On Tue, 2010-09-21 at 09:51 -0500, Robert G. (Doc) Savage wrote:

> I was expecting to see something similar to the output I got for
> the F13 kernel:
> 
>         $$$ Kernel release: 2.6.34.6-54.fc13.x86_64
>         !!! Could not find symbol: per_cpu__current_task
>         
>         A symbol required by the published exploit for CVE-2010-3081 is
>         not provided by your kernel.  The exploit would not work on your
>         system.
>         
> Thoughts?

The exploit is caused by a failure to correctly check the access and
range (for potential underflow) of a value passed to the kernel from
userspace. It's a classical exploit and it was fixed promptly.

The message you are getting above is different. All kernels provide a
number of "symbols", which are exported functions available for use by
drivers and other loadable modules. The specific incarnation of the
exploit that was being examined by the Ksplice tool looked for the
per_cpu__current_task symbol (presumably as part of the kernel stack
corruption exercise required for the exploit - I didn't check), which
isn't available on some kernels. That doesn't mean they are not
vulnerable to the compat exploit, just that they don't have this
particular symbol exported. Red Hat fixed the exploit without affecting
which symbols were or were not exported by the RHEL5 kernel, because
that was not the actual problem. Again, Ksplice did a good job with a
utility, but it is just a handy utility that helped some folks look to
see if their systems might be exploited by one version of the exploit.

Does that help?

Jon.


_______________________________________________
rhelv5-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/rhelv5-list

Reply via email to