1
0
mirror of https://git.FreeBSD.org/src.git synced 2024-12-21 11:13:30 +00:00
freebsd/lib/libc/net/rcmd.c
Daniel Eischen d201fe46e3 Remove _THREAD_SAFE and make libc thread-safe by default by
adding (weak definitions to) stubs for some of the pthread
functions.  If the threads library is linked in, the real
pthread functions will pulled in.

Use the following convention for system calls wrapped by the
threads library:
	__sys_foo - actual system call
	_foo - weak definition to __sys_foo
	foo - weak definition to __sys_foo

Change all libc uses of system calls wrapped by the threads
library from foo to _foo.  In order to define the prototypes
for _foo(), we introduce namespace.h and un-namespace.h
(suggested by bde).  All files that need to reference these
system calls, should include namespace.h before any standard
includes, then include un-namespace.h after the standard
includes and before any local includes.  <db.h> is an exception
and shouldn't be included in between namespace.h and
un-namespace.h  namespace.h will define foo to _foo, and
un-namespace.h will undefine foo.

Try to eliminate some of the recursive calls to MT-safe
functions in libc/stdio in preparation for adding a mutex
to FILE.  We have recursive mutexes, but would like to avoid
using them if possible.

Remove uneeded includes of <errno.h> from a few files.

Add $FreeBSD$ to a few files in order to pass commitprep.

Approved by:	-arch
2001-01-24 13:01:12 +00:00

766 lines
19 KiB
C

/*
* Copyright (c) 1983, 1993, 1994
* 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 University of
* California, Berkeley and its contributors.
* 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.
*
* $FreeBSD$
*/
#if defined(LIBC_SCCS) && !defined(lint)
static char sccsid[] = "@(#)rcmd.c 8.3 (Berkeley) 3/26/94";
#endif /* LIBC_SCCS and not lint */
#include "namespace.h"
#include <sys/param.h>
#include <sys/socket.h>
#include <sys/stat.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include <signal.h>
#include <fcntl.h>
#include <netdb.h>
#include <unistd.h>
#include <pwd.h>
#include <errno.h>
#include <stdio.h>
#include <ctype.h>
#include <string.h>
#ifdef YP
#include <rpc/rpc.h>
#include <rpcsvc/yp_prot.h>
#include <rpcsvc/ypclnt.h>
#endif
#include <arpa/nameser.h>
#include "un-namespace.h"
/* wrapper for KAME-special getnameinfo() */
#ifndef NI_WITHSCOPEID
#define NI_WITHSCOPEID 0
#endif
extern int innetgr __P(( const char *, const char *, const char *, const char * ));
#define max(a, b) ((a > b) ? a : b)
int __ivaliduser __P((FILE *, u_int32_t, const char *, const char *));
int __ivaliduser_af __P((FILE *,const void *, const char *, const char *,
int, int));
int __ivaliduser_sa __P((FILE *, const struct sockaddr *, socklen_t,
const char *,const char *));
static int __icheckhost __P((const struct sockaddr *, socklen_t,
const char *));
char paddr[NI_MAXHOST];
int
rcmd(ahost, rport, locuser, remuser, cmd, fd2p)
char **ahost;
u_short rport;
const char *locuser, *remuser, *cmd;
int *fd2p;
{
return rcmd_af(ahost, rport, locuser, remuser, cmd, fd2p, AF_INET);
}
int
rcmd_af(ahost, rport, locuser, remuser, cmd, fd2p, af)
char **ahost;
u_short rport;
const char *locuser, *remuser, *cmd;
int *fd2p;
int af;
{
struct addrinfo hints, *res, *ai;
struct sockaddr_storage from;
fd_set reads;
sigset_t oldmask, newmask;
pid_t pid;
int s, aport, lport, timo, error;
char c;
int refused, nres;
char num[8];
static char canonnamebuf[MAXDNAME]; /* is it proper here? */
pid = getpid();
memset(&hints, 0, sizeof(hints));
