freebsd-dev/usr.sbin/arp/arp.c
Qing Li 8eca593c5a This checkin addresses a couple of issues:
1. The "route" command allows route insertion through the interface-direct
   option "-iface". During if_attach(), an sockaddr_dl{} entry is created
   for the interface and is part of the interface address list. This
   sockaddr_dl{} entry describes the interface in detail. The "route"
   command selects this entry as the "gateway" object when the "-iface"
   option is present. The "arp" and "ndp" commands also interact with the
   kernel through the routing socket when adding and removing static L2
   entries. The static L2 information is also provided through the
   "gateway" object with an AF_LINK family type, similar to what is
   provided by the "route" command. In order to differentiate between
   these two types of operations, a RTF_LLDATA flag is introduced. This
   flag is set by the "arp" and "ndp" commands when issuing the add and
   delete commands. This flag is also set in each L2 entry returned by the
   kernel. The "arp" and "ndp" command follows a convention where a RTM_GET
   is issued first followed by a RTM_ADD/DELETE. This RTM_GET request fills
   in the fields for a "rtm" object, which is reinjected into the kernel by
   a subsequent RTM_ADD/DELETE command. The entry returend from RTM_GET
   is a prefix route, so the RTF_LLDATA flag must be specified when issuing
   the RTM_ADD/DELETE messages.

2. Enforce the convention that NET_RT_FLAGS with a 0 w_arg is the
   specification for retrieving L2 information. Also optimized the
   code logic.

Reviewed by:   julian
2008-12-26 19:45:24 +00:00

847 lines
20 KiB
C

/*
* Copyright (c) 1984, 1993
* The Regents of the University of California. All rights reserved.
*
* This code is derived from software contributed to Berkeley by
* Sun Microsystems, Inc.
*
* 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.
* 4. Neither the name of the University nor the names of its contributors
* 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.
*/
#if 0
#ifndef lint
static char const copyright[] =
"@(#) Copyright (c) 1984, 1993\n\
The Regents of the University of California. All rights reserved.\n";
#endif /* not lint */
#ifndef lint
static char const sccsid[] = "@(#)from: arp.c 8.2 (Berkeley) 1/2/94";
#endif /* not lint */
#endif
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
/*
* arp - display, set, and delete arp table entries
*/
#include <sys/param.h>
#include <sys/file.h>
#include <sys/socket.h>
#include <sys/sockio.h>
#include <sys/sysctl.h>
#include <sys/ioctl.h>
#include <sys/time.h>
#include <net/if.h>
#include <net/if_dl.h>
#include <net/if_types.h>
#include <net/route.h>
#include <net/iso88025.h>
#include <netinet/in.h>
#include <netinet/if_ether.h>
#include <arpa/inet.h>
#include <ctype.h>
#include <err.h>
#include <errno.h>
#include <netdb.h>
#include <nlist.h>
#include <paths.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <strings.h>
#include <unistd.h>
typedef void (action_fn)(struct sockaddr_dl *sdl,
struct sockaddr_inarp *s_in, struct rt_msghdr *rtm);
static int search(u_long addr, action_fn *action);
static action_fn print_entry;
static action_fn nuke_entry;
static int delete(char *host, int do_proxy);
static void usage(void);
static int set(int argc, char **argv);
static int get(char *host);
static int file(char *name);
