usr.sbin/vknetd/vknetd.c

   1 /*
   2  * Copyright (c) 2008 The DragonFly Project.  All rights reserved.
   3  *
   4  * This code is derived from software contributed to The DragonFly Project
   5  * by Matthew Dillon <dillon@backplane.com>
   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  *
  11  * 1. Redistributions of source code must retain the above copyright
  12  *    notice, this list of conditions and the following disclaimer.
  13  * 2. Redistributions in binary form must reproduce the above copyright
  14  *    notice, this list of conditions and the following disclaimer in
  15  *    the documentation and/or other materials provided with the
  16  *    distribution.
  17  * 3. Neither the name of The DragonFly Project nor the names of its
  18  *    contributors may be used to endorse or promote products derived
  19  *    from this software without specific, prior written permission.
  20  *
  21  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
  22  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
  23  * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
  24  * FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE
  25  * COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
  26  * INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING,
  27  * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
  28  * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
  29  * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
  30  * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
  31  * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  32  * SUCH DAMAGE.
  33  */
  34 /*
  35  * vknet [-cdU] [-b bridgeN] [-p socket_path] [-t tapN] [address/cidrbits]
  36  *
  37  * Create a named unix-domain socket which userland vkernels can open
  38  * to gain access to a local network.  All connections to the socket
  39  * are bridged together and the local network can also be bridged onto
  40  * a TAP interface by specifying the -t option.
  41  */
  42 #include "vknetd.h"
  43
  44 static ioinfo_t vknet_tap(const char *tapName, const char *bridgeName);
  45 static int vknet_listener(const char *pathName);
  46 static void vknet_acceptor(int net_fd);
  47 static void *vknet_io(void *arg);
  48 static int vknet_connect(const char *pathName);
  49 static void vknet_monitor(int net_fd);
  50 static void usage(void);
  51 static void writepid(void);
  52 static void cleanup(int);
  53
  54 pthread_mutex_t BridgeMutex;
  55
  56 int SecureOpt = 1;
  57 int DebugOpt = 0;
  58 int SetAddrOpt = 0;
  59 const char *pidfile = "/var/run/vknetd.pid";
  60
  61 struct in_addr NetAddress;
  62 struct in_addr NetMask;
  63
  64 int
  65 main(int ac, char **av)
  66 {
  67         const char *pathName = "/var/run/vknet";
  68         const char *tapName = "auto";
  69         const char *bridgeName = NULL;
  70         int net_fd;
  71         int connectOpt = 0;
  72         int c;
  73         ioinfo_t tap_info;
  74         pthread_t dummy_td;
  75
  76         while ((c = getopt(ac, av, "b:cdp:i:t:U")) != -1) {
  77                 switch (c) {
  78                 case 'U':
  79                         SecureOpt = 0;
  80                         break;
  81                 case 'b':
  82                         bridgeName = optarg;
  83                         break;
  84                 case 'd':
  85                         DebugOpt = 1;
  86                         break;
  87                 case 'p':
  88                         pathName = optarg;
  89                         break;
  90                 case 'i':
  91                         pidfile = optarg;
  92                         break;
  93                 case 't':
  94                         tapName = optarg;
  95                         break;
  96                 case 'c':
  97                         connectOpt = 1;
  98                         break;
  99                 default:
 100                         usage();
 101                 }
 102         }
 103         av += optind;
 104         ac -= optind;
 105         if (ac)
 106                 SetAddrOpt = 1;
 107
 108         /*
 109          * Special connect/debug mode
 110          */
 111         if (connectOpt) {
 112                 net_fd = vknet_connect(pathName);
 113                 if (net_fd < 0) {
 114                         perror("connect");
 115                         exit(1);
 116                 }
 117                 vknet_monitor(net_fd);
 118                 exit(0);
 119         }
 120
 121         /*
 122          * In secure mode (the default), a network address/mask must be
 123          * specified.  e.g. 10.1.0.0/16.  Any traffic going out the TAP
 124          * interface will be filtered.
 125          *
 126          * If non-secure mode the network address/mask is optional.
 127          */
 128         if (SecureOpt || SetAddrOpt) {
 129                 char *str;
 130                 int masklen;
 131                 u_int32_t mask;
 132
 133                 if (ac == 0 || strchr(av[0], '/') == NULL)
 134                         usage();
 135                 str = strdup(av[0]);
 136                 if (inet_pton(AF_INET, strtok(str, "/"), &NetAddress) <= 0)
 137                         usage();
 138                 masklen = strtoul(strtok(NULL, "/"), NULL, 10);
 139                 mask = (1 << (32 - masklen)) - 1;
 140                 NetMask.s_addr = htonl(~mask);
 141         }
 142
 143         /*
 144          * Normal operation, create the tap/bridge and listener.  This
 145          * part is not threaded.
