[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Tue, 5 May 2009 17:21:44 +0930
From: Mark Smith
<lnx-netdev@...22607b6285f9c5d5ea31ea9d8a7ac.nosense.org>
To: Andi Kleen <andi@...stfloor.org>
Cc: netdev@...r.kernel.org
Subject: Re: [RFC, PATCH 2.6.29.2] Ethernet V2.0 Configuration Testing
Protocol, revision 20090428
Hi Andi,
Thanks for your interest.
On Mon, 04 May 2009 11:29:55 +0200
Andi Kleen <andi@...stfloor.org> wrote:
> Mark Smith <lnx-netdev@...22607b6285f9c5d5ea31ea9d8a7ac.nosense.org> writes:
> > +
> > +4. Security
> > +
> > +ECTP was designed in the early 1980s, when protocol security was less of
> > +a concern than it is now. Consequently, there are some features of the
> > +protocol which could be abused for nefarious purposes. By default, this
> > +implementation attempts to avoid participating in them. These features
> > +could be useful for some test cases thought, so they can be enabled if
> > +required.
>
> I think security would need quite a bit more discussion. Opening new
> DOS this way sounds quite worrying, especially since this is a
> extremly obscure protocol that likely most admins don't know much
> about.
>
> Is this suspencible to ping to broadcast flood replication for example?
>
No it isn't. Forward addresses in forward messages (specifying the next
hop on the path) are prohibited by the protocol specification from
being multicast or broadcast addresses.
> Safest would probably be default to off.
>
In my implementation, this unicast forward address check can't be
switched off.
Some of the other security mitigation mechanisms I've implemented
weren't really necessary, however I felt that it'd be good to ensure
this implementation was as good a "citizen" as possible.
Thanks,
Mark.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists