From mboxrd@z Thu Jan 1 00:00:00 1970 From: Francois Romieu Subject: Re: Possible bug with r8169 driver Date: Sat, 29 Sep 2012 09:20:28 +0200 Message-ID: <20120929072028.GA3730@electric-eye.fr.zoreil.com> References: <2197741.c8FeV4Xbis@bureau> <20120927222147.GA29047@electric-eye.fr.zoreil.com> <1478421.GOy9qdQlkG@bureau> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: hayeswang , netdev@vger.kernel.org To: Nolwenn Return-path: Received: from violet.fr.zoreil.com ([92.243.8.30]:49790 "EHLO violet.fr.zoreil.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755175Ab2I2Hjg (ORCPT ); Sat, 29 Sep 2012 03:39:36 -0400 Content-Disposition: inline In-Reply-To: <1478421.GOy9qdQlkG@bureau> Sender: netdev-owner@vger.kernel.org List-ID: Nolwenn : > Le vendredi 28 septembre 2012 00:21:47 Francois Romieu a =C3=A9crit : [...] > > Can you send an 'ip -s link' before any ipv6 traffic flows, then > > after ? >=20 > Before ipv6 traffic >=20 > % ip -s link /me plugs brain and realizes that the driver does not update the device stats multicast field. The multicast stats will appear in the output of 'ethtool -S eth0'. Can you send two complete binary tcpdump captures (w/o -p option) in th= e same contexts ? Don't bother obfuscating the ipv6 address if it equals the one appearin= g in the mail headers. [RTL8168f/8111f] > > Hayes, is there by any luck something different for the 8168f regar= ding > > the layout of the multicast filtering registers ? >=20 > I don't have sufficient skills in networks to understand what do you = mean,=20 > sorry. Any command line, files comparing to get the information ? The documentation is scarce to say the least. Hayes can tell if some register layout change or chipset specifics has gone unnoticed. --=20 Ueimor