static struct rt_msghdr *rtmsg(int cmd,
struct sockaddr_inarp *dst, struct sockaddr_dl *sdl);
static int get_ether_addr(in_addr_t ipaddr, struct ether_addr *hwaddr);
static struct sockaddr_inarp *getaddr(char *host);
static int valid_type(int type);
static int nflag; /* no reverse dns lookups */
static char *rifname;
static int expire_time, flags, doing_proxy, proxy_only;
/* which function we're supposed to do */
#define F_GET 1
#define F_SET 2
#define F_FILESET 3
#define F_REPLACE 4
#define F_DELETE 5
#define SETFUNC(f) { if (func) usage(); func = (f); }
int
main(int argc, char *argv[])
{
int ch, func = 0;
int rtn = 0;
int aflag = 0; /* do it for all entries */
while ((ch = getopt(argc, argv, "andfsSi:")) != -1)
switch((char)ch) {
case 'a':
aflag = 1;
break;
case 'd':
SETFUNC(F_DELETE);
break;
case 'n':
nflag = 1;
break;
case 'S':
SETFUNC(F_REPLACE);
break;
case 's':
SETFUNC(F_SET);
break;
case 'f' :
SETFUNC(F_FILESET);
break;
case 'i':
rifname = optarg;
break;
case '?':
default:
usage();
}
argc -= optind;
argv += optind;
if (!func)
func = F_GET;
if (rifname) {
if (func != F_GET && !(func == F_DELETE && aflag))
errx(1, "-i not applicable to this operation");
if (if_nametoindex(rifname) == 0) {
if (errno == ENXIO)
errx(1, "interface %s does not exist", rifname);
else
err(1, "if_nametoindex(%s)", rifname);
}
}
switch (func) {
case F_GET:
if (aflag) {
if (argc != 0)
usage();
search(0, print_entry);
} else {
if (argc != 1)
usage();
rtn = get(argv[0]);
}
break;
case F_SET:
case F_REPLACE:
if (argc < 2 || argc > 6)
usage();
if (func == F_REPLACE)
(void)delete(argv[0], 0);
rtn = set(argc, argv) ? 1 : 0;
break;
case F_DELETE:
if (aflag) {
if (argc != 0)
usage();
search(0, nuke_entry);
} else {
if (argc == 2 && strncmp(argv[1], "pub", 3) == 0)
ch = SIN_PROXY;
else if (argc == 1)
ch = 0;
else
usage();
rtn = delete(argv[0], ch);
}
break;
case F_FILESET:
if (argc != 1)
usage();
rtn = file(argv[0]);
break;
}
return (rtn);
}
/*
* Process a file to set standard arp entries
*/
static int
file(char *name)
{
FILE *fp;
int i, retval;
char line[100], arg[5][50], *args[5], *p;
if ((fp = fopen(name, "r")) == NULL)
err(1, "cannot open %s", name);
args[0] = &arg[0][0];
args[1] = &arg[1][0];
args[2] = &arg[2][0];
args[3] = &arg[3][0];
args[4] = &arg[4][0];
retval = 0;
while(fgets(line, sizeof(line), fp) != NULL) {
if ((p = strchr(line, '#')) != NULL)
*p = '\0';
for (p = line; isblank(*p); p++);
if (*p == '\n' || *p == '\0')
continue;
i = sscanf(p, "%49s %49s %49s %49s %49s", arg[0], arg[1],
arg[2], arg[3], arg[4]);
if (i < 2) {
warnx("bad line: %s", line);
retval = 1;
continue;
}
if (set(i, args))
retval = 1;
}
fclose(fp);
return (retval);
}
/*
* Given a hostname, fills up a (static) struct sockaddr_inarp with
* the address of the host and returns a pointer to the
* structure.
*/
static struct sockaddr_inarp *
getaddr(char *host)
{
struct hostent *hp;
static struct sockaddr_inarp reply;
bzero(&reply, sizeof(reply));
reply.sin_len = sizeof(reply);
reply.sin_family = AF_INET;
reply.sin_addr.s_addr = inet_addr(host);
if (reply.sin_addr.s_addr == INADDR_NONE) {
if (!(hp = gethostbyname(host))) {
warnx("%s: %s", host, hstrerror(h_errno));
return (NULL);
}
bcopy((char *)hp->h_addr, (char *)&reply.sin_addr,
sizeof reply.sin_addr);
}
return (&reply);
}
/*
* Returns true if the type is a valid one for ARP.
*/
static int
valid_type(int type)
{
switch (type) {
case IFT_ETHER:
case IFT_FDDI:
case IFT_ISO88023:
case IFT_ISO88024:
case IFT_ISO88025:
case IFT_L2VLAN:
case IFT_BRIDGE:
return (1);
default:
return (0);
}
}
/*
* Set an individual arp entry
*/
static int
set(int argc, char **argv)
{
struct sockaddr_inarp *addr;
struct sockaddr_inarp *dst; /* what are we looking for */
struct sockaddr_dl *sdl;
struct rt_msghdr *rtm;
struct ether_addr *ea;
char *host = argv[0], *eaddr = argv[1];
struct sockaddr_dl sdl_m;
argc -= 2;
argv += 2;
bzero(&sdl_m, sizeof(sdl_m));
sdl_m.sdl_len = sizeof(sdl_m);
sdl_m.sdl_family = AF_LINK;
dst = getaddr(host);
if (dst == NULL)
return (1);
doing_proxy = flags = proxy_only = expire_time = 0;
while (argc-- > 0) {
if (strncmp(argv[0], "temp", 4) == 0) {
struct timeval tv;
gettimeofday(&tv, 0);
expire_time = tv.tv_sec + 20 * 60;
} else if (strncmp(argv[0], "pub", 3) == 0) {
flags |= RTF_ANNOUNCE;
doing_proxy = 1;
if (argc && strncmp(argv[1], "only", 3) == 0) {
proxy_only = 1;
dst->sin_other = SIN_PROXY;
argc--; argv++;
}
} else if (strncmp(argv[0], "blackhole", 9) == 0) {
if (flags & RTF_REJECT) {
printf("Choose one of blackhole or reject, not both.\n");
}
flags |= RTF_BLACKHOLE;
} else if (strncmp(argv[0], "reject", 6) == 0) {
if (flags & RTF_BLACKHOLE) {
printf("Choose one of blackhole or reject, not both.\n");
}
flags |= RTF_REJECT;
} else if (strncmp(argv[0], "trail", 5) == 0) {
/* XXX deprecated and undocumented feature */
printf("%s: Sending trailers is no longer supported\n",
host);
}
argv++;
}
ea = (struct ether_addr *)LLADDR(&sdl_m);
if (doing_proxy && !strcmp(eaddr, "auto")) {
if (!get_ether_addr(dst->sin_addr.s_addr, ea)) {
printf("no interface found for %s\n",
inet_ntoa(dst->sin_addr));
return (1);
}
sdl_m.sdl_alen = ETHER_ADDR_LEN;
} else {
struct ether_addr *ea1 = ether_aton(eaddr);
if (ea1 == NULL) {
warnx("invalid Ethernet address '%s'", eaddr);
return (1);
} else {
*ea = *ea1;
sdl_m.sdl_alen = ETHER_ADDR_LEN;
}
}
for (;;) { /* try at most twice */
rtm = rtmsg(RTM_GET, dst, &sdl_m);
if (rtm == NULL) {
warn("%s", host);
return (1);
}
addr = (struct sockaddr_inarp *)(rtm + 1);
sdl = (struct sockaddr_dl *)(SA_SIZE(addr) + (char *)addr);
if (addr->sin_addr.s_addr != dst->sin_addr.s_addr)
break;
if (sdl->sdl_family == AF_LINK &&
!(rtm->rtm_flags & RTF_GATEWAY) &&
valid_type(sdl->sdl_type) )
break;
if (doing_proxy == 0) {
printf("set: can only proxy for %s\n", host);
return (1);
}
if (dst->sin_other & SIN_PROXY) {
printf("set: proxy entry exists for non 802 device\n");
return (1);
}
dst->sin_other = SIN_PROXY;
proxy_only = 1;
}
if (sdl->sdl_family != AF_LINK) {
printf("cannot intuit interface index and type for %s\n", host);
return (1);
}
sdl_m.sdl_type = sdl->sdl_type;
sdl_m.sdl_index = sdl->sdl_index;
return (rtmsg(RTM_ADD, dst, &sdl_m) == NULL);
}
/*
* Display an individual arp entry
*/
static int
get(char *host)
{
struct sockaddr_inarp *addr;
addr = getaddr(host);
if (addr == NULL)
return (1);
if (0 == search(addr->sin_addr.s_addr, print_entry)) {
printf("%s (%s) -- no entry",
host, inet_ntoa(addr->sin_addr));
if (rifname)
printf(" on %s", rifname);
printf("\n");
return (1);
}
return (0);
}
/*
* Delete an arp entry
*/
static int
delete(char *host, int do_proxy)
{
struct sockaddr_inarp *addr, *dst;
struct rt_msghdr *rtm;
struct sockaddr_dl *sdl;
struct sockaddr_dl sdl_m;
dst = getaddr(host);
if (dst == NULL)
return (1);
dst->sin_other = do_proxy;
/*
* setup the data structure to notify the kernel
* it is the ARP entry the RTM_GET is interested
* in
*/
bzero(&sdl_m, sizeof(sdl_m));
sdl_m.sdl_len = sizeof(sdl_m);
sdl_m.sdl_family = AF_LINK;
for (;;) { /* try twice */
rtm = rtmsg(RTM_GET, dst, &sdl_m);
if (rtm == NULL) {
warn("%s", host);
return (1);
}
addr = (struct sockaddr_inarp *)(rtm + 1);
sdl = (struct sockaddr_dl *)(SA_SIZE(addr) + (char *)addr);
/*
* With the new L2/L3 restructure, the route
* returned is a prefix route. The important
* piece of information from the previous
* RTM_GET is the interface index. In the
* case of ECMP, the kernel will traverse
* the route group for the given entry.