 146          */
 147         mac_init();
 148
 149         if ((tap_info = vknet_tap(tapName, bridgeName)) == NULL) {
 150                 perror("tap: ");
 151                 exit(1);
 152         }
 153         if ((net_fd = vknet_listener(pathName)) < 0) {
 154                 perror("listener: ");
 155                 exit(1);
 156         }
 157
 158         /*
 159          * Now make us a demon and start the threads going.
 160          */
 161         if (DebugOpt == 0)
 162                 daemon(1, 0);
 163
 164         writepid();
 165
 166         signal(SIGINT, cleanup);
 167         signal(SIGHUP, cleanup);
 168         signal(SIGTERM, cleanup);
 169
 170         pthread_mutex_init(&BridgeMutex, NULL);
 171         pthread_create(&dummy_td, NULL, vknet_io, tap_info);
 172         vknet_acceptor(net_fd);
 173
 174         exit(0);
 175 }
 176
 177 #define TAPDEV_MINOR(x) ((int)((x) & 0xffff00ff))
 178
 179 static ioinfo_t
 180 vknet_tap(const char *tapName, const char *bridgeName)
 181 {
 182         struct ifreq ifr;
 183         struct ifaliasreq ifra;
 184         struct stat st;
 185         char *buf = NULL;
 186         int tap_fd;
 187         int tap_unit;
 188         int i;
 189         int s;
 190         int flags;
 191         ioinfo_t info;
 192
 193         if (strcmp(tapName, "auto") == 0) {
 194                 for (i = 0; ; ++i) {
 195                         asprintf(&buf, "/dev/tap%d", i);
 196                         tap_fd = open(buf, O_RDWR | O_NONBLOCK);
 197                         free(buf);
 198                         if (tap_fd >= 0 || errno == ENOENT)
 199                                 break;
 200                 }
 201         } else if (strncmp(tapName, "tap", 3) == 0) {
 202                 asprintf(&buf, "/dev/%s", tapName);
 203                 tap_fd = open(buf, O_RDWR | O_NONBLOCK);
 204                 free(buf);
 205         } else {
 206                 tap_fd = open(tapName, O_RDWR | O_NONBLOCK);
 207         }
 208         if (tap_fd < 0)
 209                 return(NULL);
 210
 211         /*
 212          * Figure out the tap unit number
 213          */
 214         if (fstat(tap_fd, &st) < 0) {
 215                 close(tap_fd);
 216                 return(NULL);
 217         }
 218         tap_unit = TAPDEV_MINOR(st.st_rdev);
 219
 220         /*
 221          * Setup for ioctls
 222          */
 223         fcntl(tap_fd, F_SETFL, 0);
 224         bzero(&ifr, sizeof(ifr));
 225         bzero(&ifra, sizeof(ifra));
 226         snprintf(ifr.ifr_name, sizeof(ifr.ifr_name), "tap%d", tap_unit);
 227         snprintf(ifra.ifra_name, sizeof(ifra.ifra_name), "tap%d", tap_unit);
 228
 229         s = socket(AF_INET, SOCK_DGRAM, 0);
 230
 231         /*
 232          * Set the interface address if in Secure mode.
 233          */
 234         if (SetAddrOpt) {
 235                 struct sockaddr_in *in;
 236
 237                 in = (void *)&ifra.ifra_addr;
 238                 in->sin_family = AF_INET;
 239                 in->sin_len = sizeof(ifra.ifra_addr);
 240                 in->sin_addr = NetAddress;
 241                 in = (void *)&ifra.ifra_mask;
 242                 in->sin_family = AF_INET;
 243                 in->sin_len = sizeof(ifra.ifra_mask);
 244                 in->sin_addr = NetMask;
 245                 if (ioctl(s, SIOCAIFADDR, &ifra) < 0) {
 246                         perror("Unable to set address on tap interface");
 247                         exit(1);
 248                 }
 249         }
 250
 251         /*
 252          * Turn up the interface
 253          */
 254         flags = IFF_UP;
 255         if (ioctl(s, SIOCGIFFLAGS, &ifr) >= 0) {
 256                 bzero(&ifr, sizeof(ifr));
 257                 snprintf(ifr.ifr_name, sizeof(ifr.ifr_name), "tap%d", tap_unit);
 258                 ifr.ifr_flags |= flags & 0xFFFF;
 259                 ifr.ifr_flagshigh |= flags >> 16;
 260                 if (ioctl(s, SIOCSIFFLAGS, &ifr) < 0) {
 261                         perror("Unable to set IFF_UP on tap interface");
 262                         exit(1);
 263                 }
 264         }
 265
 266         if (bridgeName) {
 267                 struct ifbreq ifbr;
 268                 struct ifdrv ifd;
 269
 270                 /*
 271                  * Create the bridge if necessary.
