Re: CVE Request New-djbdns: dnscache: potential cache poisoning

[Michael Samuel <mik () miknet net>]

On 17 February 2014 19:33, P J P <ppandit () redhat com> wrote:

>   Hello,
>
> +-- On Wed, 12 Feb 2014, P J P wrote --+
> | +-- On Wed, 12 Feb 2014, Michael Samuel wrote --+
> | | >  -> http://www.openwall.com/lists/oss-security/2014/02/11/7
> | | The same issue, different result.
> |
> |   Yes, true. Thank you for confirming.
>
>   So, does this qualify for a CVE?


I think I've come around to a yes for this one. Pushing attacker-chosen
entries
out of the cache after only 100 packets is clearly not what the admin
wants.  It
makes a secondary attack (DNS over UDP blind cache poisoning) much more
viable than it was.

I can think of some DoS scenarios where this vector would assist another
attack.

Regards,
  Michael