new version from Vernon Schryver (vjs@sgi.com); see ../timed/CHANGES for details
[unix-history] / usr / src / usr.sbin / timed / timed / timed.c
* Copyright (c) 1985 Regents of the University of California.
* All rights reserved.
*
* %sccs.include.redist.c%
*/
#ifndef lint
char copyright[] =
"@(#) Copyright (c) 1985 Regents of the University of California.\n\
All rights reserved.\n";
#endif /* not lint */
#ifndef lint
static char sccsid[] = "@(#)timed.c 5.1 (Berkeley) %G%";
#endif /* not lint */
#ifdef sgi
#ident "$Revision: 1.25 $"
#endif /* sgi */
#define TSPTYPES
#include "globals.h"
#include <net/if.h>
#include <sys/file.h>
#include <sys/ioctl.h>
#include <setjmp.h>
#include "pathnames.h"
#include <math.h>
#include <sys/types.h>
#include <sys/times.h>
#ifdef sgi
#include <unistd.h>
#include <sys/syssgi.h>
#include <sys/schedctl.h>
#endif /* sgi */
int trace = 0;
int sock, sock_raw = -1;
int status = 0;
u_short sequence; /* sequence number */
long delay1;
long delay2;
int nslavenets; /* nets were I could be a slave */
int nmasternets; /* nets were I could be a master */
int nignorednets; /* ignored nets */
int nnets; /* nets I am connected to */
FILE *fd; /* trace file FD */
jmp_buf jmpenv;
struct netinfo *nettab = 0;
struct netinfo *slavenet;
int Mflag;
int justquit = 0;
int debug;
static struct nets {
char *name;
long net;
struct nets *next;
} *nets = 0;
struct hosttbl hosttbl[NHOSTS+1]; /* known hosts */
static struct goodhost { /* hosts that we trust */
char name[MAXHOSTNAMELEN+1];
struct goodhost *next;
char perm;
} *goodhosts;
static char *goodgroup; /* net group of trusted hosts */
static void checkignorednets __P((void));
static void pickslavenet __P((struct netinfo *));
static void add_good_host __P((char *, int));
#ifdef sgi
char *timetrim_fn;
char *timetrim_wpat = "long timetrim = %ld;\ndouble tot_adj = %.0f;\ndouble tot_ticks = %.0f;\n/* timed version 2 */\n";
char *timetrim_rpat = "long timetrim = %ld;\ndouble tot_adj = %lf;\ndouble tot_ticks = %lf;";
long timetrim;
double tot_adj, hr_adj; /* totals in nsec */
double tot_ticks, hr_ticks;
int bufspace = 60*1024;
#endif
/*
* The timedaemons synchronize the clocks of hosts in a local area network.
* One daemon runs as master, all the others as slaves. The master
* performs the task of computing clock differences and sends correction
* values to the slaves.
* Slaves start an election to choose a new master when the latter disappears
* because of a machine crash, network partition, or when killed.
* A resolution protocol is used to kill all but one of the masters
* that happen to exist in segments of a partitioned network when the
* network partition is fixed.