 272                  */
 273                 bzero(&ifr, sizeof(ifr));
 274                 snprintf(ifr.ifr_name, sizeof(ifr.ifr_name), "%s", bridgeName);
 275                 if (ioctl(s, SIOCIFCREATE, &ifr) < 0) {
 276                         if (errno != EEXIST) {
 277                                 perror("Unable to create bridge interface");
 278                                 exit(1);
 279                         }
 280                 }
 281
 282
 283                 /*
 284                  * Add the tap interface to the bridge
 285                  */
 286                 bzero(&ifbr, sizeof(ifbr));
 287                 snprintf(ifbr.ifbr_ifsname, sizeof(ifbr.ifbr_ifsname),
 288                          "tap%d", tap_unit);
 289
 290                 bzero(&ifd, sizeof(ifd));
 291                 snprintf(ifd.ifd_name, sizeof(ifd.ifd_name), "%s", bridgeName);
 292                 ifd.ifd_cmd = BRDGADD;
 293                 ifd.ifd_len = sizeof(ifbr);
 294                 ifd.ifd_data = &ifbr;
 295
 296                 if (ioctl(s, SIOCSDRVSPEC, &ifd) < 0) {
 297                         if (errno != EEXIST) {
 298                                 perror("Unable to add tap ifc to bridge!");
 299                                 exit(1);
 300                         }
 301                 }
 302         }
 303
 304         close(s);
 305         info = malloc(sizeof(*info));
 306         bzero(info, sizeof(*info));
 307         info->fd = tap_fd;
 308         info->istap = 1;
 309         return(info);
 310 }
 311
 312 #undef TAPDEV_MINOR
 313
 314 static int
 315 vknet_listener(const char *pathName)
 316 {
 317         struct sockaddr_un sunx;
 318         int net_fd;
 319         int len;
 320         gid_t gid;
 321         struct group *grp;
 322
 323         /*
 324          * Group access to our named unix domain socket.
 325          */
 326         if ((grp = getgrnam("vknet")) == NULL) {
 327                 fprintf(stderr, "The 'vknet' group must exist\n");
 328                 exit(1);
 329         }
 330         gid = grp->gr_gid;
 331         endgrent();
 332
 333         /*
 334          * Socket setup
 335          */
 336         snprintf(sunx.sun_path, sizeof(sunx.sun_path), "%s", pathName);
 337         len = offsetof(struct sockaddr_un, sun_path[strlen(sunx.sun_path)]);
 338         ++len;  /* include nul */
 339         sunx.sun_family = AF_UNIX;
 340         sunx.sun_len = len;
 341
 342         net_fd = socket(AF_UNIX, SOCK_SEQPACKET, 0);
 343         if (net_fd < 0)
 344                 return(-1);
 345         remove(pathName);
 346         if (bind(net_fd, (void *)&sunx, len) < 0) {
 347                 close(net_fd);
 348                 return(-1);
 349         }
 350         if (listen(net_fd, 1024) < 0) {
 351                 close(net_fd);
 352                 return(-1);
 353         }
 354         if (chown(pathName, (uid_t)-1, gid) < 0) {
 355                 close(net_fd);
 356                 return(-1);
 357         }
 358         if (chmod(pathName, 0660) < 0) {
 359                 close(net_fd);
 360                 return(-1);
 361         }
 362         return(net_fd);
 363 }
 364
 365 static
 366 void
 367 vknet_acceptor(int net_fd)
 368 {
 369         struct sockaddr_un sunx;
 370         pthread_t dummy_td;
 371         int sunx_len;
 372         int rfd;
 373         ioinfo_t info;
 374
 375         for (;;) {
 376                 sunx_len = sizeof(sunx);
 377                 rfd = accept(net_fd, (void *)&sunx, &sunx_len);
 378                 if (rfd < 0)
 379                         break;
 380                 info = malloc(sizeof(*info));
 381                 bzero(info, sizeof(*info));
 382                 info->fd = rfd;
 383                 info->istap = 0;
 384                 pthread_create(&dummy_td, NULL, vknet_io, info);
 385         }
 386 }
 387
 388 /*
 389  * This I/O thread implements the core of the bridging code.
