From: Stas Sergeev Subject: Re: [5/6] mvneta: implement SGMII-based in-band link state signaling Date: Thu, 09 Jul 2015 13:11:10 +0300 Message-ID: <559E48BE.5010801@list.ru> References: <55155D9D.6070800@list.ru> <559D5520.3030700@list.ru> <20150709111941.6a49091f@free-electrons.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: netdev@vger.kernel.org, Linux kernel , Stas Sergeev To: Thomas Petazzoni , Sebastien Rannou Return-path: Received: from smtp31.i.mail.ru ([94.100.177.91]:59399 "EHLO smtp31.i.mail.ru" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751154AbbGIKLS (ORCPT ); Thu, 9 Jul 2015 06:11:18 -0400 In-Reply-To: <20150709111941.6a49091f@free-electrons.com> Sender: netdev-owner@vger.kernel.org List-ID: 09.07.2015 12:19, Thomas Petazzoni =D0=BF=D0=B8=D1=88=D0=B5=D1=82: > Sebastien, Stas, >=20 > On Thu, 9 Jul 2015 11:03:26 +0200 (CEST), Sebastien Rannou wrote: >> On Wed, 8 Jul 2015, Stas Sergeev wrote: >> >>> What is there? A phy chip, or something else? >> >> It's "something else", there's a phy which aggregates 4xSGMIIs to >> 1xQSGMII, we are on the media side here, the MAC side is connected >> to the switch through QSGMII. >> >>> Perhaps some DT property should be added to explicitly >>> enable the use of the inband status... >> >> Yes, that would be fine. >=20 > Isn't it a bit weird to need a new DT property for this? Shouldn't > fixed-link always imply this inband status thing? That's how it is currently implemented. I thought its safe. But what if the device on the other end does not generate the inband status? I think this device is doing the wrong thing, but nevertheless we have a regression at hands. Currently the link status cannot be specified for fixed-link, at all. What I am going to code up, is the new property, like this: fixed-link { link =3D "up" | "down" | "auto"; }; "auto" will mean the inband status. Looks like a simple solution.