oss-security - Re: openssl security issue or not? (CVE Request?)

Follow @Openwall on Twitter for new release announcements and other news

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <OF5EE2B5C5.F11C9695-ON862579CA.005960A4-862579CA.005A991A@us.ibm.com>
Date: Fri, 23 Mar 2012 11:29:33 -0500
From: Ivan Nestlerode <inestlerode@...ibm.com>
To: Marcus Meissner <meissner@...e.de>
Cc: OSS Security List <oss-security@...ts.openwall.com>
Subject: Re: openssl security issue or not? (CVE Request?)

Marcus Meissner <meissner@...e.de> wrote on 03/23/2012 11:13:20 AM:
> From: Marcus Meissner <meissner@...e.de>
> To: OSS Security List <oss-security@...ts.openwall.com>
> Cc: Ivan Nestlerode/Cambridge/IBM@...US
> Date: 03/23/2012 11:13 AM
> Subject: openssl security issue or not? (CVE Request?)
> 
> Hi folks, Ivan,
> 
> This patch:
> http://cvs.openssl.org/chngview?cn=22161
> fixes a decrypt error return values and according to the changelog
> "detects symmetric crypto errors" 
> 
> I am not sure if this counts as security issue in the end, but "not
> detecting a failed decrypt" seems to me like it is a security issue.
> 
> Any comments?
> 
> Ciao, Marcus
> (also https://bugzilla.novell.com/show_bug.cgi?id=749210 ) 

Marcus,

I don't think that this change needs separate tracking as a security issue 
since there is no guarantee that you can detect symmetric decryption 
errors anyway (it is not that hard for random gibberish to decrypt with 
correct PKCS#5 padding).  This change made its way into their CVS first, 
but it was actually meant to go in along with another change that is a 
security fix and already has a CVE entry associated with it 
(CVE-2012-0884):
http://cvs.openssl.org/chngview?cn=22238

If a Linux distribution picks up the fix for CVE-2012-0884 then they will 
want to pick up change 22161 at the same time since the fix for the 
security vulnerability will generally cause symmetric decryption errors 
when it kicks in and things get very confusing for the end user without 
change 22161 (they will frequently get junk results with a success return 
value instead of an error return value).  Both were reported to OpenSSL at 
the same time, but the CVE change was submitted later because it was more 
complicated and required more review and discussion.

Hope that helps,
-Ivan

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.