分类： WINDOWS
2009-06-15 09:14:15
========================================================================
ModSecurity (Core Rules) HPP Filter Bypass Vulnerability
========================================================================
Affected Software : ModSecurity <= 2.5.9 using ModSecurity Core Rules <= 2.5-1.6.1
Author : Lavakumar Kuppan - lavakumar[dot]in[at]gmail[dot]com
Advisory URL :
Severity : High
Local/Remote : Remote
[Vulnerability Details]
Modsecurity is an Open source Web Application firewall which runs as an Apache
module. It has a comprehensive set of rules called 'ModSecurity Core Rules' for common web application
attacks like SQL Injection, Cross-Site Scripting etc.
It is possible to bypass the ModSecurity Core Rules due to the difference in behaviour
of ModSecurity and ASP/ASP.NET applications in handling duplicate HTTP GET/POST/Cookie
parameters. Using duplicate parameters has been termed as HTTP Parameter Pollution by Luca Carettoni
and Stefano Di Paola.
When multiple GET/POST/Cookie parameters of the same name are passed in the HTTP request
to ASP and ASP.NET applications they are treated as an array collection.
This leads to the values being concatenated with a comma inbetween them.
For example when the following query is sent to the server:
-----------------------------
POST /index.aspx?a=1&a=2
Host:
Cookie: a=5; a=6
Content-Length: 7
a=3&a=4
-----------------------------
The server side interpretation of this data is as follows:
Request.Params["a"] --> "1,2,3,4,5,6" ( if "a" was registered as a server-side control ) (ASP.NET Only)
Request.Params["a"] --> "1,2,5,6" ( if "a" was not registered as a server-side control ) (ASP.NET Only)
Request.QueryString["a"] --> "1,2" (ASP and ASP.NET)
Request.Form["a"] --> "3,4" (ASP and ASP.NET)
This behaviour is unique to ASP and ASP.NET applications and ModSecurity does not interpret this data in the
same way. When dealt with multiple parameters of the same name ModSecurity matches the value of each instance
of the parameter seperately against its rule base. Incase of the above example ModSecurity would run '1' against
the rule set first then '2' and so on till '6'.
Since data is interpreted differently by the Web Application and the Firewall this produces intresting possibilities
for a filter bypass scenario.
This theory was tested against the SQL Injection rule base of ModSecurity Core Rules and was found to bypass the
default-enabled rule set successfully.
The following request is blocked by ModSecurity as this matches its Generic SQL Injection Attack rule.
1,2,3 from table
ModSecurity Interpretation:
value = select 1,2,3 from table
Web Application Interpretation:
value = select 1,2,3 from table
However the same payload can be sent to the server by splitting it using duplicate parameters like below.
1&value=2,3 from table
ModSecurity Interpretation:
value = select 1
value = 2,3 from table
Web Application Interpretation:
value select 1,2,3 from table
The attack can be made more flexible by using the inline comment feature in MS SQL servers.
/*&value=*/1,2,3/*&value=*/from/*&value=*/table
ModSecurity Interpretation:
value=select/*
value=*/1,2,3/*
value=*/from/*
value=*/table
Web Application Interpretation:
value = select/*,*/1,2,3/*,*/from/*,*/table
This technique could possibly be extended to exploit other types of Web Application vulnerabilities as well.
Refer the whitepaper 'Split and Join' (see references) for more details on this attack.
[Fix Information]
N/A
[References]
/Split_and_Join.pdf
http://www.owasp.org/images/b/ba/AppsecEU09_CarettoniDiPaola_v0.8.pdf
[Legal Notices]
The information in the advisory is believed to be accurate at the
time of publishing based on currently available information.
This information is provided as-is, as a free service to the community.
There are no warranties with regard to this information.
The author does not accept any liability for any direct,
indirect, or consequential loss or damage arising from use of,
or reliance on, this information.
Permission is hereby granted for the redistribution of this alert,
provided that the content is not altered in any way, except
reformatting, and that due credit is given.
This vulnerability has been disclosed in accordance with the RFP
Full-Disclosure Policy v2.0, available at:
