1 /*- 2 * Copyright (c) 1982, 1986, 1993 3 * The Regents of the University of California. All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 4. Neither the name of the University nor the names of its contributors 14 * may be used to endorse or promote products derived from this software 15 * without specific prior written permission. 16 * 17 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 20 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 27 * SUCH DAMAGE. 28 * 29 * @(#)if_loop.c 8.2 (Berkeley) 1/9/95 30 * $FreeBSD$ 31 */ 32 33 /* 34 * Loopback interface driver for protocol testing and timing. 35 */ 36 37 #include "opt_atalk.h" 38 #include "opt_inet.h" 39 #include "opt_inet6.h" 40 41 #include <sys/param.h> 42 #include <sys/systm.h> 43 #include <sys/kernel.h> 44 #include <sys/mbuf.h> 45 #include <sys/module.h> 46 #include <machine/bus.h> 47 #include <sys/rman.h> 48 #include <sys/socket.h> 49 #include <sys/sockio.h> 50 #include <sys/sysctl.h> 51 52 #include <net/if.h> 53 #include <net/if_var.h> 54 #include <net/if_clone.h> 55 #include <net/if_types.h> 56 #include <net/netisr.h> 57 #include <net/route.h> 58 #include <net/bpf.h> 59 #include <net/vnet.h> 60 61 #ifdef INET 62 #include <netinet/in.h> 63 #include <netinet/in_var.h> 64 #endif 65 66 #ifdef INET6 67 #ifndef INET 68 #include <netinet/in.h> 69 #endif 70 #include <netinet6/in6_var.h> 71 #include <netinet/ip6.h> 72 #endif 73 74 #ifdef NETATALK 75 #include <netatalk/at.h> 76 #include <netatalk/at_var.h> 77 #endif 78 79 #include <security/mac/mac_framework.h> 80 81 #ifdef TINY_LOMTU 82 #define LOMTU (1024+512) 83 #elif defined(LARGE_LOMTU) 84 #define LOMTU 131072 85 #else 86 #define LOMTU 16384 87 #endif 88 89 #define LO_CSUM_FEATURES (CSUM_IP | CSUM_TCP | CSUM_UDP | CSUM_SCTP) 90 #define LO_CSUM_FEATURES6 (CSUM_TCP_IPV6 | CSUM_UDP_IPV6 | CSUM_SCTP_IPV6) 91 #define LO_CSUM_SET (CSUM_DATA_VALID | CSUM_DATA_VALID_IPV6 | \ 92 CSUM_PSEUDO_HDR | \ 93 CSUM_IP_CHECKED | CSUM_IP_VALID | \ 94 CSUM_SCTP_VALID) 95 96 int loioctl(struct ifnet *, u_long, caddr_t); 97 static void lortrequest(int, struct rtentry *, struct rt_addrinfo *); 98 int looutput(struct ifnet *ifp, struct mbuf *m, 99 const struct sockaddr *dst, struct route *ro); 100 static int lo_clone_create(struct if_clone *, int, caddr_t); 101 static void lo_clone_destroy(struct ifnet *); 102 103 VNET_DEFINE(struct ifnet *, loif); /* Used externally */ 104 105 #ifdef VIMAGE 106 static VNET_DEFINE(struct if_clone *, lo_cloner); 107 #define V_lo_cloner VNET(lo_cloner) 108 #endif 109 110 static struct if_clone *lo_cloner; 111 static const char loname[] = "lo"; 112 113 static void 114 lo_clone_destroy(struct ifnet *ifp) 115 { 116 117 #ifndef VIMAGE 118 /* XXX: destroying lo0 will lead to panics. */ 119 KASSERT(V_loif != ifp, ("%s: destroying