From mboxrd@z Thu Jan 1 00:00:00 1970 From: Benjamin Herrenschmidt Subject: Re: [PATCH v2 00/13] ftgmac100: Rework batch 1 - Link & Interrupts Date: Fri, 07 Apr 2017 07:46:58 +1000 Message-ID: <1491515218.4166.124.camel@kernel.crashing.org> References: <20170405022853.8238-1-benh@kernel.crashing.org> <2344b1ea-f8bc-7491-4662-b8a3e94fb94e@gmail.com> <1491371595.4166.77.camel@kernel.crashing.org> <1491373880.4166.79.camel@kernel.crashing.org> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 8bit To: Florian Fainelli , netdev@vger.kernel.org Return-path: Received: from gate.crashing.org ([63.228.1.57]:37154 "EHLO gate.crashing.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754065AbdDFVrM (ORCPT ); Thu, 6 Apr 2017 17:47:12 -0400 In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Thu, 2017-04-06 at 12:46 -0700, Florian Fainelli wrote: > > I thought a while ago we could add some dev flag to prevent the link > > watch from doing that, but never got to look into it myself and > > apparently neither did Gavin. > > It sounds like a similar situation to e.g: 802.1x, you want to let some > frames make it through and you need a working link for that, but you > can't quite flag the device as UP entirely. Maybe you can find a way to > re-use IFF_LOWER_UP to that purpose and only set that flag based on the > NC-SI frames indicating link state? Possibly yes, I wasn't familiar with the difference between  "carrier" and IFF_LOWER_UP so far. I'll have a look or have Gavin do so. It would be nice to reflect that properly so things like fallover or bonding do the right thing. On the other hand I yet have to see a system where the BMC has more than one NC-SI link to a NIC (what they usually have is that NC-SI is connected to multiple NICs which is a different matter and should be handled within the NC-SI stack). Anyway, I'll look into it further down the track along with other improvements I'd like to do to the NC-SI stack such as properly setting up multicast and vlan filters in the peer NIC and various cleanups. Cheers, Ben.