Re: Linux kernel futex local privilege escalation (CVE-2014-3153)

[Greg KH <greg () kroah com>]

On Fri, Jun 06, 2014 at 11:58:46AM -0400, Rich Felker wrote:
> On Fri, Jun 06, 2014 at 05:43:28PM +0200, rf () q-leap de wrote:
> >     Greg> There is someone still maintaining 3.12-stable, why not rely
> >     Greg> on those releases if you want that kernel version, instead of
> >     Greg> rolling your own?
> >
> > We thankfully do rely on that as our base. In this case though, the
> > patches haven't been ported until this moment. And I can't wait for them
> > to appear since there is no time-line when that will happen ...
>
> Indeed. This is probably the biggest security flaw in Linux in the
> past 5 years (if not the biggest ever) since it allows a full kernel
> compromise even from extremely tight sandboxes. In my opinion, the way
> the announcement was handled was really unprofessional. There should
> have been fixes prepared for, and/or committed into the git repos for,
> all currently maintained releases/branches at the time of the
> announcement. Anything else leaves everybody but users of the big
> mainstream distros scrambling to figure out how to get a
> non-vulnerable kernel that's compatible with their current setups.

That was planned, but something happened which caused the issue to
"leak" much too early.  It was not intentional at all, but rather a
human error.  The parties involved are very sorry about it, there was no
malicious intention at all involved.

Stuff happens, sorry.

greg k-h