*
* Authors: Riccardo Gusella & Stefano Zatti
*
* overhauled at Silicon Graphics
*/
int
main(argc, argv)
int argc;
char *argv[];
{
int on;
int ret;
int nflag, iflag;
struct timeval ntime;
struct servent *srvp;
char buf[BUFSIZ], *cp, *cplim;
struct ifconf ifc;
struct ifreq ifreq, ifreqf, *ifr;
register struct netinfo *ntp;
struct netinfo *ntip;
struct netinfo *savefromnet;
struct netent *nentp;
struct nets *nt;
struct sockaddr_in server;
u_short port;
char c;
extern char *optarg;
extern int optind, opterr;
#ifdef sgi
FILE *timetrim_st;
#endif
#define IN_MSG "timed: -i and -n make no sense together\n"
#ifdef sgi
struct tms tms;
#define USAGE "timed: [-dtM] [-i net|-n net] [-F host1 host2 ...] [-G netgp] [-P trimfile]\n"
#else
#ifdef HAVENIS
#define USAGE "timed: [-dtM] [-i net|-n net] [-F host1 host2 ...] [-G netgp]\n"
#else
#define USAGE "timed: [-dtM] [-i net|-n net] [-F host1 host2 ...]\n"
#endif /* HAVENIS */
#endif /* sgi */
#ifdef lint
ntip = NULL;
#endif
on = 1;
nflag = OFF;
iflag = OFF;
#ifdef sgi
if (0 > syssgi(SGI_GETTIMETRIM, &timetrim)) {
perror("timed: syssgi(GETTIMETRIM)");
timetrim = 0;
}
tot_ticks = hr_ticks = times(&tms);
#endif /* sgi */
opterr = 0;
while ((c = getopt(argc, argv, "Mtdn:i:F:G:P:")) != EOF) {
switch (c) {
case 'M':
Mflag = 1;
break;
case 't':
trace = 1;
break;
case 'n':
if (iflag) {
fprintf(stderr, IN_MSG);
exit(1);
} else {
nflag = ON;
addnetname(optarg);
}
break;
case 'i':
if (nflag) {
fprintf(stderr, IN_MSG);
exit(1);
} else {
iflag = ON;
addnetname(optarg);
}
break;
case 'F':
add_good_host(optarg,1);
while (optind < argc && argv[optind][0] != '-')
add_good_host(argv[optind++], 1);
break;
case 'd':
debug = 1;
break;
case 'G':
if (goodgroup != 0) {
fprintf(stderr,"timed: only one net group\n");
exit(1);
}
goodgroup = optarg;
break;
#ifdef sgi
case 'P':
timetrim_fn = optarg;
timetrim_st = fopen(timetrim_fn, "r+");
if (0 == timetrim_st) {
if (errno != ENOENT) {
(void)fprintf(stderr,"timed: ");
perror(timetrim_fn);
timetrim_fn = 0;
}
} else {
int i;
long trim;
double adj, ticks;
i = fscanf(timetrim_st, timetrim_rpat,
&trim, &adj, &ticks);
if (i < 1
|| trim > MAX_TRIM
|| trim < -MAX_TRIM
|| i == 2
|| (i == 3
&& trim != rint(adj*CLK_TCK/ticks))) {
if (trace && i != EOF)
(void)fprintf(stderr,
"timed: unrecognized contents in %s\n",
timetrim_fn);
} else {
if (0 > syssgi(SGI_SETTIMETRIM,
trim)) {
perror("timed: syssgi(SETTIMETRIM)");
} else {
timetrim = trim;
}
if (i == 3) {
tot_adj = adj;
tot_ticks -= ticks;
}
}
(void)fclose(timetrim_st);
}
break;
#endif /* sgi */
default:
fprintf(stderr, USAGE);
exit(1);
break;
}
}
if (optind < argc) {
fprintf(stderr, USAGE);
exit(1);
}
if (gethostname(hostname, sizeof(hostname) - 1) < 0) {
perror("gethostname");
exit(1);
}
self.l_bak = &self;
self.l_fwd = &self;
self.h_bak = &self;
self.h_fwd = &self;
self.head = 1;
self.good = 1;
if (goodhosts != 0) /* trust ourself */
add_good_host(hostname,1);
srvp = getservbyname("timed", "udp");
if (srvp == 0) {
fprintf(stderr, "unknown service 'timed/udp'\n");
exit(1);
}
port = srvp->s_port;
