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_inet.h" 38 #include "opt_inet6.h" 39 40 #include <sys/param.h> 41 #include <sys/systm.h> 42 #include <sys/kernel.h> 43 #include <sys/mbuf.h> 44 #include <sys/module.h> 45 #include <machine/bus.h> 46 #include <sys/rman.h> 47 #include <sys/socket.h> 48 #include <sys/sockio.h> 49 #include <sys/sysctl.h> 50 51 #include <net/if.h> 52 #include <net/if_var.h> 53 #include <net/if_clone.h> 54 #include <net/if_types.h> 55 #include <net/netisr.h> 56 #include <net/route.h> 57 #include <net/bpf.h> 58 #include <net/vnet.h> 59 60 #ifdef INET 61 #include <netinet/in.h> 62 #include <netinet/in_var.h> 63 #endif 64 65 #ifdef INET6 66 #ifndef INET 67 #include <netinet/in.h> 68 #endif 69 #include <netinet6/in6_var.h> 70 #include <netinet/ip6.h> 71 #endif 72 73 #include <net/rt_nhops.h> 74 75 #include <security/mac/mac_framework.h> 76 77 #ifdef TINY_LOMTU 78 #define LOMTU (1024+512) 79 #elif defined(LARGE_LOMTU) 80 #define LOMTU 131072 81 #else 82 #define LOMTU 16384 83 #endif 84 85 #define LO_CSUM_FEATURES (CSUM_IP | CSUM_TCP | CSUM_UDP | CSUM_SCTP) 86 #define LO_CSUM_FEATURES6 (CSUM_TCP_IPV6 | CSUM_UDP_IPV6 | CSUM_SCTP_IPV6) 87 #define LO_CSUM_SET (CSUM_DATA_VALID | CSUM_DATA_VALID_IPV6 | \ 88 CSUM_PSEUDO_HDR | \ 89 CSUM_IP_CHECKED | CSUM_IP_VALID | \ 90 CSUM_SCTP_VALID) 91 92 int loioctl(struct ifnet *, u_long, caddr_t); 93 int looutput(struct ifnet *ifp, struct mbuf *m, 94 const struct sockaddr *dst, struct nhop_info *ni); 95 static int lo_clone_create(struct if_clone *, int, caddr_t); 96 static void lo_clone_destroy(struct ifnet *); 97 98 VNET_DEFINE(struct ifnet *, loif); /* Used externally */ 99 100 #ifdef VIMAGE 101 static VNET_DEFINE(struct if_clone *, lo_cloner); 102 #define V_lo_cloner VNET(lo_cloner) 103 #endif 104 105 static struct if_clone *lo_cloner; 106 static const char loname[] = "lo"; 107 108 static void 109 lo_clone_destroy(struct ifnet *ifp) 110 { 111 112 #ifndef VIMAGE 113 /* XXX: destroying lo0 will lead to panics. */ 114 KASSERT(V_loif != ifp, ("%s: destroying lo0", __func__)); 115 #endif 116 117 bpfdetach(ifp); 118 if_detach(ifp); 119 if_free(ifp); 120 } 121 122 static int 123 lo_clone_create(struct if_clone *ifc, int unit, caddr_t params) 124 { 125 struct ifnet *ifp; 126 127 ifp = if_alloc(IFT_LOOP); 128 if (ifp == NULL) 129 return (ENOSPC); 130 131 if_initname(ifp, loname, unit); 132 ifp->if_mtu = LOMTU; 133 ifp->if_flags = IFF_LOOPBACK | IFF_MULTICAST; 134 ifp->if_ioctl = loioctl; 135 ifp->if_output = looutput; 136 ifp->if_snd.ifq_maxlen = ifqmaxlen; 137 ifp->if_capabilities = ifp->if_capenable = 138 IFCAP_HWCSUM | IFCAP_HWCSUM_IPV6; 139 ifp->if_hwassist = LO_CSUM_FEATURES | LO_CSUM_FEATURES6; 140 if_attach(ifp); 141 bpfattach(ifp, DLT_NULL, sizeof(u_int32_t)); 142 if (V_loif == NULL) 143 V_loif = ifp; 144 145 return (0); 146 } 147 148 static void 149 vnet_loif_init(const void *unused __unused) 150 { 151 152 #ifdef VIMAGE 153 lo_cloner = if_clone_simple(loname, lo_clone_create, lo_clone_destroy, 154 1); 155 V_lo_cloner = lo_cloner; 156 #else 157 lo_cloner = if_clone_simple(loname, lo_clone_create, lo_clone_destroy, 158 1); 159 #endif 160 } 161 VNET_SYSINIT(vnet_loif_init, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY, 162 vnet_loif_init, NULL); 163 164 #ifdef VIMAGE 165 static void 166 vnet_loif_uninit(const void *unused __unused) 167 { 168 169 if_clone_detach(V_lo_cloner); 170 V_loif = NULL; 171 } 172 VNET_SYSUNINIT(vnet_loif_uninit, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY, 173 vnet_loif_uninit, NULL); 174 #endif 175 176 static int 177 loop_modevent(module_t mod, int type, void *data) 178 { 179 180 switch (type) { 181 case MOD_LOAD: 182 break; 183 184 case MOD_UNLOAD: 185 printf("loop module unload - not possible for this module type\n"); 186 return (EINVAL); 187 188 default: 189 return (EOPNOTSUPP); 190 } 191 return (0); 192 } 193 194 static moduledata_t loop_mod = { 195 "if_lo", 196 loop_modevent, 197 0 198 }; 199 200 DECLARE_MODULE(if_lo, loop_mod, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY); 201 202 int 203 looutput(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst, 204 struct nhop_info *ni) 205 { 206 uint32_t af; 207 uint32_t nh_flags; 208 #ifdef MAC 209 int error; 210 #endif 211 212 M_ASSERTPKTHDR(m); /* check if we have the packet header */ 213 214 nh_flags = 0; 215 af = AF_UNSPEC; 216 if (ni != NULL && ni->ni_nh != NULL) { 217 nh_flags = ni->ni_nh->nh_flags; 218 af = ni->ni_family; 219 } else if (dst != NULL) 220 af = dst->sa_family; 221 222 #ifdef MAC 223 error = mac_ifnet_check_transmit(ifp, m); 224 if (error) { 225 m_freem(m); 226 return (error); 227 } 228 #endif 229 230 if (nh_flags & (NHF_REJECT | NHF_BLACKHOLE)) { 231 m_freem(m); 232 /* XXX: RTF_HOST */ 233 return (nh_flags & NHF_BLACKHOLE ? 0 : EHOSTUNREACH); 234 } 235 236 if_inc_counter(ifp, IFCOUNTER_OPACKETS, 1); 237 if_inc_counter(ifp, IFCOUNTER_OBYTES, m->m_pkthdr.len); 238 239 /* BPF writes need to be handled specially. */ 240 if (af == AF_UNSPEC && dst != NULL) 241 bcopy(dst->sa_data, &af, sizeof(af)); 242 243 #if 1 /* XXX */ 244 switch (af) { 245 case AF_INET: 246 if (ifp->if_capenable & IFCAP_RXCSUM) { 247 m->m_pkthdr.csum_data = 0xffff; 248 m->m_pkthdr.csum_flags = LO_CSUM_SET; 249 } 250 m->m_pkthdr.csum_flags &= ~LO_CSUM_FEATURES; 251 break; 252 case AF_INET6: 253 #if 0 254 /* 255 * XXX-BZ for now always claim the checksum is good despite 256 * any interface flags. This is a workaround for 9.1-R and 257 * a proper solution ought to be sought later. 258 */ 259 if (ifp->if_capenable & IFCAP_RXCSUM_IPV6) { 260 m->m_pkthdr.csum_data = 0xffff; 261 m->m_pkthdr.csum_flags = LO_CSUM_SET; 262 } 263 #else 264 m->m_pkthdr.csum_data = 0xffff; 265 m->m_pkthdr.csum_flags = LO_CSUM_SET; 266 #endif 267 m->m_pkthdr.csum_flags &= ~LO_CSUM_FEATURES6; 268 break; 269 default: 270 printf("looutput: af=%d unexpected\n", af); 271 m_freem(m); 272 return (EAFNOSUPPORT); 273 } 274 #endif 275 return (if_simloop(ifp, m, af, 0)); 276 } 277 278 /* 279 * if_simloop() 280 * 281 * This function is to support software emulation of hardware loopback, 282 * i.e., for interfaces with the IFF_SIMPLEX attribute. Since they can't 283 * hear their own broadcasts, we create a copy of the packet that we 284 * would normally receive via a hardware loopback. 285 * 286 * This function expects the packet to include the media header of length hlen. 287 */ 288 int 289 if_simloop(struct ifnet *ifp, struct mbuf *m, int af, int hlen) 290 { 291 int isr; 292 293 M_ASSERTPKTHDR(m); 294 m_tag_delete_nonpersistent(m); 295 m->m_pkthdr.rcvif = ifp; 296 297 #ifdef MAC 298 mac_ifnet_create_mbuf(ifp, m); 299 #endif 300 301 /* 302 * Let BPF see incoming packet in the following manner: 303 * - Emulated packet loopback for a simplex interface 304 * (net/if_ethersubr.c) 305 * -> passes it to ifp's BPF 306 * - IPv4/v6 multicast packet loopback (netinet(6)/ip(6)_output.c) 307 * -> not passes it to any BPF 308 * - Normal packet loopback from myself to myself (net/if_loop.c) 309 * -> passes to lo0's BPF (even in case of IPv6, where ifp!