*/
if (sdl->sdl_family == AF_LINK &&
!(rtm->rtm_flags & RTF_GATEWAY) &&
valid_type(sdl->sdl_type) ) {
addr->sin_addr.s_addr = dst->sin_addr.s_addr;
break;
}
if (dst->sin_other & SIN_PROXY) {
fprintf(stderr, "delete: cannot locate %s\n",host);
return (1);
}
dst->sin_other = SIN_PROXY;
}
rtm->rtm_flags |= RTF_LLDATA;
if (rtmsg(RTM_DELETE, dst, NULL) != NULL) {
printf("%s (%s) deleted\n", host, inet_ntoa(addr->sin_addr));
return (0);
}
return (1);
}
/*
* Search the arp table and do some action on matching entries
*/
static int
search(u_long addr, action_fn *action)
{
int mib[6];
size_t needed;
char *lim, *buf, *newbuf, *next;
struct rt_msghdr *rtm;
struct sockaddr_inarp *sin2;
struct sockaddr_dl *sdl;
char ifname[IF_NAMESIZE];
int st, found_entry = 0;
mib[0] = CTL_NET;
mib[1] = PF_ROUTE;
mib[2] = 0;
mib[3] = AF_INET;
mib[4] = NET_RT_FLAGS;
#ifdef RTF_LLINFO
mib[5] = RTF_LLINFO;
#else
mib[5] = 0;
#endif
if (sysctl(mib, 6, NULL, &needed, NULL, 0) < 0)
err(1, "route-sysctl-estimate");
if (needed == 0) /* empty table */
return 0;
buf = NULL;
for (;;) {
newbuf = realloc(buf, needed);
if (newbuf == NULL) {
if (buf != NULL)
free(buf);
errx(1, "could not reallocate memory");
}
buf = newbuf;
st = sysctl(mib, 6, buf, &needed, NULL, 0);
if (st == 0 || errno != ENOMEM)
break;
needed += needed / 8;
}
if (st == -1)
err(1, "actual retrieval of routing table");
lim = buf + needed;
for (next = buf; next < lim; next += rtm->rtm_msglen) {
rtm = (struct rt_msghdr *)next;
sin2 = (struct sockaddr_inarp *)(rtm + 1);
sdl = (struct sockaddr_dl *)((char *)sin2 + SA_SIZE(sin2));
if (rifname && if_indextoname(sdl->sdl_index, ifname) &&
strcmp(ifname, rifname))
continue;
if (addr) {
if (addr != sin2->sin_addr.s_addr)
continue;
found_entry = 1;
}
(*action)(sdl, sin2, rtm);
}
free(buf);
return (found_entry);
}
/*
* Display an arp entry
*/
static void
print_entry(struct sockaddr_dl *sdl,
struct sockaddr_inarp *addr, struct rt_msghdr *rtm)
{
const char *host;
struct hostent *hp;
struct iso88025_sockaddr_dl_data *trld;
char ifname[IF_NAMESIZE];
int seg;
if (nflag == 0)
hp = gethostbyaddr((caddr_t)&(addr->sin_addr),
sizeof addr->sin_addr, AF_INET);
else
hp = 0;
if (hp)
host = hp->h_name;
else {
host = "?";
if (h_errno == TRY_AGAIN)
nflag = 1;
}
printf("%s (%s) at ", host, inet_ntoa(addr->sin_addr));
if (sdl->sdl_alen) {
if ((sdl->sdl_type == IFT_ETHER ||
sdl->sdl_type == IFT_L2VLAN ||
