/* -*- Mode: c; tab-width: 8; indent-tabs-mode: 1; c-basic-offset: 8; -*- */ /* * Copyright (c) 1994, 1995, 1996, 1997, 1998 * The Regents of the University of California. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. All advertising materials mentioning features or use of this software * must display the following acknowledgement: * This product includes software developed by the Computer Systems * Engineering Group at Lawrence Berkeley Laboratory. * 4. Neither the name of the University nor of the Laboratory may be used * to endorse or promote products derived from this software without * specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ #ifndef lint static const char rcsid[] _U_ = "@(#) $Header: /tcpdump/master/libpcap/fad-gifc.c,v 1.8.2.2 2005/06/29 06:43:31 guy Exp $ (LBL)"; #endif #ifdef HAVE_CONFIG_H #include "config.h" #endif #include #include #include #include #ifdef HAVE_SYS_SOCKIO_H #include #endif #include /* concession to AIX */ struct mbuf; /* Squelch compiler warnings on some platforms for */ struct rtentry; /* declarations in */ #include #include #include #include #include #include #include #include #include #include "pcap-int.h" #ifdef HAVE_OS_PROTO_H #include "os-proto.h" #endif /* * This is fun. * * In older BSD systems, socket addresses were fixed-length, and * "sizeof (struct sockaddr)" gave the size of the structure. * All addresses fit within a "struct sockaddr". * * In newer BSD systems, the socket address is variable-length, and * there's an "sa_len" field giving the length of the structure; * this allows socket addresses to be longer than 2 bytes of family * and 14 bytes of data. * * Some commercial UNIXes use the old BSD scheme, some use the RFC 2553 * variant of the old BSD scheme (with "struct sockaddr_storage" rather * than "struct sockaddr"), and some use the new BSD scheme. * * Some versions of GNU libc use neither scheme, but has an "SA_LEN()" * macro that determines the size based on the address family. Other * versions don't have "SA_LEN()" (as it was in drafts of RFC 2553 * but not in the final version). * * We assume that a UNIX that doesn't have "getifaddrs()" and doesn't have * SIOCGLIFCONF, but has SIOCGIFCONF, uses "struct sockaddr" for the * address in an entry returned by SIOCGIFCONF. */ #ifndef SA_LEN #ifdef HAVE_SOCKADDR_SA_LEN #define SA_LEN(addr) ((addr)->sa_len) #else /* HAVE_SOCKADDR_SA_LEN */ #define SA_LEN(addr) (sizeof (struct sockaddr)) #endif /* HAVE_SOCKADDR_SA_LEN */ #endif /* SA_LEN */ /* * This is also fun. * * There is no ioctl that returns the amount of space required for all * the data that SIOCGIFCONF could return, and if a buffer is supplied * that's not large enough for all the data SIOCGIFCONF could return, * on at least some platforms it just returns the data that'd fit with * no indication that there wasn't enough room for all the data, much * less an indication of how much more room is required. * * The only way to ensure that we got all the data is to pass a buffer * large enough that the amount of space in the buffer *not* filled in * is greater than the largest possible entry. * * We assume that's "sizeof(ifreq.ifr_name)" plus 255, under the assumption * that no address is more than 255 bytes (on systems where the "sa_len" * field in a "struct sockaddr" is 1 byte, e.g. newer BSDs, that's the * case, and addresses are unlikely to be bigger than that in any case). */ #define MAX_SA_LEN 255 #ifdef HAVE_PROC_NET_DEV /* * Get