 390  */
 391 static
 392 void *
 393 vknet_io(void *arg)
 394 {
 395         ioinfo_t info = arg;
 396         bridge_t bridge;
 397         u_int8_t *pkt;
 398         int bytes;
 399
 400         pthread_detach(pthread_self());
 401
 402         /*
 403          * Assign as a bridge slot using our thread id.
 404          */
 405         pthread_mutex_lock(&BridgeMutex);
 406         bridge = bridge_add(info);
 407         pthread_mutex_unlock(&BridgeMutex);
 408
 409         /*
 410          * Read packet loop.  Writing is handled by the bridge code.
 411          */
 412         pkt = malloc(MAXPKT);
 413         while ((bytes = read(info->fd, pkt, MAXPKT)) > 0) {
 414                 pthread_mutex_lock(&BridgeMutex);
 415                 bridge_packet(bridge, pkt, bytes);
 416                 pthread_mutex_unlock(&BridgeMutex);
 417         }
 418
 419         /*
 420          * Cleanup
 421          */
 422         pthread_mutex_lock(&BridgeMutex);
 423         bridge_del(bridge);
 424         pthread_mutex_unlock(&BridgeMutex);
 425
 426         close(info->fd);
 427         free(pkt);
 428         pthread_exit(NULL);
 429 }
 430
 431 /*
 432  * Debugging
 433  */
 434 static int
 435 vknet_connect(const char *pathName)
 436 {
 437         struct sockaddr_un sunx;
 438         int len;
 439         int net_fd;
 440
 441         snprintf(sunx.sun_path, sizeof(sunx.sun_path), "%s", pathName);
 442         len = offsetof(struct sockaddr_un, sun_path[strlen(sunx.sun_path)]);
 443         ++len;  /* include nul */
 444         sunx.sun_family = AF_UNIX;
 445         sunx.sun_len = len;
 446
 447         net_fd = socket(AF_UNIX, SOCK_SEQPACKET, 0);
 448         if (net_fd < 0)
 449                 return(-1);
 450         if (connect(net_fd, (void *)&sunx, len) < 0) {
 451                 close(net_fd);
 452                 return(-1);
 453         }
 454         return(net_fd);
 455 }
 456
 457 static void
 458 vknet_monitor(int net_fd)
 459 {
 460         u_int8_t *pkt;
 461         int bytes;
 462         int i;
 463
 464         pkt = malloc(MAXPKT);
 465         while ((bytes = read(net_fd, pkt, MAXPKT)) > 0) {
 466                 printf("%02x:%02x:%02x:%02x:%02x:%02x <- "
 467                        "%02x:%02x:%02x:%02x:%02x:%02x",
 468                        pkt[0], pkt[1], pkt[2], pkt[3], pkt[4], pkt[5],
 469                        pkt[6], pkt[7], pkt[8], pkt[9], pkt[10], pkt[11]);
 470                 for (i = 12; i < bytes; ++i) {
 471                         if (((i - 12) & 15) == 0) {
 472                                 printf("\n\t");
 473                         }
 474                         printf(" %02x", pkt[i]);
 475                 }
 476                 printf("\n");
 477         }
 478         free(pkt);
 479 }
 480
 481 /*
 482  * Misc
 483  */
 484 static void
 485 writepid(void)
 486 {
 487         FILE *pf;
 488
 489         if ((pf = fopen(pidfile, "w+")) == NULL)
 490                 errx(1, "Failed to create pidfile %s", pidfile);
 491
 492         if ((fprintf(pf, "%d\n", getpid())) < 1)
 493                 err(1, "fprintf");
 494
 495         fclose(pf);
 496 }
 497
 498 static void
 499 cleanup(int __unused sig)
 500 {
 501         if (pidfile)
 502                 unlink(pidfile);
 503 }
 504
 505 static
 506 void
 507 usage(void)
 508 {
 509         fprintf(stderr, "usage: vknet [-cdU] [-b bridgeN] [-p socket_path] [-i pidfile] [-t tapN] [address/cidrbits]\n");
 510         fprintf(stderr, "address/cidrbits must be specified in default secure mode.\n");
 511         exit(1);
 512 }
 513