sdl->sdl_type == IFT_BRIDGE) &&
sdl->sdl_alen == ETHER_ADDR_LEN)
printf("%s", ether_ntoa((struct ether_addr *)LLADDR(sdl)));
else {
int n = sdl->sdl_nlen > 0 ? sdl->sdl_nlen + 1 : 0;
printf("%s", link_ntoa(sdl) + n);
}
} else
printf("(incomplete)");
if (if_indextoname(sdl->sdl_index, ifname) != NULL)
printf(" on %s", ifname);
if (rtm->rtm_rmx.rmx_expire == 0)
printf(" permanent");
if (addr->sin_other & SIN_PROXY)
printf(" published (proxy only)");
if (rtm->rtm_flags & RTF_ANNOUNCE)
printf(" published");
switch(sdl->sdl_type) {
case IFT_ETHER:
printf(" [ethernet]");
break;
case IFT_ISO88025:
printf(" [token-ring]");
trld = SDL_ISO88025(sdl);
if (trld->trld_rcf != 0) {
printf(" rt=%x", ntohs(trld->trld_rcf));
for (seg = 0;
seg < ((TR_RCF_RIFLEN(trld->trld_rcf) - 2 ) / 2);
seg++)
printf(":%x", ntohs(*(trld->trld_route[seg])));
}
break;
case IFT_FDDI:
printf(" [fddi]");
break;
case IFT_ATM:
printf(" [atm]");
break;
case IFT_L2VLAN:
printf(" [vlan]");
break;
case IFT_IEEE1394:
printf(" [firewire]");
break;
case IFT_BRIDGE:
printf(" [bridge]");
break;
default:
break;
}
printf("\n");
}
/*
* Nuke an arp entry
*/
static void
nuke_entry(struct sockaddr_dl *sdl __unused,
struct sockaddr_inarp *addr, struct rt_msghdr *rtm __unused)
{
char ip[20];
snprintf(ip, sizeof(ip), "%s", inet_ntoa(addr->sin_addr));
(void)delete(ip, 0);
}
static void
usage(void)
{
fprintf(stderr, "%s\n%s\n%s\n%s\n%s\n%s\n%s\n",
"usage: arp [-n] [-i interface] hostname",
" arp [-n] [-i interface] -a",
" arp -d hostname [pub]",
" arp -d [-i interface] -a",
" arp -s hostname ether_addr [temp] [reject | blackhole] [pub [only]]",
" arp -S hostname ether_addr [temp] [reject | blackhole] [pub [only]]",
" arp -f filename");
exit(1);
}
static struct rt_msghdr *
rtmsg(int cmd, struct sockaddr_inarp *dst, struct sockaddr_dl *sdl)
{
static int seq;
int rlen;
int l;
struct sockaddr_in so_mask, *som = &so_mask;
static int s = -1;
static pid_t pid;
static struct {
struct rt_msghdr m_rtm;
char m_space[512];
} m_rtmsg;
struct rt_msghdr *rtm = &m_rtmsg.m_rtm;
char *cp = m_rtmsg.m_space;
if (s < 0) { /* first time: open socket, get pid */
s = socket(PF_ROUTE, SOCK_RAW, 0);
if (s < 0)
err(1, "socket");
pid = getpid();
}
bzero(&so_mask, sizeof(so_mask));
so_mask.sin_len = 8;
so_mask.sin_addr.s_addr = 0xffffffff;
errno = 0;
/*
* XXX RTM_DELETE relies on a previous RTM_GET to fill the buffer
* appropriately.
