Re: Hacker pierces hardware firewalls with web page.

I don't see any need for them to do anything like this. It seems far 
simpler to me.

As an example:

To support FTP packet filter based firewalls watch the data stream of any 
connections going to port 21 for the string where the client tells the 
server what IP and port to connect back to, the firewall then re-writes 
the packet to an IP and port on the firewall, and sets up a NAT rule on 
the firewall so that traffic from the server to that IP and port on the 
firewall gets directed to the IP and port the client specified.

If javascript issues a HTTP request to port 21 on the bad guy's server and 
in the request sends the string that the firewall is looking for, the 
firewall will helpfully setup a NAT and allow that server to connect to a 
port on the client as specified by the string that the javascript code 
specified.

There is no need to exploit UPnP or anything other than the fact that the 
firewall is watching for a specific string (without understanding and 
enforcing the full protocol) and then opening up a incoming port.

SIP has similar features, but instead of running on a standard low port 
that coudl be blocked by the browser, it is running on a high port that is 
much more risky for the browser to block access to.

David Lang

On Tue, 12 Jan 2010, Farrukh Haroon wrote:

> Perhaps they are exploiting UPnP in some obscure way to achieve this?
>
> Regards
>
> Farrukh
>
> On Tue, Jan 12, 2010 at 5:51 AM, <[email protected]> wrote:
>
>> I've seen several other posts where people make use of browser exploits to
>> trick the browser into submitting a form to the router/firewall, and if the
>> router has the default password, the attacker can then configure the
>> firewall any way they want.
>>
>> This sounds a little different. This sounds like it is exploiting standard
>> protocols.
>>
>> With FTP the client connect to the server, then at the start of a file
>> transfer the client tells the server what port to connect to on the client.
>> A 'helpful' firewall will watch for this message and reconfigure itself to
>> allow traffic to that port. IIRC for FTP this data connection is one-way
>> (with acks flowing the other way), but with SIP the port is used for data in
>> both directions.
>>
>> This sounds like the attacker is managing to use javascript to make a
>> connection out that the firewall thinks is a protocol like this, and then by
>> specifying the port they want to attack, tricking the firewall into opening
>> that port up so that it can be attacked from the server the javascript
>> connected to.
>>
>> David Lang
>>
>>
>>
>>
>> On Fri, 8 Jan 2010, R. DuFresne wrote:
>>
>>  -----BEGIN PGP SIGNED MESSAGE-----
>>> Hash: SHA1
>>>
>>>
>>>
>>> In reading this, I get the impression this is not a fault in the firewalls
>>> themselves, but more an issue with the configuration of firewalls having
>>> been 'tested' by this hacker.  Am I wrong in reading this news in that
>>> fashion?::
>>>
>>>
>>> January 6, The Register - (International) Hacker pierces hardware
>>> firewalls with web page. On January 5, a hacker demonstrated a way to
>>> identify a browser's geographical location by exploiting weaknesses in many
>>> WiFi
>>> routers. Now, the same hacker is back with a simple method to penetrate
>>> hardware firewalls using little more than some javascript embedded in a
>>> webpage. By luring victims to a malicious link, the attacker can access
>>> virtually any service on their machine, even when it's behind certain
>>> routers that automatically block it to the outside world. The method has
>>> been tested on a Belkin N1 Vision Wireless router, and the hacker says he
>>> suspects other devices are also vulnerable. "What this means is I can
>>> penetrate their firewall/router and connect to the port that I specified,
>>> even though the firewall should never forward that port," the hacker told
>>> the Register. "This defeats that security by visiting a simple web page. No
>>> authentication, XSS, user input, etc. is required." The hacker's
>>> proof-ofconcept page forces the visitor to submit a hidden form on port
>>> 6667, the standard port for internet relay chat. Using a hidden value, the
>>> form surreptitiously coerces the victim to establish a DCC, or direct
>>> client-to-client, connection. Vulnerable  routers will then automatically
>>> forward DCC traffic to the victim's internal system, and using what's known
>>> as NAT traversal an attacker can  access any port that's open on the local
>>> system. For the hack to work, the visitor must have an application such as
>>> file transfer protocol or session initiation protocol running on his
>>> machine. The hack does not guarantee an attacker will be able to compromise
>>> that service, but it does give the  attacker the ability to probe it in the
>>> hope of finding a weak password or a vulnerability that will expose data or
>>> system resources. Source:
>>> http://www.theregister.co.uk/2010/01/06/web_based_firewall_attack/
>>>
>>>
>>>
>>>
>>> Thanks,
>>>
>>>
>>> Ron DuFresne
>>> - -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>>       admin & senior security consultant:  sysinfo.com
>>>                       http://sysinfo.com
>>> Key fingerprint = 9401 4B13 B918 164C 647A  E838 B2DF AFCC 94B0 6629
>>>
>>> These things happened. They were glorious and they changed the world...,
>>> and then we fucked up the endgame.    --Charlie Wilson
>>> -----BEGIN PGP SIGNATURE-----
>>> Version: GnuPG v1.4.5 (GNU/Linux)
>>>
>>> iD8DBQFLR3qist+vzJSwZikRAotcAJ9fHEWAOm2N5xFKww7/wA9O+YYdeACfZUEZ
>>> uZciRDQsRu1kZZQUZctPwmY=
>>> =KCsu
>>> -----END PGP SIGNATURE-----
>>> _______________________________________________
>>> firewall-wizards mailing list
>>> [email protected]
>>> https://listserv.icsalabs.com/mailman/listinfo/firewall-wizards
>>>
>>>  _______________________________________________
>> firewall-wizards mailing list
>> [email protected]
>> https://listserv.icsalabs.com/mailman/listinfo/firewall-wizards
>>
>