From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ivan Frederiks Subject: Re: Realtek RTL8102E registers Date: Wed, 06 Nov 2013 14:13:14 +0400 Message-ID: <527A163A.9010409@speechpro.com> References: <526FCEAA.30102@speechpro.com> <20131029214059.GA26496@electric-eye.fr.zoreil.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Cc: Linux r8169 crew To: Francois Romieu Return-path: Received: from spb2.speechpro.com ([149.126.18.58]:59366 "EHLO spb2.speechpro.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756185Ab3KFKN2 (ORCPT ); Wed, 6 Nov 2013 05:13:28 -0500 In-Reply-To: <20131029214059.GA26496@electric-eye.fr.zoreil.com> Sender: netdev-owner@vger.kernel.org List-ID: Hello Francois, I found solution for the issue that I described earlier: one has to set correct MAC-address before bringing up the interface. On my board RTL8102E works without EEPROM, so initial MAC-address value is unpredictable. In many cases this value has multicast bit set to 1. This fact breaks driver and (or) TCP stack operation: * r8169 reports that link is down * r8101 reports that link is OK, but Linux TCP stack ignores packets with multicast destination MAC Hope this info would be useful. BTW, I suppose that driver could have some kind of protection against such situations. On 30.10.2013 01:40, Francois Romieu wrote: > Ivan Frederiks : > [...] >> Maybe you have access to RTL8102E register description ? > > No. > > Realtek's hardware documentation policy isn't exactly opened. Your hardware > vendor may help. > > [...] >> Issue description: > [snip] > > kernel version and r81xx XID value as it should appear in dmesg would > be welcome. > -- Best regards, Ivan Frederiks Embedded developer Speech Technology Center Phone: +7-812-331-0665, ext. 6123, 6942 Fax: +7-812-327-9297