Information Security News mailing list archives

Re: MS Patches Trio of Windows Flaws


From: InfoSec News <isn () c4i org>
Date: Fri, 18 Jul 2003 02:46:54 -0500 (CDT)

Forwarded from: Mark Bernard <mbernard () nbnet nb ca>

Dear Associates,

By now we are all perhaps to familiar with the phase
'buffer-over-flow'. But as Microsoft apparently is systematically
working through each and every open buffer has anyone stopped to ask
the question, what will be the draw backs of these changes? Could
these changes inherently create a domino affect and/or lead to more
vulnerabilities or different vulnerabilities? Maybe even more sever
vulnerabilities.

When I did write code once upon a time, we used many, many different
ways to pass information from one application to another. A buffer
zone only represented one type of memory allocation device.
Controlling a buffer will limit the flexibility of that particular
application no doubt while reducing risk. However, what about all the
other allocated fields within the very same buffer or what about the
other types memory allocation devices such as arrays, indexes, etc...

It would be interesting to know how many of these on going problems
are truly bran new and how many are being created by each new patch.
Something to think about.

Regards,
Mark E. S. Bernard, CISM, CISSP



----- Original Message ----- 
From: "InfoSec News" <isn () c4i org>
To: <isn () attrition org>
Sent: Thursday, July 17, 2003 4:45 AM
Subject: [ISN] MS Patches Trio of Windows Flaws


http://www.eweek.com/article2/0,3959,1187632,00.asp

By Dennis Fisher
July 9, 2003

Microsoft Corp. on Wednesday issued patches for three flaws in
various versions of Windows, two of which give attackers the ability
to run whatever code they wish on vulnerable machines.

The most serious of the vulnerabilities affects all currently
supported versions of Windows, from Windows 98 up through Windows
Server 2003. The problem lies in the HTML converter, which allows
users to handle HTML files. A vulnerability results from the way the
converter handles conversion requests during cut-and-paste
operations.

An attacker who could create a special conversion request could
cause the converter to fail in a way that enables the attacker to
execute code on the user's machine. The code would run with the
user's privileges. The patch for this flaw is here [1].

The second vulnerability affects Windows NT 4.0, Windows 2000 and XP
Professional and results from a buffer overrun in a portion of the
operating system that handles Server Message Block requests. When
the Windows server receives SMB packets, it fails to validate the
length of the buffer established by the packet. As a result, an
attacker could use a malicious SMB request to overrun the buffer,
which would cause one of three things to happen: data corruption, a
system failure or code execution.

However, in order to exploit this flaw, the attacker would need to
be authenticated to the server. The patch for this issue is located
here [2].

The third flaw affects only Windows 2000 and results because the
Windows Utility Manager handles some messages incorrectly. The
control that provides the list of accessibility options to the user
doesn't validate Windows messages sent to it. This allows one
interactive process to use a specific message to cause the Utility
Manager to execute a callback function to the address of its choice.

Because the Utility Manager runs at a higher privilege level, this
would allow the inferior process with a way to use the Utility
Manager's privilege set. An attacker able to exploit this would have
complete control over the compromised system, Microsoft said in its
bulletin. But, this flaw cannot be exploited remotely.

The patch for this vulnerability is here [3].

[1] http://www.microsoft.com/security/security_bulletins/ms03-023.asp
[2] http://www.microsoft.com/security/security_bulletins/ms03-024.asp
[3] http://www.microsoft.com/security/security_bulletins/ms03-025.asp



-
ISN is currently hosted by Attrition.org

To unsubscribe email majordomo () attrition org with 'unsubscribe isn'
in the BODY of the mail.


Current thread: