From mboxrd@z Thu Jan 1 00:00:00 1970 From: Neil Horman Subject: Re: [PATCH RFC] r8169: straighten out overlength frame detection (v3) Date: Sat, 30 Jan 2010 16:50:26 -0500 Message-ID: <20100130215026.GA6146@localhost.localdomain> References: <20100108.160252.189352309.davem@davemloft.net> <1263088638.2480.210.camel@localhost> <20100110235017.GA8959@electric-eye.fr.zoreil.com> <20100110.224504.247142858.davem@davemloft.net> <20100112001646.GA7017@electric-eye.fr.zoreil.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: David Miller , ben@decadent.org.uk, eric.dumazet@gmail.com, netdev@vger.kernel.org To: Francois Romieu Return-path: Received: from charlotte.tuxdriver.com ([70.61.120.58]:37056 "EHLO smtp.tuxdriver.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753014Ab0A3Vuv (ORCPT ); Sat, 30 Jan 2010 16:50:51 -0500 Content-Disposition: inline In-Reply-To: <20100112001646.GA7017@electric-eye.fr.zoreil.com> Sender: netdev-owner@vger.kernel.org List-ID: On Tue, Jan 12, 2010 at 01:16:46AM +0100, Francois Romieu wrote: > On Sun, Jan 10, 2010 at 10:45:04PM -0800, David Miller wrote: > [...] > > In my opinion if you reset, you give them more power. > > > > Instead of just dropping the next few frames, you allow them > > to cause a drop of how ever many RX frames can arrive during > > the reset period _PLUS_ the amount of other RX frames which > > were in the receive ring at the point of detection. > > Sure. > > The fragmented frame test has been moved a few lines up, before checking > for the status bits. Exceedingly small frames are detected and dropped > as well. > > I have spent the evening disrupting the communication. As a challenger > one does not even need complete control : send a few random packets and > the card goes to neverland. :o/ > > I'll do some tests limiting the crap at the first packet. > Any further results you can share with us Francois? Thanks Neil