Mailing-List:	contact cygwin-help AT cygwin DOT com; run by ezmlm
List-Subscribe:	<mailto:cygwin-subscribe AT cygwin DOT com>
List-Archive:	<http://sources.redhat.com/ml/cygwin/>
List-Post:	<mailto:cygwin AT cygwin DOT com>
List-Help:	<mailto:cygwin-help AT cygwin DOT com>, <http://sources.redhat.com/ml/#faqs>
Sender:	cygwin-owner AT cygwin DOT com
Delivered-To:	mailing list cygwin AT cygwin DOT com
Message-Id:	<3.0.5.32.20020222220653.007e8450@pop.ne.mediaone.net>
X-Sender:	phumblet AT pop DOT ne DOT mediaone DOT net (Unverified)
X-Mailer:	QUALCOMM Windows Eudora Pro Version 3.0.5 (32)
Date:	Fri, 22 Feb 2002 22:06:53 -0500
To:	Corinna Vinschen <cygwin AT cygwin DOT com>
From:	"Pierre A. Humblet" <Pierre DOT Humblet AT ieee DOT org>
Subject:	Re: More security issues
In-Reply-To:	<20020214101329.O23094@cygbert.vinschen.de>
References:	<3 DOT 0 DOT 5 DOT 32 DOT 20020213155051 DOT 007cce50 AT pop DOT ne DOT mediaone DOT net>
	<3 DOT 0 DOT 5 DOT 32 DOT 20020210143455 DOT 007f2100 AT pop DOT ne DOT mediaone DOT net>
	<3 DOT 0 DOT 5 DOT 32 DOT 20020210143455 DOT 007f2100 AT pop DOT ne DOT mediaone DOT net>
	<3 DOT 0 DOT 5 DOT 32 DOT 20020213155051 DOT 007cce50 AT pop DOT ne DOT mediaone DOT net>
Mime-Version:	1.0

Hi Corinna

At 10:13 AM 2/14/2002 +0100, you wrote:

>The sec_user() call in CreateProcess()
>was never intended to set the default DACL (I didn't even know
>that something like that exists when I added that) but to set the
>permissions to access the process.  
<snip>
Yes, and in the case of DuplicateTokenEx(), the permissions of the
new primary token. However the sd's of a new process TOKEN and of a 
new impersonation token are always initialized from the default in the 
(parent) process token. I think I now understand what's going on.
The confusion between the impersonated sid and the original sid 
that we have observed in LookupAccountSid() is also present in the 
token sd, but things work out all right, for some reason. I will 
send you some patches shortly.
<snip>

>When I implemented this, the fork/exec implementation was pretty
>different from today.  As far as I rememeber, the code which copied
>data from one process to the other needed access under the 2nd SID.
>This could qualify for some code which could be pretty useless
>today.  E.g. your observation that RevertToSelf() could be dropped,
>probably.

I am still looking at that. On 2001-10-31 you added RevertToSelf() in 
dtable.cc (dtable::vfork_child_dup). Do you remember why?

Pierre


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Bug reporting:         http://cygwin.com/bugs.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

- Raw text -