest = loop_rtrequest; 465 /* 466 * Everything else is done at a higher level. 467 */ 468 break; 469 470 case SIOCSIFMTU: 471 if ((unsigned)ifr->ifr_mtu > LOMTU_MAX) 472 error = EINVAL; 473 else if ((error = ifioctl_common(ifp, cmd, data)) == ENETRESET){ 474 error = 0; 475 } 476 break; 477 478 case SIOCADDMULTI: 479 case SIOCDELMULTI: 480 if (ifr == NULL) { 481 error = EAFNOSUPPORT; /* XXX */ 482 break; 483 } 484 switch (ifreq_getaddr(cmd, ifr)->sa_family) { 485 486 #ifdef INET 487 case AF_INET: 488 break; 489 #endif 490 #ifdef INET6 491 case AF_INET6: 492 break; 493 #endif 494 495 default: 496 error = EAFNOSUPPORT; 497 break; 498 } 499 break; 500 501 default: 502 error = ifioctl_common(ifp, cmd, data); 503 } 504 return (error); 505 } 506 507 /* 508 * Module infrastructure 509 */ 510 #include "if_module.h" 511 512 IF_MODULE(MODULE_CLASS_DRIVER, loop, NULL) 513