oss-sec mailing list archives

Re: CVE request: phpMyAdmin < 3.1.1.0 (SQL injection through XSRF on several pages )


From: Tomas Hoger <thoger () redhat com>
Date: Wed, 17 Dec 2008 15:40:01 +0100

On Tue, 16 Dec 2008 20:52:42 -0500 (EST) "Steven M. Christey"
<coley () linus mitre org> wrote:

Two separate CVE's are assigned, one for the original milw0rm exploit
and the other for the unspecified vectors implied by the implied
"XSRF on several pages" in the PMASA-2008-10 advisory.

Are those really separate issues?  I believe that -5622 was assigned
because of the following mention in the upstream ChangeLog:
  - [security] possible XSRF on several pages

However, that ChangeLog entry was added as a not too good description
of the fix for the SQL injection described in milw0rm 7382.  Upstream
commit is referenced in the PMASA-2008-10:
http://phpmyadmin.svn.sourceforge.net/viewvc/phpmyadmin?view=rev&revision=12100

I think that the wording here is bit confusing.  I think this kind of
flaw would normally be described "privileged / logged-in user SQL
injection".  Though as this is SQL DB management application, once you
are logged in, you can execute SQL commands using the standard
application features, rather than having to find some privileged user
SQL injection flaw.   Hence this is only exploitable via CSRF-like
methods.  Not sure if the CSRF term is right there, as this "CSRF" does
not seem to do any harm without SQL injection flaw.

Or were there any other reasons for split?

-- 
Tomas Hoger / Red Hat Security Response Team


Current thread: