Bugtraq mailing list archives

Resources consumption in Reptile webserver daily version


From: "Donato Ferrante" <fdonato () autistici org>
Date: Sat, 24 Jan 2004 18:41:40 +0100

                           Donato Ferrante


Application:  Reptile Web Server
              http://sourceforge.net/projects/reptilews

Version:      daily version

Bug:          resources consumption

Author:       Donato Ferrante
              e-mail: fdonato () autistici org
              web:    www.autistici.org/fdonato



xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

1. Description
2. The bug
3. The code
4. The fix



xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

----------------
1. Description:
----------------

Vendor's Description:

"Reptile is a web server made in Python. It supports server side
scripting with "Embedded Python", PHP, and CGI scripts. It has an
integrated HTML/XML validator that checks the pages before publication
and others handy features."



xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

------------
2. The bug:
------------

The program doesn't well manage the user input string.
In fact it waits the HTTP version. So an attacker can consume a lot of
CPU resources, sending crafted strings.



xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

-------------
3. The code:
-------------

To test the vulnerability simply send to the webserver some (about 10)
strings like:

GET index.htm

without specify the HTTP* at the end of the GET request, and where
the requested file must be avaible in the public_html directory.



xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

------------
4. The fix:
------------

No fix.
Reptile Web Server is no more supported.



xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx


Current thread: