Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 01 Feb 2012 13:53:40 +0100
From: Berke Viktor <berkeviktor@....com>
To: oss-security@...ts.openwall.com
Subject: Re: CVE Request (two ids) -- Xchat-WDK (prior 1499-4 [2012-01-18])
 and Xchat-v2.8.6 on Maemo architecture -- Heap-based buffer overflow by processing
 UTF-8 line from server containing characters outside BMP

Hello,

Here are my notes:

- Apparently only Windows versions are affected, no Linux ones. I 
haven't tested Maemo but I'd be suprised if it would crash.
- Not all non-BMP characters crash, only a specific range. See the patch 
you linked for details.

if ((suspect >= 0x1D173 && suspect <= 0x1D17A)
			|| (suspect >= 0xE0001 && suspect <= 0xE007F))

- As for your assumption that private messages would still crash, I 
already made a correction in the bug tracker, but for your reference:

"ANY received text will be filtered correctly, be it private or public 
message or anything else. XChat-WDK will ONLY crash if you paste the 
malicious characters into your own client's input box, for which the 
only reason could be to intentionally crash yourself. This obviously 
can't be prevented in XChat-WDK, only if GTK+ fixes it."

That is, current versions of XChat-WDK are immune to these kinds of attacks.

Regards,

Viktor

On 2012.02.01. 11:55, Jan Lieskovsky wrote:
> Hello Kurt, Steve, Viktor, vendors,
>
> a heap-based buffer overflow flaw was found in the way xchat, graphical IRC
> chat client, processed one line of text received from the server, when
> the text
> contained Unicode characters and some of the characters were outside of the
> Basic Multilingual Plane (BMP). A remote attacker could provide a
> specially-crafted Unicode string as a xchat channel or private message,
> which
> once processed would lead to denial of service (xchat client crash), or,
> potentially arbitrary code execution with the privileges of the user
> running
> xchat client.
>
> This issue has been successfully reproduced on Xchat-WDK versions prior to:
> * 1499-4 (2012-01-18)
>
> add Non-BMP plugin to avoid client crashes
>
> version. Also Joerg Reisenweber reports, this deficiency to have been
> exploited
> in the past on Xchat-v2.8.6 versions, as being used on Maemo architecture.
>
> The following Linux based xchat versions have been investigated against
> presence
> of this issue:
> * xchat-v2.6.6,
> * xchat-v2.8.6,
> * xchat-v2.8.8
>
> on various architectures (i386, x86_64, ppc64) with various versions of
> gtk2 library:
> * gtk-v2.10.4,
> * gtk-v2.18.9,
> * gtk-v2.24.7,
> * gtk-v2.14.7
>
> and presence of this flaw has not been observed on those Linux versions,
> which makes
> us think it is some Microsoft Windows 7 / Maemo architecture specific
> feature, which
> makes this issue to be visible on those Xchat derivatives.
>
> References:
> [1] http://code.google.com/p/xchat-wdk/issues/detail?id=132
> [2] http://code.google.com/p/xchat-wdk/issues/detail?id=134
> [3] http://code.google.com/p/xchat-wdk/issues/detail?id=135
> [4] https://bugzilla.redhat.com/show_bug.cgi?id=786391
>
> Xchat-WDK upstream changelog:
> [5] http://www.xchat-wdk.org/home/changelog
> part:
> * 1499-4 (2012-01-18)
>
> add Non-BMP plugin to avoid client crashes
>
> Particular Xchat-WDK upstream patch:
> [6] http://lwsitu.com/xchat/replace_non-bmp.diff
>
> Could you allocate two CVE ids for these flaws? (assuming two ids are
> necessary, because Xchat-WDK for MS Windows 7 case and Xchat-v2.8.6 for
> Maemo case can / should be considered as different source code bases).
>
> Steve, please advise if one id is sufficient or two should be used?
>
> Also, for the Xchat-WDK case it looks that v1499-6 corrected the issue
> for channel messages, but the issue is still present for 'private messages'
> case:
> [7] http://code.google.com/p/xchat-wdk/issues/detail?id=132#c33
> [8] http://code.google.com/p/xchat-wdk/issues/detail?id=132#c34
> [9] http://code.google.com/p/xchat-wdk/issues/detail?id=132#c36
>
> Though this assumption needs to be verified / confirmed yet.
> Viktor, could you please confirm or disprove it?
>
> If that assumption would have shown as valid, a third CVE identifier
> would need to be assigned yet for the incomplete Xchat-WDK v1499-6
> fix yet (addressing the issue for 'channel messages' case, but not
> for 'private messages' case).
>
> Thank you && Regards, Jan.
> --
> Jan iankko Lieskovsky / Red Hat Security Response Team

Powered by blists - more mailing lists

Please check out the Open Source Software Security Wiki, which is counterpart to this mailing list.

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.