hints.ai_flags = AI_CANONNAME;
hints.ai_family = af;
hints.ai_socktype = SOCK_STREAM;
hints.ai_protocol = 0;
(void)snprintf(num, sizeof(num), "%d", ntohs(rport));
error = getaddrinfo(*ahost, num, &hints, &res);
if (error) {
fprintf(stderr, "rcmd: getaddrinfo: %s\n",
gai_strerror(error));
if (error == EAI_SYSTEM)
fprintf(stderr, "rcmd: getaddrinfo: %s\n",
strerror(errno));
return (-1);
}
if (res->ai_canonname
&& strlen(res->ai_canonname) + 1 < sizeof(canonnamebuf)) {
strncpy(canonnamebuf, res->ai_canonname, sizeof(canonnamebuf));
*ahost = canonnamebuf;
}
nres = 0;
for (ai = res; ai; ai = ai->ai_next)
nres++;
ai = res;
refused = 0;
sigemptyset(&newmask);
sigaddset(&newmask, SIGURG);
_sigprocmask(SIG_BLOCK, (const sigset_t *)&newmask, &oldmask);
for (timo = 1, lport = IPPORT_RESERVED - 1;;) {
s = rresvport_af(&lport, ai->ai_family);
if (s < 0) {
if (errno != EAGAIN && ai->ai_next) {
ai = ai->ai_next;
continue;
}
if (errno == EAGAIN)
(void)fprintf(stderr,
"rcmd: socket: All ports in use\n");
else
(void)fprintf(stderr, "rcmd: socket: %s\n",
strerror(errno));
freeaddrinfo(res);
_sigprocmask(SIG_SETMASK, (const sigset_t *)&oldmask,
NULL);
return (-1);
}
_fcntl(s, F_SETOWN, pid);
if (_connect(s, ai->ai_addr, ai->ai_addrlen) >= 0)
break;
(void)_close(s);
if (errno == EADDRINUSE) {
lport--;
continue;
}
if (errno == ECONNREFUSED)
refused = 1;
if (ai->ai_next == NULL && (!refused || timo > 16)) {
(void)fprintf(stderr, "%s: %s\n",
*ahost, strerror(errno));
freeaddrinfo(res);
_sigprocmask(SIG_SETMASK, (const sigset_t *)&oldmask,
NULL);
return (-1);
}
if (nres > 1) {
int oerrno = errno;
getnameinfo(ai->ai_addr, ai->ai_addrlen,
paddr, sizeof(paddr),
NULL, 0,
NI_NUMERICHOST|NI_WITHSCOPEID);
(void)fprintf(stderr, "connect to address %s: ",
paddr);
errno = oerrno;
perror(0);
}
if ((ai = ai->ai_next) == NULL) {
/* refused && timo <= 16 */
struct timespec time_to_sleep, time_remaining;
time_to_sleep.tv_sec = timo;
time_to_sleep.tv_nsec = 0;
(void)_nanosleep(&time_to_sleep, &time_remaining);
timo *= 2;
ai = res;
refused = 0;
}
if (nres > 1) {
getnameinfo(ai->ai_addr, ai->ai_addrlen,
paddr, sizeof(paddr),
NULL, 0,
NI_NUMERICHOST|NI_WITHSCOPEID);
fprintf(stderr, "Trying %s...\n", paddr);
}
}
lport--;
if (fd2p == 0) {
_write(s, "", 1);
lport = 0;
} else {
char num[8];
int s2 = rresvport_af(&lport, ai->ai_family), s3;
int len = ai->ai_addrlen;
int nfds;
if (s2 < 0)
goto bad;
_listen(s2, 1);
(void)snprintf(num, sizeof(num), "%d", lport);
if (_write(s, num, strlen(num)+1) != strlen(num)+1) {
(void)fprintf(stderr,
"rcmd: write (setting up stderr): %s\n",
strerror(errno));
(void)_close(s2);
goto bad;
}
nfds = max(s, s2)+1;
if(nfds > FD_SETSIZE) {
fprintf(stderr, "rcmd: too many files\n");
(void)_close(s2);
goto bad;
}
again:
FD_ZERO(&reads);
FD_SET(s, &reads);
FD_SET(s2, &reads);
errno = 0;
if (_select(nfds, &reads, 0, 0, 0) < 1 || !FD_ISSET(s2, &reads)){
if (errno != 0)
(void)fprintf(stderr,
"rcmd: select (setting up stderr): %s\n",
strerror(errno));
else
(void)fprintf(stderr,
"select: protocol failure in circuit setup\n");
(void)_close(s2);
goto bad;
}
s3 = _accept(s2, (struct sockaddr *)&from, &len);
switch (from.ss_family) {
case AF_INET:
aport = ntohs(((struct sockaddr_in *)&from)->sin_port);
break;
#ifdef INET6
case AF_INET6:
aport = ntohs(((struct sockaddr_in6 *)&from)->sin6_port);
break;
#endif
default:
aport = 0; /* error */
break;
}
/*
* XXX careful for ftp bounce attacks. If discovered, shut them
* down and check for the real auxiliary channel to connect.
*/
if (aport == 20) {
_close(s3);
goto again;
}
(void)_close(s2);
if (s3 < 0) {
(void)fprintf(stderr,
"rcmd: accept: %s\n", strerror(errno));
lport = 0;
goto bad;
}
*fd2p = s3;
if (aport >= IPPORT_RESERVED || aport < IPPORT_RESERVED / 2) {
(void)fprintf(stderr,
"socket: protocol failure in circuit setup.\n");
goto bad2;
}
}
(void)_write(s, locuser, strlen(locuser)+1);
(void)_write(s, remuser, strlen(remuser)+1);
(void)_write(s, cmd, strlen(cmd)+1);
if (_read(s, &c, 1) != 1) {
(void)fprintf(stderr,
"rcmd: %s: %s\n", *ahost, strerror(errno));
goto bad2;
}
if (c != 0) {
while (_read(s, &c, 1) == 1) {
(void)_write(STDERR_FILENO, &c, 1);
if (c == '\n')
break;
}
goto bad2;
}
_sigprocmask(SIG_SETMASK, (const sigset_t *)&oldmask, NULL);
freeaddrinfo(res);
return (s);
bad2:
if (lport)
(void)_close(*fd2p);
bad:
(void)_close(s);
_sigprocmask(SIG_SETMASK, (const sigset_t *)&oldmask, NULL);
freeaddrinfo(res);
return (-1);
}
int
rresvport(port)
int *port;
{
return rresvport_af(port, AF_INET);
}
int
rresvport_af(alport, family)
int *alport, family;
{
int i, s, len, err;
struct sockaddr_storage ss;
u_short *sport;
memset(&ss, 0, sizeof(ss));
ss.ss_family = family;
switch (family) {
case AF_INET:
((struct sockaddr *)&ss)->sa_len = sizeof(struct sockaddr_in);
sport = &((struct sockaddr_in *)&ss)->sin_port;
((struct sockaddr_in *)&ss)->sin_addr.s_addr = INADDR_ANY;
break;
#ifdef INET6
case AF_INET6:
((struct sockaddr *)&ss)->sa_len = sizeof(struct sockaddr_in6);
sport = &((struct sockaddr_in6 *)&ss)->sin6_port;
((struct sockaddr_in6 *)&ss)->sin6_addr = in6addr_any;
break;
#endif
default:
errno = EAFNOSUPPORT;
return -1;
}
s = _socket(ss.ss_family, SOCK_STREAM, 0);
if (s < 0)
return (-1);
#if 0 /* compat_exact_traditional_rresvport_semantics */
sin.sin_port = htons((u_short)*alport);
if (_bind(s, (struct sockaddr *)&sin, sizeof(sin)) >= 0)
return (s);
if (errno != EADDRINUSE) {
(void)_close(s);
return (-1);
}
#endif
*sport = 0;
if (bindresvport_sa(s, (struct sockaddr *)&ss) == -1) {
(void)_close(s);
return (-1);
}
*alport = (int)ntohs(*sport);
return (s);
}
int __check_rhosts_file = 1;
char *__rcmd_errstr;
int
ruserok(rhost, superuser, ruser, luser)
const char *rhost, *ruser, *luser;
int superuser;
{
struct addrinfo hints, *res, *r;
int error;
memset(&hints, 0, sizeof(hints));
hints.ai_family = PF_UNSPEC;
hints.ai_socktype = SOCK_DGRAM; /*dummy*/
error = getaddrinfo(rhost, "0", &hints, &res);
if (error)
return (-1);
for (r = res; r; r = r->ai_next) {
if (iruserok_sa(r->ai_addr, r->ai_addrlen, superuser, ruser,
luser) == 0) {
freeaddrinfo(res);
return (0);
}
}
freeaddrinfo(res);
return (-1);
}
/*
* New .rhosts strategy: We are passed an ip address. We spin through
* hosts.equiv and .rhosts looking for a match. When the .rhosts only
* has ip addresses, we don't have to trust a nameserver. When it
* contains hostnames, we spin through the list of addresses the nameserver
* gives us and look for a match.
*
* Returns 0 if ok, -1 if not ok.
*/
int
iruserok(raddr, superuser, ruser, luser)
unsigned long raddr;
int superuser;
const char *ruser, *luser;
{
struct sockaddr_in sin;
memset(&sin, 0, sizeof(sin));
sin.sin_family = AF_INET;
sin.sin_len = sizeof(struct sockaddr_in);
memcpy(&sin.sin_addr, &raddr, sizeof(sin.sin_addr));
return iruserok_sa((struct sockaddr *)&sin, sin.sin_len, superuser,
ruser, luser);
}
/*
* AF independent extension of iruserok.
*
* Returns 0 if ok, -1 if not ok.
*/
int
iruserok_sa(ra, rlen, superuser, ruser, luser)
const void *ra;
int rlen;
int superuser;
const char *ruser, *luser;
{
register char *cp;
struct stat sbuf;
struct passwd *pwd;
FILE *hostf;
uid_t uid;
int first;
char pbuf[MAXPATHLEN];
const struct sockaddr *raddr;
struct sockaddr_storage ss;
/* avoid alignment issue */
if (rlen > sizeof(ss))
return(-1);
memcpy(&ss, ra, rlen);
raddr = (struct sockaddr *)&ss;
first = 1;
hostf = superuser ? NULL : fopen(_PATH_HEQUIV, "r");
again:
if (hostf) {
if (__ivaliduser_sa(hostf, raddr, rlen, luser, ruser) == 0) {
(void)fclose(hostf);
return (0);
}
(void)fclose(hostf);
}
if (first == 1 && (__check_rhosts_file || superuser)) {
first = 0;
if ((pwd = getpwnam(luser)) == NULL)
return (-1);
(void)strcpy(pbuf, pwd->pw_dir);
(void)strcat(pbuf, "/.rhosts");
/*
* Change effective uid while opening .rhosts. If root and
* reading an NFS mounted file system, can't read files that
* are protected read/write owner only.
*/
uid = geteuid();
(void)seteuid(pwd->pw_uid);
hostf = fopen(pbuf, "r");
(void)seteuid(uid);
if (hostf == NULL)
return (-1);
/*
* If not a regular file, or is owned by someone other than
* user or root or if writeable by anyone but the owner, quit.
*/
cp = NULL;
if (lstat(pbuf, &sbuf) < 0)
cp = ".rhosts lstat failed";
else if (!S_ISREG(sbuf.st_mode))
cp = ".rhosts not regular file";
else if (_fstat(fileno(hostf), &sbuf) < 0)
cp = ".rhosts fstat failed";
else if (sbuf.st_uid && sbuf.st_uid != pwd->pw_uid)
cp = "bad .rhosts owner";
else if (sbuf.st_mode & (S_IWGRP|S_IWOTH))
cp = ".rhosts writeable by other than owner";
/* If there were any problems, quit. */
if (cp) {
__rcmd_errstr = cp;
(void)fclose(hostf);
return (-1);
}
goto again;
}
return (-1);
}
/*
* XXX
* Don't make static, used by lpd(8).
