netdev - Re: [PATCH 1/2] net: Toeplitz library functions

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CA+mtBx8Nm_wQcizqD7ku5xONurqQaGZpqX0AafUrC7+4uU34xQ@mail.gmail.com>
Date:	Tue, 24 Sep 2013 12:04:25 -0700
From:	Tom Herbert <therbert@...gle.com>
To:	Jesse Brandeburg <jesse.brandeburg@...el.com>
Cc:	David Miller <davem@...hat.com>,
	David Laight <David.Laight@...lab.com>,
	Linux Netdev List <netdev@...r.kernel.org>
Subject: Re: [PATCH 1/2] net: Toeplitz library functions

On Tue, Sep 24, 2013 at 11:48 AM, Jesse Brandeburg
<jesse.brandeburg@...el.com> wrote:
> On Tue, 24 Sep 2013 14:03:12 -0400 David Miller <davem@...hat.com> wrote:
> ...
>> >> For security reasons we absolutely cannot use it for that purpose,
>> >> please stop claiming this.
>> >>
>> >> Any hash function which an attacker can reproduce is attackable.
>> >
> ...
>> > that should be addressed.  It is possible to DoS attack through the
>> > steering mechanism.
>>
>> All of them are using a fixed, defined, key.
>
> We selected the fixed key on purpose.  The existing mechanisms built
> into the stack for preventing the impact of DOS attacks like NAPI
> polling will prevent any actual damage even if someone sends lots of
> packets on a single flow.  If someone overflows a receive queue that
> CPU runs until it can't keep up and then hardware drops further
> packets.  In this case even with a randomized seed key any single flow
> can still be targeted at a queue, which is no different than a single
> queue adapter.
>
> I'm not convinced there is an actual impact in practice.
>
But what is the purpose of using a fixed key, what are the benefits?
A clever attacker could attack a specific queue with different
4-tuples.  While this might be similar to a single flow attack, it
will be much harder to detect and be used to attack multiple servers
simultaneously in the say way (since they all share the same key).  I
don't think we could deploy a NIC with a static RSS key.
--
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