[26624] in bugtraq
RE: White paper: Exploiting the Win32 API.
daemon@ATHENA.MIT.EDU (Kenn Humborg)
Sat Aug 10 21:58:21 2002
From: "Kenn Humborg" <kenn@bluetree.ie>
To: "Chris Calabrese" <chris_calabrese@yahoo.com>, <bugtraq@securityfocus.com>
Date: Wed, 7 Aug 2002 18:18:15 +0100
Message-ID: <NBBBIGEGHIGMPCNKHCECIEGBEAAA.kenn@bluetree.ie>
MIME-Version: 1.0
Content-Type: text/plain;
charset="us-ascii"
Content-Transfer-Encoding: 7bit
In-Reply-To: <20020807133813.31892.qmail@web13304.mail.yahoo.com>
> So let me get this straight.
>
> Allowing unpriveleged processes to send control messages to priveleged
> processes is not a flaw in the Win32 API because there is a mechanism
> for applications to protect themselves from this type of attack
> (alternate Windows Stations/Desktops).
>
> But the mechanism effectively prevents the priveleged processes from
> providing a GUI because the user won't be able to actually see the
> alternate Windows Stations/Desktops without some kind of Station
> switching tool, and/or extra training in how to do this.
>
> So, the result is that no applications actually use this mechanism.
>
> What part of "this is broken" doesn't make sense?
Well, there is a Right Way of controlling privileged processes
from the user's desktop. Simply banging a window up on the
desktop is not the Right Way.
There should be a separate UI/management tool which runs under
the user's credentials on his desktop and uses some IPC mechanism
to control the privileged process (RPC, DCOM, named pipes, tcp
sockets, whatever). This IPC interface is the security boundary.
To make an analogy in the Unix world, it would be like a deamon
running as root opening an xterm on the users desktop to manage
it. Nobody would say "X is broken" in this case - I think we'd
all agree that the app is broken.
Later,
Kenn