*
* Returns 0 if ok, -1 if not ok.
*/
int
__ivaliduser(hostf, raddr, luser, ruser)
FILE *hostf;
u_int32_t raddr;
const char *luser, *ruser;
{
struct sockaddr_in sin;
memset(&sin, 0, sizeof(sin));
sin.sin_family = AF_INET;
sin.sin_len = sizeof(struct sockaddr_in);
memcpy(&sin.sin_addr, &raddr, sizeof(sin.sin_addr));
return __ivaliduser_sa(hostf, (struct sockaddr *)&sin, sin.sin_len,
luser, ruser);
}
/*
* Returns 0 if ok, -1 if not ok.
*
* XXX obsolete API.
*/
int
__ivaliduser_af(hostf, raddr, luser, ruser, af, len)
FILE *hostf;
const void *raddr;
const char *luser, *ruser;
int af, len;
{
struct sockaddr *sa = NULL;
struct sockaddr_in *sin = NULL;
#ifdef INET6
struct sockaddr_in6 *sin6 = NULL;
#endif
struct sockaddr_storage ss;
memset(&ss, 0, sizeof(ss));
switch (af) {
case AF_INET:
if (len != sizeof(sin->sin_addr))
return -1;
sin = (struct sockaddr_in *)&ss;
sin->sin_family = AF_INET;
sin->sin_len = sizeof(struct sockaddr_in);
memcpy(&sin->sin_addr, raddr, sizeof(sin->sin_addr));
break;
#ifdef INET6
case AF_INET6:
if (len != sizeof(sin6->sin6_addr))
return -1;
/* you will lose scope info */
sin6 = (struct sockaddr_in6 *)&ss;
sin6->sin6_family = AF_INET6;
sin6->sin6_len = sizeof(struct sockaddr_in6);
memcpy(&sin6->sin6_addr, raddr, sizeof(sin6->sin6_addr));
break;
#endif
default:
return -1;
}
sa = (struct sockaddr *)&ss;
return __ivaliduser_sa(hostf, sa, sa->sa_len, luser, ruser);
}
/*
* Returns 0 if ok, -1 if not ok.
*/
int
__ivaliduser_sa(hostf, raddr, salen, luser, ruser)
FILE *hostf;
const struct sockaddr *raddr;
socklen_t salen;
const char *luser, *ruser;
{
register char *user, *p;
int ch;
char buf[MAXHOSTNAMELEN + 128]; /* host + login */
char hname[MAXHOSTNAMELEN];
/* Presumed guilty until proven innocent. */
int userok = 0, hostok = 0;
int h_error;
#ifdef YP
char *ypdomain;
if (yp_get_default_domain(&ypdomain))
ypdomain = NULL;
#else
#define ypdomain NULL
#endif
/* We need to get the damn hostname back for netgroup matching. */
if (getnameinfo(raddr, salen, hname, sizeof(hname), NULL, 0,
NI_NAMEREQD) != 0)
return (-1);
while (fgets(buf, sizeof(buf), hostf)) {
p = buf;
/* Skip lines that are too long. */
if (strchr(p, '\n') == NULL) {
while ((ch = getc(hostf)) != '\n' && ch != EOF);
continue;
}
if (*p == '\n' || *p == '#') {
/* comment... */
continue;
}
while (*p != '\n' && *p != ' ' && *p != '\t' && *p != '\0') {
*p = isupper((unsigned char)*p) ? tolower((unsigned char)*p) : *p;
p++;
}
if (*p == ' ' || *p == '\t') {
*p++ = '\0';
while (*p == ' ' || *p == '\t')
p++;
user = p;
while (*p != '\n' && *p != ' ' &&
*p != '\t' && *p != '\0')
p++;
} else
user = p;
*p = '\0';
/*
* Do +/- and +@/-@ checking. This looks really nasty,
* but it matches SunOS's behavior so far as I can tell.
*/
switch(buf[0]) {
case '+':
if (!buf[1]) { /* '+' matches all hosts */
hostok = 1;
break;
}
if (buf[1] == '@') /* match a host by netgroup */
hostok = innetgr((char *)&buf[2],
(char *)&hname, NULL, ypdomain);
else /* match a host by addr */
hostok = __icheckhost(raddr, salen,
(char *)&buf[1]);
break;
case '-': /* reject '-' hosts and all their users */
if (buf[1] == '@') {
if (innetgr((char *)&buf[2],
(char *)&hname, NULL, ypdomain))
return(-1);
} else {
if (__icheckhost(raddr, salen,
(char *)&buf[1]))
return(-1);
}
break;
default: /* if no '+' or '-', do a simple match */
hostok = __icheckhost(raddr, salen, buf);
break;
}
switch(*user) {
case '+':
if (!*(user+1)) { /* '+' matches all users */
userok = 1;
break;
}
if (*(user+1) == '@') /* match a user by netgroup */
userok = innetgr(user+2, NULL, ruser, ypdomain);
else /* match a user by direct specification */
userok = !(strcmp(ruser, user+1));
break;
case '-': /* if we matched a hostname, */
if (hostok) { /* check for user field rejections */
if (!*(user+1))
return(-1);
if (*(user+1) == '@') {
if (innetgr(user+2, NULL,
ruser, ypdomain))
return(-1);
} else {
if (!strcmp(ruser, user+1))
return(-1);
}
}
break;
default: /* no rejections: try to match the user */
if (hostok)
userok = !(strcmp(ruser,*user ? user : luser));
break;
}
if (hostok && userok)
return(0);
}
return (-1);
}
/*
* Returns "true" if match, 0 if no match.
*
* NI_WITHSCOPEID is useful for comparing sin6_scope_id portion
* if af == AF_INET6.
*/
static int
__icheckhost(raddr, salen, lhost)
const struct sockaddr *raddr;
socklen_t salen;
const char *lhost;
{
struct sockaddr_in sin;
struct sockaddr_in6 *sin6;
struct addrinfo hints, *res, *r;
int error;
char h1[NI_MAXHOST], h2[NI_MAXHOST];
if (raddr->sa_family == AF_INET6) {
sin6 = (struct sockaddr_in6 *)raddr;
if (IN6_IS_ADDR_V4MAPPED(&sin6->sin6_addr)) {
memset(&sin, 0, sizeof(sin));
sin.sin_family = AF_INET;
sin.sin_len = sizeof(struct sockaddr_in);
memcpy(&sin.sin_addr, &sin6->sin6_addr.s6_addr[12],
sizeof(sin.sin_addr));
raddr = (struct sockaddr *)&sin;
salen = sin.sin_len;
}
}
h1[0] = '\0';
if (getnameinfo(raddr, salen, h1, sizeof(h1), NULL, 0,
NI_NUMERICHOST | NI_WITHSCOPEID) != 0)
return (0);
/* Resolve laddr into sockaddr */
memset(&hints, 0, sizeof(hints));
hints.ai_family = raddr->sa_family;
hints.ai_socktype = SOCK_DGRAM; /*XXX dummy*/
res = NULL;
error = getaddrinfo(lhost, "0", &hints, &res);
if (error)
return (0);
for (r = res; r ; r = r->ai_next) {
h2[0] = '\0';
if (getnameinfo(r->ai_addr, r->ai_addrlen, h2, sizeof(h2),
NULL, 0, NI_NUMERICHOST | NI_WITHSCOPEID) != 0)
continue;
if (strcmp(h1, h2) == 0) {
freeaddrinfo(res);
return (1);
}
}
/* No match. */
freeaddrinfo(res);
return (0);
}