server.sin_port = srvp->s_port;
server.sin_family = AF_INET;
sock = socket(AF_INET, SOCK_DGRAM, 0);
if (sock < 0) {
perror("socket");
exit(1);
}
if (setsockopt(sock, SOL_SOCKET, SO_BROADCAST, (char *)&on,
sizeof(on)) < 0) {
perror("setsockopt");
exit(1);
}
if (bind(sock, (struct sockaddr*)&server, sizeof(server))) {
if (errno == EADDRINUSE)
fprintf(stderr,"timed: time daemon already running\n");
else
perror("bind");
exit(1);
}
#ifdef sgi
/*
* handle many slaves with our buffer
*/
if (0 > setsockopt(sock, SOL_SOCKET, SO_RCVBUF, (char*)&bufspace,
sizeof(bufspace))) {
perror("setsockopt");
exit(1);
}
#endif /* sgi */
/* choose a unique seed for random number generation */
(void)gettimeofday(&ntime, 0);
srandom(ntime.tv_sec + ntime.tv_usec);
sequence = random(); /* initial seq number */
#ifndef sgi
/* rounds kernel variable time to multiple of 5 ms. */
ntime.tv_sec = 0;
ntime.tv_usec = -((ntime.tv_usec/1000) % 5) * 1000;
(void)adjtime(&ntime, (struct timeval *)0);
#endif /* sgi */
for (nt = nets; nt; nt = nt->next) {
nentp = getnetbyname(nt->name);
if (nentp == 0) {
nt->net = inet_network(nt->name);
if (nt->net != INADDR_NONE)
nentp = getnetbyaddr(nt->net, AF_INET);
}
if (nentp != 0) {
nt->net = nentp->n_net;
} else if (nt->net == INADDR_NONE) {
fprintf(stderr, "timed: unknown net %s\n", nt->name);
exit(1);
} else if (nt->net == INADDR_ANY) {
fprintf(stderr, "timed: bad net %s\n", nt->name);
exit(1);
} else {
fprintf(stderr,
"timed: warning: %s unknown in /etc/networks\n",
nt->name);
}
if (0 == (nt->net & 0xff000000))
nt->net <<= 8;
if (0 == (nt->net & 0xff000000))
nt->net <<= 8;
if (0 == (nt->net & 0xff000000))
nt->net <<= 8;
}
ifc.ifc_len = sizeof(buf);
ifc.ifc_buf = buf;
if (ioctl(sock, SIOCGIFCONF, (char *)&ifc) < 0) {
perror("timed: get interface configuration");
exit(1);
}
ntp = NULL;
#ifdef sgi
#define size(p) (sizeof(*ifr) - sizeof(ifr->ifr_name)) /* XXX hack. kludge */
#else
#define size(p) max((p).sa_len, sizeof(p))
#endif
cplim = buf + ifc.ifc_len; /*skip over if's with big ifr_addr's */
for (cp = buf; cp < cplim;
cp += sizeof (ifr->ifr_name) + size(ifr->ifr_addr)) {
ifr = (struct ifreq *)cp;
if (ifr->ifr_addr.sa_family != AF_INET)
continue;
if (!ntp)
ntp = (struct netinfo*)malloc(sizeof(struct netinfo));
bzero(ntp,sizeof(*ntp));
ntp->my_addr=((struct sockaddr_in *)&ifr->ifr_addr)->sin_addr;
ntp->status = NOMASTER;
ifreq = *ifr;
ifreqf = *ifr;
if (ioctl(sock, SIOCGIFFLAGS, (char *)&ifreqf) < 0) {
perror("get interface flags");
continue;
}
if ((ifreqf.ifr_flags & IFF_UP) == 0)
continue;
if ((ifreqf.ifr_flags & IFF_BROADCAST) == 0 &&
(ifreqf.ifr_flags & IFF_POINTOPOINT) == 0) {
continue;
}
if (ioctl(sock, SIOCGIFNETMASK, (char *)&ifreq) < 0) {
perror("get netmask");
continue;
}
ntp->mask = ((struct sockaddr_in *)
&ifreq.ifr_addr)->sin_addr.s_addr;
if (ifreqf.ifr_flags & IFF_BROADCAST) {
if (ioctl(sock, SIOCGIFBRDADDR, (char *)&ifreq) < 0) {
perror("get broadaddr");
continue;
}
ntp->dest_addr = *(struct sockaddr_in *)&ifreq.ifr_broadaddr;
/* What if the broadcast address is all ones?
* So we cannot just mask ntp->dest_addr. */
ntp->net = ntp->my_addr;
ntp->net.s_addr &= ntp->mask;
} else {
if (ioctl(sock, SIOCGIFDSTADDR,
(char *)&ifreq) < 0) {
perror("get destaddr");
continue;
}
ntp->dest_addr = *(struct sockaddr_in *)&ifreq.ifr_dstaddr;
ntp->net = ntp->dest_addr.sin_addr;
}
ntp->dest_addr.sin_port = port;
for (nt = nets; nt; nt = nt->next) {
if (ntp->net.s_addr == nt->net)
break;
}
if (nflag && !nt || iflag && nt)
continue;
ntp->next = NULL;
if (nettab == NULL) {
nettab = ntp;
} else {
ntip->next = ntp;
}
ntip = ntp;
ntp = NULL;
}
if (ntp)
(void) free((char *)ntp);
if (nettab == NULL) {
fprintf(stderr, "timed: no network usable\n");
exit(1);
}
#ifdef sgi
(void)schedctl(RENICE,0,10); /* run fast to get good time */
/* ticks to delay before responding to a broadcast */
delay1 = casual(0, CLK_TCK/10);
#else
/* microseconds to delay before responding to a broadcast */
delay1 = casual(1, 100*1000);
#endif /* sgi */
/* election timer delay in secs. */
delay2 = casual(MINTOUT, MAXTOUT);
#ifndef DEBUG
if (fork())
exit(0);
{ int s;
for (s = getdtablesize(); s >= 0; --s)
(void) close(s);
(void) open("/dev/null", 0);
(void) dup2(0, 1);
(void) dup2(0, 2);
s = open("/dev/tty", 2);
if (s >= 0) {
(void) close(s);
}
}
#endif
#ifdef sgi
(void)_daemonize(debug ? _DF_NOFORK|_DF_NOCHDIR : 0, sock, -1, -1);
#else
if (!debug)
daemon(debug, 0);
#endif /* sgi */
if (trace)
traceon();
openlog("timed", LOG_CONS|LOG_PID, LOG_DAEMON);
/*
* keep returning here
*/
ret = setjmp(jmpenv);
savefromnet = fromnet;
setstatus();
if (Mflag) {
switch (ret) {
case 0:
checkignorednets();
pickslavenet(0);
break;
case 1:
/* Just lost our master */
if (slavenet != 0)
slavenet->status = election(slavenet);
if (!slavenet || slavenet->status == MASTER) {
checkignorednets();
pickslavenet(0);
} else {
makeslave(slavenet); /* prune extras */
}
break;
case 2:
/* Just been told to quit */
justquit = 1;
pickslavenet(savefromnet);
break;
}
setstatus();
if (!(status & MASTER) && sock_raw != -1) {
/* sock_raw is not being used now */
(void)close(sock_raw);
sock_raw = -1;
}
if (status == MASTER)
master();
else
slave();
} else {
if (sock_raw != -1) {
(void)close(sock_raw);
sock_raw = -1;
}
if (ret) {
/* we just lost our master or were told to quit */
justquit = 1;
}
for (ntp = nettab; ntp != NULL; ntp = ntp->next) {
if (ntp->status == MASTER)
rmnetmachs(ntp);
ntp->status = NOMASTER;
}
checkignorednets();
pickslavenet(0);
setstatus();
slave();
}
/* NOTREACHED */
#ifdef lint
return(0);
#endif
}
/*
* suppress an upstart, untrustworthy, self-appointed master
*/
void
suppress(addr, name,net)
struct sockaddr_in *addr;
char *name;
struct netinfo *net;
{
struct sockaddr_in tgt;
char tname[MAXHOSTNAMELEN];
struct tsp msg;
static struct timeval wait;
if (trace)
fprintf(fd, "suppress: %s\n", name);
tgt = *addr;
(void)strcpy(tname, name);
while (0 != readmsg(TSP_ANY, ANYADDR, &wait, net)) {
if (trace)
fprintf(fd, "suppress:\tdiscarded packet from %s\n",
name);
}
syslog(LOG_NOTICE, "suppressing false master %s", tname);
msg.tsp_type = TSP_QUIT;
(void)strcpy(msg.tsp_name, hostname);
(void)acksend(&msg, &tgt, tname, TSP_ACK, 0, 1);
}
void
lookformaster(ntp)
struct netinfo *ntp;
{
struct tsp resp, conflict, *answer;
struct timeval ntime;
char mastername[MAXHOSTNAMELEN];
struct sockaddr_in masteraddr;
get_goodgroup(0);
ntp->status = SLAVE;
/* look for master */
resp.tsp_type = TSP_MASTERREQ;
(void)strcpy(resp.tsp_name, hostname);
answer = acksend(&resp, &ntp->dest_addr, ANYADDR,
TSP_MASTERACK, ntp, 0);
if (answer != 0 && !good_host_name(answer->tsp_name)) {
suppress(&from, answer->tsp_name, ntp);
ntp->status = NOMASTER;
answer = 0;
}
if (answer == 0) {
/*
* Various conditions can cause conflict: races between
* two just started timedaemons when no master is
* present, or timedaemons started during an election.
* A conservative approach is taken. Give up and became a
* slave, postponing election of a master until first
* timer expires.
*/
ntime.tv_sec = ntime.tv_usec = 0;
answer = readmsg(TSP_MASTERREQ, ANYADDR, &ntime, ntp);
if (answer != 0) {
if (!good_host_name(answer->tsp_name)) {
suppress(&from, answer->tsp_name, ntp);
ntp->status = NOMASTER;
}
return;
}
ntime.tv_sec = ntime.tv_usec = 0;
answer = readmsg(TSP_MASTERUP, ANYADDR, &ntime, ntp);
if (answer != 0) {
if (!good_host_name(answer->tsp_name)) {
suppress(&from, answer->tsp_name, ntp);
ntp->status = NOMASTER;
}
return;
}
ntime.tv_sec = ntime.tv_usec = 0;
answer = readmsg(TSP_ELECTION, ANYADDR, &ntime, ntp);
if (answer != 0) {
if (!good_host_name(answer->tsp_name)) {
suppress(&from, answer->tsp_name, ntp);
ntp->status = NOMASTER;
}
return;
}
if (Mflag)
ntp->status = MASTER;
else
ntp->status = NOMASTER;
return;
}
ntp->status = SLAVE;
(void)strcpy(mastername, answer->tsp_name);
masteraddr = from;
/*
* If network has been partitioned, there might be other
* masters; tell the one we have just acknowledged that
* it has to gain control over the others.
*/
ntime.tv_sec = 0;
ntime.tv_usec = 300000;
answer = readmsg(TSP_MASTERACK, ANYADDR, &ntime, ntp);
/*
* checking also not to send CONFLICT to ack'ed master
* due to duplicated MASTERACKs
*/
if (answer != NULL &&
strcmp(answer->tsp_name, mastername) != 0) {
conflict.tsp_type = TSP_CONFLICT;
(void)strcpy(conflict.tsp_name, hostname);
if (!acksend(&conflict, &masteraddr, mastername,
TSP_ACK, 0, 0)) {
syslog(LOG_ERR,
"error on sending TSP_CONFLICT");
}
}
}
/*
* based on the current network configuration, set the status, and count
* networks;
*/
void
setstatus()
{
struct netinfo *ntp;
status = 0;
nmasternets = nslavenets = nnets = nignorednets = 0;
if (trace)
fprintf(fd, "Net status:\n");
for (ntp = nettab; ntp != NULL; ntp = ntp->next) {
switch ((int)ntp->status) {
case MASTER:
nmasternets++;
break;
case SLAVE:
nslavenets++;
break;
case NOMASTER:
case IGNORE:
nignorednets++;
break;
}
if (trace) {
fprintf(fd, "\t%-16s", inet_ntoa(ntp->net));
switch ((int)ntp->status) {
case NOMASTER:
fprintf(fd, "NOMASTER\n");
break;
case MASTER:
fprintf(fd, "MASTER\n");
break;
case SLAVE:
fprintf(fd, "SLAVE\n");
break;
case IGNORE:
fprintf(fd, "IGNORE\n");
break;
default:
fprintf(fd, "invalid state %d\n",
(int)ntp->status);
break;
}
}
nnets++;
status |= ntp->status;
}
status &= ~IGNORE;
if (trace)
fprintf(fd,
"\tnets=%d masters=%d slaves=%d ignored=%d delay2=%d\n",
nnets, nmasternets, nslavenets, nignorednets, delay2);
}
void
makeslave(net)
struct netinfo *net;
{
register struct netinfo *ntp;
for (ntp = nettab; ntp != NULL; ntp = ntp->next) {
if (ntp->status == SLAVE && ntp != net)
ntp->status = IGNORE;
}
slavenet = net;
}
/*
* Try to become master over ignored nets..
*/
static void
checkignorednets()
{
register struct netinfo *ntp;
for (ntp = nettab; ntp != NULL; ntp = ntp->next) {
if (!Mflag && ntp->status == SLAVE)
break;
if (ntp->status == IGNORE || ntp->status == NOMASTER) {
lookformaster(ntp);
if (!Mflag && ntp->status == SLAVE)
break;
}
}
}
/*
* choose a good network on which to be a slave
* The ignored networks must have already been checked.
* Take a hint about for a good network.
*/
static void
pickslavenet(ntp)
struct netinfo *ntp;
{
if (slavenet != 0 && slavenet->status == SLAVE) {
makeslave(slavenet); /* prune extras */
return;
}
if (ntp == 0 || ntp->status != SLAVE) {
for (ntp = nettab; ntp != 0; ntp = ntp->next) {
if (ntp->status == SLAVE)
break;
}
}
makeslave(ntp);
}
/*
* returns a random number in the range [inf, sup]
*/
long
casual(inf, sup)
long inf, sup;
{
double value;
value = ((double)(random() & 0x7fffffff)) / (0x7fffffff*1.0);
return(inf + (sup - inf)*value);
}
char *
date()
{
#ifdef sgi
struct timeval tv;
static char tm[32];
(void)gettimeofday(&tv, (struct timezone *)0);
(void)cftime(tm, "%D %T", &tv.tv_sec);
return (tm);
#else
struct timeval tv;
(void)gettimeofday(&tv, (struct timezone *)0);
return (ctime(&tv.tv_sec));
#endif /* sgi */
}
void
addnetname(name)
char *name;
{
register struct nets **netlist = &nets;
while (*netlist)
netlist = &((*netlist)->next);
*netlist = (struct nets *)malloc(sizeof **netlist);
if (*netlist == 0) {
fprintf(stderr,"malloc failed\n");
exit(1);
}
bzero((char *)*netlist, sizeof(**netlist));
(*netlist)->name = name;
}
/* note a host as trustworthy */
static void
add_good_host(name, perm)
char *name;
int perm; /* 1=not part of the netgroup */
{
register struct goodhost *ghp;
register struct hostent *hentp;
ghp = (struct goodhost*)malloc(sizeof(*ghp));
if (!ghp) {
syslog(LOG_ERR, "malloc failed");
exit(1);
}
bzero((char*)ghp, sizeof(*ghp));
(void)strncpy(&ghp->name[0], name, sizeof(ghp->name));
ghp->next = goodhosts;
ghp->perm = perm;
goodhosts = ghp;
hentp = gethostbyname(name);
if (0 == hentp && perm)
(void)fprintf(stderr, "unknown host %s\n", name);
}
/* update our image of the net-group of trustworthy hosts
*/
void
get_goodgroup(force)
int force;
{
# define NG_DELAY (30*60*CLK_TCK) /* 30 minutes */
static unsigned long last_update = -NG_DELAY;
unsigned long new_update;
struct hosttbl *htp;
struct goodhost *ghp, **ghpp;
char *mach, *usr, *dom;
struct tms tm;
/* if no netgroup, then we are finished */
if (goodgroup == 0 || !Mflag)
return;
/* Do not chatter with the netgroup master too often.
*/
new_update = times(&tm);
if (new_update < last_update + NG_DELAY
&& !force)
return;
last_update = new_update;
/* forget the old temporary entries */
ghpp = &goodhosts;
while (0 != (ghp = *ghpp)) {
if (!ghp->perm) {
*ghpp = ghp->next;
free((char*)ghp);
} else {
ghpp = &ghp->next;
}
}
#ifdef HAVENIS
/* quit now if we are not one of the trusted masters
*/
if (!innetgr(goodgroup, &hostname[0], 0,0)) {
if (trace)
(void)fprintf(fd, "get_goodgroup: %s not in %s\n",
&hostname[0], goodgroup);
return;
}
if (trace)
(void)fprintf(fd, "get_goodgroup: %s in %s\n",
&hostname[0], goodgroup);
/* mark the entire netgroup as trusted */
(void)setnetgrent(goodgroup);
while (getnetgrent(&mach,&usr,&dom)) {
if (0 != mach)
add_good_host(mach,0);
}
(void)endnetgrent();
/* update list of slaves */
for (htp = self.l_fwd; htp != &self; htp = htp->l_fwd) {
htp->good = good_host_name(&htp->name[0]);
}
#endif /* HAVENIS */
}
/* see if a machine is trustworthy
*/
int /* 1=trust hp to change our date */
good_host_name(name)
char *name;
{
register struct goodhost *ghp = goodhosts;
register char c;
if (!ghp || !Mflag) /* trust everyone if no one named */
return 1;
c = *name;
do {
if (c == ghp->name[0]
&& !strcasecmp(name, ghp->name))
return 1; /* found him, so say so */
} while (0 != (ghp = ghp->next));
if (!strcasecmp(name,hostname)) /* trust ourself */
return 1;
return 0; /* did not find him */
}