*/
if (cmd == RTM_DELETE)
goto doit;
bzero((char *)&m_rtmsg, sizeof(m_rtmsg));
rtm->rtm_flags = flags;
rtm->rtm_version = RTM_VERSION;
switch (cmd) {
default:
errx(1, "internal wrong cmd");
case RTM_ADD:
rtm->rtm_addrs |= RTA_GATEWAY;
rtm->rtm_rmx.rmx_expire = expire_time;
rtm->rtm_inits = RTV_EXPIRE;
rtm->rtm_flags |= (RTF_HOST | RTF_STATIC | RTF_LLDATA);
dst->sin_other = 0;
if (doing_proxy) {
if (proxy_only)
dst->sin_other = SIN_PROXY;
else {
rtm->rtm_addrs |= RTA_NETMASK;
rtm->rtm_flags &= ~RTF_HOST;
}
}
/* FALLTHROUGH */
case RTM_GET:
rtm->rtm_addrs |= RTA_DST;
}
#define NEXTADDR(w, s) \
if ((s) != NULL && rtm->rtm_addrs & (w)) { \
bcopy((s), cp, sizeof(*(s))); cp += SA_SIZE(s);}
NEXTADDR(RTA_DST, dst);
NEXTADDR(RTA_GATEWAY, sdl);
NEXTADDR(RTA_NETMASK, som);
rtm->rtm_msglen = cp - (char *)&m_rtmsg;
doit:
l = rtm->rtm_msglen;
rtm->rtm_seq = ++seq;
rtm->rtm_type = cmd;
if ((rlen = write(s, (char *)&m_rtmsg, l)) < 0) {
if (errno != ESRCH || cmd != RTM_DELETE) {
warn("writing to routing socket");
return (NULL);
}
}
do {
l = read(s, (char *)&m_rtmsg, sizeof(m_rtmsg));
} while (l > 0 && (rtm->rtm_seq != seq || rtm->rtm_pid != pid));
if (l < 0)
warn("read from routing socket");
return (rtm);
}
/*
* get_ether_addr - get the hardware address of an interface on the
* the same subnet as ipaddr.
*/
#define MAX_IFS 32
static int
get_ether_addr(in_addr_t ipaddr, struct ether_addr *hwaddr)
{
struct ifreq *ifr, *ifend, *ifp;
in_addr_t ina, mask;
struct sockaddr_dl *dla;
struct ifreq ifreq;
struct ifconf ifc;
struct ifreq ifs[MAX_IFS];
int sock;
int retval = 0;
sock = socket(AF_INET, SOCK_DGRAM, 0);
if (sock < 0)
err(1, "socket");
ifc.ifc_len = sizeof(ifs);
ifc.ifc_req = ifs;
if (ioctl(sock, SIOCGIFCONF, &ifc) < 0) {
warnx("ioctl(SIOCGIFCONF)");
goto done;
}
#define NEXTIFR(i) \
((struct ifreq *)((char *)&(i)->ifr_addr \
+ MAX((i)->ifr_addr.sa_len, sizeof((i)->ifr_addr))) )
/*
* Scan through looking for an interface with an Internet
* address on the same subnet as `ipaddr'.
*/
ifend = (struct ifreq *)(ifc.ifc_buf + ifc.ifc_len);
for (ifr = ifc.ifc_req; ifr < ifend; ifr = NEXTIFR(ifr) ) {
if (ifr->ifr_addr.sa_family != AF_INET)
continue;
strncpy(ifreq.ifr_name, ifr->ifr_name,
sizeof(ifreq.ifr_name));
ifreq.ifr_addr = ifr->ifr_addr;
/*
* Check that the interface is up,
* and not point-to-point or loopback.
*/
if (ioctl(sock, SIOCGIFFLAGS, &ifreq) < 0)
continue;
if ((ifreq.ifr_flags &
(IFF_UP|IFF_BROADCAST|IFF_POINTOPOINT|
IFF_LOOPBACK|IFF_NOARP))
!= (IFF_UP|IFF_BROADCAST))
continue;
/*
* Get its netmask and check that it's on
* the right subnet.
*/
if (ioctl(sock, SIOCGIFNETMASK, &ifreq) < 0)
continue;
mask = ((struct sockaddr_in *)
&ifreq.ifr_addr)->sin_addr.s_addr;
ina = ((struct sockaddr_in *)
&ifr->ifr_addr)->sin_addr.s_addr;
if ((ipaddr & mask) == (ina & mask))
break; /* ok, we got it! */
}
if (ifr >= ifend)
goto done;
/*
* Now scan through again looking for a link-level address
* for this interface.
*/
ifp = ifr;
for (ifr = ifc.ifc_req; ifr < ifend; ifr = NEXTIFR(ifr))
if (strcmp(ifp->ifr_name, ifr->ifr_name) == 0 &&
ifr->ifr_addr.sa_family == AF_LINK)
break;
if (ifr >= ifend)
goto done;
/*
* Found the link-level address - copy it out
*/
dla = (struct sockaddr_dl *) &ifr->ifr_addr;
memcpy(hwaddr, LLADDR(dla), dla->sdl_alen);
printf("using interface %s for proxy with address ",
ifp->ifr_name);
printf("%s\n", ether_ntoa(hwaddr));
retval = dla->sdl_alen;
done:
close(sock);
return (retval);
}