lo0", __func__)); 120 #endif 121 122 bpfdetach(ifp); 123 if_detach(ifp); 124 if_free(ifp); 125 } 126 127 static int 128 lo_clone_create(struct if_clone *ifc, int unit, caddr_t params) 129 { 130 struct ifnet *ifp; 131 132 ifp = if_alloc(IFT_LOOP); 133 if (ifp == NULL) 134 return (ENOSPC); 135 136 if_initname(ifp, loname, unit); 137 ifp->if_mtu = LOMTU; 138 ifp->if_flags = IFF_LOOPBACK | IFF_MULTICAST; 139 ifp->if_ioctl = loioctl; 140 ifp->if_output = looutput; 141 ifp->if_snd.ifq_maxlen = ifqmaxlen; 142 ifp->if_capabilities = ifp->if_capenable = 143 IFCAP_HWCSUM | IFCAP_HWCSUM_IPV6; 144 ifp->if_hwassist = LO_CSUM_FEATURES | LO_CSUM_FEATURES6; 145 if_attach(ifp); 146 bpfattach(ifp, DLT_NULL, sizeof(u_int32_t)); 147 if (V_loif == NULL) 148 V_loif = ifp; 149 150 return (0); 151 } 152 153 static void 154 vnet_loif_init(const void *unused __unused) 155 { 156 157 #ifdef VIMAGE 158 lo_cloner = if_clone_simple(loname, lo_clone_create, lo_clone_destroy, 159 1); 160 V_lo_cloner = lo_cloner; 161 #else 162 lo_cloner = if_clone_simple(loname, lo_clone_create, lo_clone_destroy, 163 1); 164 #endif 165 } 166 VNET_SYSINIT(vnet_loif_init, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY, 167 vnet_loif_init, NULL); 168 169 #ifdef VIMAGE 170 static void 171 vnet_loif_uninit(const void *unused __unused) 172 { 173 174 if_clone_detach(V_lo_cloner); 175 V_loif = NULL; 176 } 177 VNET_SYSUNINIT(vnet_loif_uninit, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY, 178 vnet_loif_uninit, NULL); 179 #endif 180 181 static int 182 loop_modevent(module_t mod, int type, void *data) 183 { 184 185 switch (type) { 186 case MOD_LOAD: 187 break; 188 189 case MOD_UNLOAD: 190 printf("loop module unload - not possible for this module type\n"); 191 return (EINVAL); 192 193 default: 194 return (EOPNOTSUPP); 195 } 196 return (0); 197 } 198 199 static moduledata_t loop_mod = { 200 "if_lo", 201 loop_modevent, 202 0 203 }; 204 205 DECLARE_MODULE(if_lo, loop_mod, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY); 206 207 int 208 looutput(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst, 209 struct route *ro) 210 { 211 u_int32_t af; 212 struct rtentry *rt = NULL; 213 #ifdef MAC 214 int error; 215 #endif 216 217 M_ASSERTPKTHDR(m); /* check if we have the packet header */ 218 219 if (ro != NULL) 220 rt = ro->ro_rt; 221 #ifdef MAC 222 error = mac_ifnet_check_transmit(ifp, m); 223 if (error) { 224 m_freem(m); 225 return (error); 226 } 227 #endif 228 229 if (rt && rt->rt_flags & (RTF_REJECT|RTF_BLACKHOLE)) { 230 m_freem(m); 231 return (rt->rt_flags & RTF_BLACKHOLE ? 0 : 232 rt->rt_flags & RTF_HOST ? EHOSTUNREACH : ENETUNREACH); 233 } 234 235 ifp->if_opackets++; 236 ifp->if_obytes += m->m_pkthdr.len; 237 238 /* BPF writes need to be handled specially. */ 239 if (dst->sa_family == AF_UNSPEC) 240 bcopy(dst->sa_data, &af, sizeof(af)); 241 else 242 af = dst->sa_family; 243 244 #if 1 /* XXX */ 245 switch (af) { 246 case AF_INET: 247 if (ifp->if_capenable & IFCAP_RXCSUM) { 248 m->m_pkthdr.csum_data = 0xffff; 249 m->m_pkthdr.csum_flags = LO_CSUM_SET; 250 } 251 m->m_pkthdr.csum_flags &= ~LO_CSUM_FEATURES; 252 break; 253 case AF_INET6: 254 #if 0 255 /* 256 * XXX-BZ for now always claim the checksum is good despite 257 * any interface flags. This is a workaround for 9.1-R and 258 * a proper solution ought to be sought later. 259 */ 260 if (ifp->if_capenable & IFCAP_RXCSUM_IPV6) { 261 m->m_pkthdr.csum_data = 0xffff; 262 m->m_pkthdr.csum_flags = LO_CSUM_SET; 263 } 264 #else 265 m->m_pkthdr.csum_data = 0xffff; 266 m->m_pkthdr.csum_flags = LO_CSUM_SET; 267 #endif 268 m->m_pkthdr.csum_flags &= ~LO_CSUM_FEATURES6; 269 break; 270 case AF_APPLETALK: 271 break; 272 default: 273 printf("looutput: af=%d unexpected\n", af); 274 m_freem(m); 275 return (EAFNOSUPPORT); 276 } 277 #endif 278 return (if_simloop(ifp, m, af, 0)); 279 } 280 281 /* 282 * if_simloop() 283 * 284 * This function is to support software emulation of hardware loopback, 285 * i.e., for interfaces with the IFF_SIMPLEX attribute. Since they can't 286 * hear their own broadcasts, we create a copy of the packet that we 287 * would normally receive via a hardware loopback. 288 * 289 * This function expects the packet to include the media header of length hlen. 290 */ 291 int 292 if_simloop(struct ifnet *ifp, struct mbuf *m, int af, int hlen) 293 { 294 int isr; 295 296 M_ASSERTPKTHDR(m); 297 m_tag_delete_nonpersistent(m); 298 m->m_pkthdr.rcvif = ifp; 299 300 #ifdef MAC 301 mac_ifnet_create_mbuf(ifp, m); 302 #endif 303 304 /* 305 * Let BPF see incoming packet in the following manner: 306 * - Emulated packet loopback for a simplex interface 307 * (net/if_ethersubr.c) 308 * -> passes it to ifp's BPF 309 * - IPv4/v6 multicast packet loopback (netinet(6)/ip(6)_output.c) 310 * -> not passes it to any BPF 311 * - Normal packet loopback from myself to myself (net/if_loop.c) 312 * -> passes to lo0's BPF (even in case of IPv6, where ifp!=lo0) 313 */ 314 if (hlen > 0) { 315 if (bpf_peers_present(ifp->if_bpf)) { 316 bpf_mtap(ifp->if_bpf, m); 317 } 318 } else { 319 if (bpf_peers_present(V_loif->if_bpf)) { 320 if ((m->m_flags & M_MCAST) == 0 || V_loif == ifp) { 321 /* XXX beware sizeof(af) != 4 */ 322 u_int32_t af1 = af; 323 324 /* 325 * We need to prepend the address family. 