from "/proc/net/dev" all interfaces listed there; if they're * already in the list of interfaces we have, that won't add another * instance, but if they're not, that'll add them. * * We don't bother getting any addresses for them; it appears you can't * use SIOCGIFADDR on Linux to get IPv6 addresses for interfaces, and, * although some other types of addresses can be fetched with SIOCGIFADDR, * we don't bother with them for now. * * We also don't fail if we couldn't open "/proc/net/dev"; we just leave * the list of interfaces as is. */ static int scan_proc_net_dev(pcap_if_t **devlistp, int fd, char *errbuf) { FILE *proc_net_f; char linebuf[512]; int linenum; unsigned char *p; char name[512]; /* XXX - pick a size */ char *q, *saveq; struct ifreq ifrflags; int ret = 0; proc_net_f = fopen("/proc/net/dev", "r"); if (proc_net_f == NULL) return (0); for (linenum = 1; fgets(linebuf, sizeof linebuf, proc_net_f) != NULL; linenum++) { /* * Skip the first two lines - they're headers. */ if (linenum <= 2) continue; p = &linebuf[0]; /* * Skip leading white space. */ while (*p != '\0' && isspace(*p)) p++; if (*p == '\0' || *p == '\n') continue; /* blank line */ /* * Get the interface name. */ q = &name[0]; while (*p != '\0' && !isspace(*p)) { if (*p == ':') { /* * This could be the separator between a * name and an alias number, or it could be * the separator between a name with no * alias number and the next field. * * If there's a colon after digits, it * separates the name and the alias number, * otherwise it separates the name and the * next field. */ saveq = q; while (isdigit(*p)) *q++ = *p++; if (*p != ':') { /* * That was the next field, * not the alias number. */ q = saveq; } break; } else *q++ = *p++; } *q = '\0'; /* * Get the flags for this interface, and skip it if * it's not up. */ strncpy(ifrflags.ifr_name, name, sizeof(ifrflags.ifr_name)); if (ioctl(fd, SIOCGIFFLAGS, (char *)&ifrflags) < 0) { if (errno == ENXIO) continue; (void)snprintf(errbuf, PCAP_ERRBUF_SIZE, "SIOCGIFFLAGS: %.*s: %s", (int)sizeof(ifrflags.ifr_name), ifrflags.ifr_name, pcap_strerror(errno)); ret = -1; break; } if (!(ifrflags.ifr_flags & IFF_UP)) continue; /* * Add an entry for this interface, with no addresses. */ if (pcap_add_if(devlistp, name, ifrflags.ifr_flags, NULL, errbuf) == -1) { /* * Failure. */ ret = -1; break; } } if (ret != -1) { /* * Well, we didn't fail for any other reason; did we * fail due to an error reading the file? */ if (ferror(proc_net_f)) { (void)snprintf(errbuf, PCAP_ERRBUF_SIZE, "Error reading /proc/net/dev: %s", pcap_strerror(errno)); ret = -1; } } (void)fclose(proc_net_f); return (ret); } #endif /* HAVE_PROC_NET_DEV */ /* * Get a list of all interfaces that are up and that we can open. * Returns -1 on error, 0 otherwise. * The list, as returned through "alldevsp", may be null if no interfaces * were up and could be opened. * * This is the implementation used on platforms that have SIOCGIFCONF but * don't have any other mechanism for getting a list of interfaces. * * XXX - or platforms that have other, better mechanisms but for which * we don't yet have code to use that mechanism; I think there's a better * way on Linux, for example. */ int pcap_findalldevs(pcap_if_t **alldevsp, char *errbuf) { pcap_if_t *devlist = NULL; register int fd; register struct ifreq *ifrp, *ifend, *ifnext; int n; struct ifconf ifc; char *buf = NULL; unsigned buf_size; #if defined (HAVE_SOLARIS) || defined (HAVE_HPUX10_20_OR_LATER) char *p, *q; #endif struct ifreq ifrflags, ifrnetmask, ifrbroadaddr, ifrdstaddr; struct sockaddr *netmask, *broadaddr, *dstaddr; size_t netmask_size, broadaddr_size, dstaddr_size; int ret = 0; /* * Create a socket from which to fetch the list of interfaces. */ fd = socket(AF_INET, SOCK_DGRAM, 0); if (fd < 0) { (void)snprintf(errbuf, PCAP_ERRBUF_SIZE, "socket: %s", pcap_strerror(errno)); return (-1); } /* * Start with an 8K buffer, and keep growing the buffer until * we have more than "sizeof(ifrp->ifr_name) + MAX_SA_LEN" * bytes left over in the buffer or we fail to get the * interface list for some reason other than EINVAL (which is * presumed here to mean "buffer is too small"). */ buf_size = 8192; for (;;) { buf = malloc(buf_size); if (buf == NULL) { (void)snprintf(errbuf, PCAP_ERRBUF_SIZE, "malloc: %s", pcap_strerror(errno)); (void)close(fd); return (-1); } ifc.ifc_len = buf_size; ifc.ifc_buf = buf; memset(buf, 0, buf_size); if (ioctl(fd, SIOCGIFCONF, (char *)&ifc) < 0 && errno != EINVAL) { (void)snprintf(errbuf, PCAP_ERRBUF_SIZE, "SIOCGIFCONF: %s", pcap_strerror(errno)); (void)close(fd); free(buf); return (-1); } if (ifc.ifc_len < buf_size && (buf_size - ifc.ifc_len) > sizeof(ifrp->ifr_name) + MAX_SA_LEN) break; free(buf); buf_size *= 2; } ifrp = (struct ifreq *)buf; ifend = (struct ifreq *)(buf + ifc.ifc_len); for (; ifrp < ifend; ifrp = ifnext) { /* * XXX - what if this isn't an IPv4 address? Can * we still get the netmask, etc. with ioctls on * an IPv4 socket? * * The answer is probably platform-dependent, and * if the answer is "no" on more than one platform, * the way you work around it is probably platform- * dependent as well. */ n = SA_LEN(&ifrp->ifr_addr) + sizeof(ifrp->ifr_name); if (n < sizeof(*ifrp)) ifnext = ifrp + 1; else ifnext = (struct ifreq *)((char *)ifrp + n); /* * XXX - The 32-bit compatibility layer for Linux on IA-64 * is slightly broken. It correctly converts the structures * to and from kernel land from 64 bit to 32 bit but * doesn't update ifc.ifc_len, leaving it larger than the * amount really used. This means we read off the end * of the buffer and encounter an interface with an * "empty" name. Since this is highly unlikely to ever * occur in a valid case we can just finish looking for * interfaces if we see an empty name. */ if (!(*ifrp->ifr_name)) break; /* * Skip entries that begin with "dummy". * XXX - what are these? Is this Linux-specific? * Are there platforms on which we shouldn't do this? */ if (strncmp(ifrp->ifr_name, "dummy", 5) == 0) continue; /* * Get the flags for this interface, and skip it if it's * not up. */ strncpy(ifrflags.ifr_name, ifrp->ifr_name, sizeof(ifrflags.ifr_name)); if (ioctl(fd, SIOCGIFFLAGS, (char *)&ifrflags) < 0) { if (errno == ENXIO) continue; (void)snprintf(errbuf, PCAP_ERRBUF_SIZE, "SIOCGIFFLAGS: %.*s: %s", (int)sizeof(ifrflags.ifr_name), ifrflags.ifr_name, pcap_strerror(errno)); ret = -1; break; } if (!(ifrflags.ifr_flags & IFF_UP)) continue; /* * Get the netmask for this address on this interface. */ strncpy(ifrnetmask.ifr_name, ifrp->ifr_name, sizeof(ifrnetmask.ifr_name)); memcpy(&ifrnetmask.ifr_addr, &ifrp->ifr_addr, sizeof(ifrnetmask.ifr_addr)); if (ioctl(fd, SIOCGIFNETMASK, (char *)&ifrnetmask) < 0) { if (errno == EADDRNOTAVAIL) { /* * Not available. */ netmask = NULL; netmask_size = 0; } else { (void)snprintf(errbuf, PCAP_ERRBUF_SIZE, "SIOCGIFNETMASK: %.