=lo0) 310 */ 311 if (hlen > 0) { 312 if (bpf_peers_present(ifp->if_bpf)) { 313 bpf_mtap(ifp->if_bpf, m); 314 } 315 } else { 316 if (bpf_peers_present(V_loif->if_bpf)) { 317 if ((m->m_flags & M_MCAST) == 0 || V_loif == ifp) { 318 /* XXX beware sizeof(af) != 4 */ 319 u_int32_t af1 = af; 320 321 /* 322 * We need to prepend the address family. 323 */ 324 bpf_mtap2(V_loif->if_bpf, &af1, sizeof(af1), m); 325 } 326 } 327 } 328 329 /* Strip away media header */ 330 if (hlen > 0) { 331 m_adj(m, hlen); 332 #ifndef __NO_STRICT_ALIGNMENT 333 /* 334 * Some archs do not like unaligned data, so 335 * we move data down in the first mbuf. 336 */ 337 if (mtod(m, vm_offset_t) & 3) { 338 KASSERT(hlen >= 3, ("if_simloop: hlen too small")); 339 bcopy(m->m_data, 340 (char *)(mtod(m, vm_offset_t) 341 - (mtod(m, vm_offset_t) & 3)), 342 m->m_len); 343 m->m_data -= (mtod(m,vm_offset_t) & 3); 344 } 345 #endif 346 } 347 348 /* Deliver to upper layer protocol */ 349 switch (af) { 350 #ifdef INET 351 case AF_INET: 352 isr = NETISR_IP; 353 break; 354 #endif 355 #ifdef INET6 356 case AF_INET6: 357 m->m_flags |= M_LOOP; 358 isr = NETISR_IPV6; 359 break; 360 #endif 361 default: 362 printf("if_simloop: can't handle af=%d\n", af); 363 m_freem(m); 364 return (EAFNOSUPPORT); 365 } 366 if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1); 367 if_inc_counter(ifp, IFCOUNTER_IBYTES, m->m_pkthdr.len); 368 netisr_queue(isr, m); /* mbuf is free'd on failure. */ 369 return (0); 370 } 371 372 /* 373 * Process an ioctl request. 374 */ 375 /* ARGSUSED */ 376 int 377 loioctl(struct ifnet *ifp, u_long cmd, caddr_t data) 378 { 379 struct ifreq *ifr = (struct ifreq *)data; 380 int error = 0, mask; 381 382 switch (cmd) { 383 case SIOCSIFADDR: 384 ifp->if_flags |= IFF_UP; 385 ifp->if_drv_flags |= IFF_DRV_RUNNING; 386 /* 387 * Everything else is done at a higher level. 388 */ 389 break; 390 391 case SIOCADDMULTI: 392 case SIOCDELMULTI: 393 if (ifr == 0) { 394 error = EAFNOSUPPORT; /* XXX */ 395 break; 396 } 397 switch (ifr->ifr_addr.sa_family) { 398 399 #ifdef INET 400 case AF_INET: 401 break; 402 #endif 403 #ifdef INET6 404 case AF_INET6: 405 break; 406 #endif 407 408 default: 409 error = EAFNOSUPPORT; 410 break; 411 } 412 break; 413 414 case SIOCSIFMTU: 415 ifp->if_mtu = ifr->ifr_mtu; 416 break; 417 418 case SIOCSIFFLAGS: 419 break; 420 421 case SIOCSIFCAP: 422 mask = ifp->if_capenable ^ ifr->ifr_reqcap; 423 if ((mask & IFCAP_RXCSUM) != 0) 424 ifp->if_capenable ^= IFCAP_RXCSUM; 425 if ((mask & IFCAP_TXCSUM) != 0) 426 ifp->if_capenable ^= IFCAP_TXCSUM; 427 if ((mask & IFCAP_RXCSUM_IPV6) != 0) { 428 #if 0 429 ifp->if_capenable ^= IFCAP_RXCSUM_IPV6; 430 #else 431 error = EOPNOTSUPP; 432 break; 433 #endif 434 } 435 if ((mask & IFCAP_TXCSUM_IPV6) != 0) { 436 #if 0 437 ifp->if_capenable ^= IFCAP_TXCSUM_IPV6; 438 #else 439 error = EOPNOTSUPP; 440 break; 441 #endif 442 } 443 ifp->if_hwassist = 0; 444 if (ifp->if_capenable & IFCAP_TXCSUM) 445 ifp->if_hwassist = LO_CSUM_FEATURES; 446 #if 0 447 if (ifp->if_capenable & IFCAP_TXCSUM_IPV6) 448 ifp->if_hwassist |= LO_CSUM_FEATURES6; 449 #endif 450 break; 451 452 default: 453 error = EINVAL; 454 } 455 return (error); 456 } 457