326 */ 327 bpf_mtap2(V_loif->if_bpf, &af1, sizeof(af1), m); 328 } 329 } 330 } 331 332 /* Strip away media header */ 333 if (hlen > 0) { 334 m_adj(m, hlen); 335 #ifndef __NO_STRICT_ALIGNMENT 336 /* 337 * Some archs do not like unaligned data, so 338 * we move data down in the first mbuf. 339 */ 340 if (mtod(m, vm_offset_t) & 3) { 341 KASSERT(hlen >= 3, ("if_simloop: hlen too small")); 342 bcopy(m->m_data, 343 (char *)(mtod(m, vm_offset_t) 344 - (mtod(m, vm_offset_t) & 3)), 345 m->m_len); 346 m->m_data -= (mtod(m,vm_offset_t) & 3); 347 } 348 #endif 349 } 350 351 /* Deliver to upper layer protocol */ 352 switch (af) { 353 #ifdef INET 354 case AF_INET: 355 isr = NETISR_IP; 356 break; 357 #endif 358 #ifdef INET6 359 case AF_INET6: 360 m->m_flags |= M_LOOP; 361 isr = NETISR_IPV6; 362 break; 363 #endif 364 #ifdef NETATALK 365 case AF_APPLETALK: 366 isr = NETISR_ATALK2; 367 break; 368 #endif 369 default: 370 printf("if_simloop: can't handle af=%d\n", af); 371 m_freem(m); 372 return (EAFNOSUPPORT); 373 } 374 ifp->if_ipackets++; 375 ifp->if_ibytes += m->m_pkthdr.len; 376 netisr_queue(isr, m); /* mbuf is free'd on failure. */ 377 return (0); 378 } 379 380 /* ARGSUSED */ 381 static void 382 lortrequest(int cmd, struct rtentry *rt, struct rt_addrinfo *info) 383 { 384 385 RT_LOCK_ASSERT(rt); 386 rt->rt_mtu = rt->rt_ifp->if_mtu; 387 } 388 389 /* 390 * Process an ioctl request. 391 */ 392 /* ARGSUSED */ 393 int 394 loioctl(struct ifnet *ifp, u_long cmd, caddr_t data) 395 { 396 struct ifaddr *ifa; 397 struct ifreq *ifr = (struct ifreq *)data; 398 int error = 0, mask; 399 400 switch (cmd) { 401 case SIOCSIFADDR: 402 ifp->if_flags |= IFF_UP; 403 ifp->if_drv_flags |= IFF_DRV_RUNNING; 404 ifa = (struct ifaddr *)data; 405 ifa->ifa_rtrequest = lortrequest; 406 /* 407 * Everything else is done at a higher level. 408 */ 409 break; 410 411 case SIOCADDMULTI: 412 case SIOCDELMULTI: 413 if (ifr == 0) { 414 error = EAFNOSUPPORT; /* XXX */ 415 break; 416 } 417 switch (ifr->ifr_addr.sa_family) { 418 419 #ifdef INET 420 case AF_INET: 421 break; 422 #endif 423 #ifdef INET6 424 case AF_INET6: 425 break; 426 #endif 427 428 default: 429 error = EAFNOSUPPORT; 430 break; 431 } 432 break; 433 434 case SIOCSIFMTU: 435 ifp->if_mtu = ifr->ifr_mtu; 436 break; 437 438 case SIOCSIFFLAGS: 439 break; 440 441 case SIOCSIFCAP: 442 mask = ifp->if_capenable ^ ifr->ifr_reqcap; 443 if ((mask & IFCAP_RXCSUM) != 0) 444 ifp->if_capenable ^= IFCAP_RXCSUM; 445 if ((mask & IFCAP_TXCSUM) != 0) 446 ifp->if_capenable ^= IFCAP_TXCSUM; 447 if ((mask & IFCAP_RXCSUM_IPV6) != 0) { 448 #if 0 449 ifp->if_capenable ^= IFCAP_RXCSUM_IPV6; 450 #else 451 error = EOPNOTSUPP; 452 break; 453 #endif 454 } 455 if ((mask & IFCAP_TXCSUM_IPV6) != 0) { 456 #if 0 457 ifp->if_capenable ^= IFCAP_TXCSUM_IPV6; 458 #else 459 error = EOPNOTSUPP; 460 break; 461 #endif 462 } 463 ifp->if_hwassist = 0; 464 if (ifp->if_capenable & IFCAP_TXCSUM) 465 ifp->if_hwassist = LO_CSUM_FEATURES; 466 #if 0 467 if (ifp->if_capenable & IFCAP_TXCSUM_IPV6) 468 ifp->if_hwassist |= LO_CSUM_FEATURES6; 469 #endif 470 break; 471 472 default: 473 error = EINVAL; 474 } 475 return (error); 476 } 477