*s: %s", (int)sizeof(ifrnetmask.ifr_name), ifrnetmask.ifr_name, pcap_strerror(errno)); ret = -1; break; } } else { netmask = &ifrnetmask.ifr_addr; netmask_size = SA_LEN(netmask); } /* * Get the broadcast address for this address on this * interface (if any). */ if (ifrflags.ifr_flags & IFF_BROADCAST) { strncpy(ifrbroadaddr.ifr_name, ifrp->ifr_name, sizeof(ifrbroadaddr.ifr_name)); memcpy(&ifrbroadaddr.ifr_addr, &ifrp->ifr_addr, sizeof(ifrbroadaddr.ifr_addr)); if (ioctl(fd, SIOCGIFBRDADDR, (char *)&ifrbroadaddr) < 0) { if (errno == EADDRNOTAVAIL) { /* * Not available. */ broadaddr = NULL; broadaddr_size = 0; } else { (void)snprintf(errbuf, PCAP_ERRBUF_SIZE, "SIOCGIFBRDADDR: %.*s: %s", (int)sizeof(ifrbroadaddr.ifr_name), ifrbroadaddr.ifr_name, pcap_strerror(errno)); ret = -1; break; } } else { broadaddr = &ifrbroadaddr.ifr_broadaddr; broadaddr_size = SA_LEN(broadaddr); } } else { /* * Not a broadcast interface, so no broadcast * address. */ broadaddr = NULL; broadaddr_size = 0; } /* * Get the destination address for this address on this * interface (if any). */ if (ifrflags.ifr_flags & IFF_POINTOPOINT) { strncpy(ifrdstaddr.ifr_name, ifrp->ifr_name, sizeof(ifrdstaddr.ifr_name)); memcpy(&ifrdstaddr.ifr_addr, &ifrp->ifr_addr, sizeof(ifrdstaddr.ifr_addr)); if (ioctl(fd, SIOCGIFDSTADDR, (char *)&ifrdstaddr) < 0) { if (errno == EADDRNOTAVAIL) { /* * Not available. */ dstaddr = NULL; dstaddr_size = 0; } else { (void)snprintf(errbuf, PCAP_ERRBUF_SIZE, "SIOCGIFDSTADDR: %.*s: %s", (int)sizeof(ifrdstaddr.ifr_name), ifrdstaddr.ifr_name, pcap_strerror(errno)); ret = -1; break; } } else { dstaddr = &ifrdstaddr.ifr_dstaddr; dstaddr_size = SA_LEN(dstaddr); } } else { /* * Not a point-to-point interface, so no destination * address. */ dstaddr = NULL; dstaddr_size = 0; } #if defined (HAVE_SOLARIS) || defined (HAVE_HPUX10_20_OR_LATER) /* * If this entry has a colon followed by a number at * the end, it's a logical interface. Those are just * the way you assign multiple IP addresses to a real * interface, so an entry for a logical interface should * be treated like the entry for the real interface; * we do that by stripping off the ":" and the number. */ p = strchr(ifrp->ifr_name, ':'); if (p != NULL) { /* * We have a ":"; is it followed by a number? */ q = p + 1; while (isdigit((unsigned char)*q)) q++; if (*q == '\0') { /* * All digits after the ":" until the end. * Strip off the ":" and everything after * it. */ *p = '\0'; } } #endif /* * Add information for this address to the list. */ if (add_addr_to_iflist(&devlist, ifrp->ifr_name, ifrflags.ifr_flags, &ifrp->ifr_addr, SA_LEN(&ifrp->ifr_addr), netmask, netmask_size, broadaddr, broadaddr_size, dstaddr, dstaddr_size, errbuf) < 0) { ret = -1; break; } } free(buf); #ifdef HAVE_PROC_NET_DEV if (ret != -1) { /* * We haven't had any errors yet; now read "/proc/net/dev", * and add to the list of interfaces all interfaces listed * there that we don't already have, because, on Linux, * SIOCGIFCONF reports only interfaces with IPv4 addresses, * so you need to read "/proc/net/dev" to get the names of * the rest of the interfaces. */ ret = scan_proc_net_dev(&devlist, fd, errbuf); } #endif (void)close(fd); if (ret != -1) { /* * We haven't had any errors yet; do any platform-specific * operations to add devices. */ if (pcap_platform_finddevs(&devlist, errbuf) < 0) ret = -1; } if (ret == -1) { /* * We had an error; free the list we've been constructing. */ if (devlist != NULL) { pcap_freealldevs(devlist); devlist = NULL; } } *alldevsp = devlist; return (ret); }