From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jakub Kicinski Subject: Re: [PATCH net-next v2 RFC 7/8] devlink: Add a generic port parameter Date: Wed, 12 Dec 2018 11:01:51 -0800 Message-ID: <20181212110151.3150c50e@cakuba.netronome.com> References: <1544518008-15289-1-git-send-email-vasundhara-v.volam@broadcom.com> <1544518008-15289-8-git-send-email-vasundhara-v.volam@broadcom.com> <20181211101839.01d817ac@cakuba.netronome.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: Vasundhara Volam , David Miller , Jiri Pirko , Netdev To: Michael Chan Return-path: Received: from mail-qt1-f196.google.com ([209.85.160.196]:33258 "EHLO mail-qt1-f196.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726235AbeLLTB4 (ORCPT ); Wed, 12 Dec 2018 14:01:56 -0500 Received: by mail-qt1-f196.google.com with SMTP id l11so21771942qtp.0 for ; Wed, 12 Dec 2018 11:01:56 -0800 (PST) In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Tue, 11 Dec 2018 23:35:31 -0800, Michael Chan wrote: > On Tue, Dec 11, 2018 at 10:18 AM Jakub Kicinski wrote: > > On Tue, 11 Dec 2018 14:16:47 +0530, Vasundhara Volam wrote: > > > wake-on-lan - Enables Wake on Lan for this port. If enabled, > > > the controller asserts a wake pin based on the wake-on-lan type. > > > > > > Cc: Jiri Pirko > > > Signed-off-by: Vasundhara Volam > > > > As explained previously I think it's a very bad idea to add existing > > configuration options to devlink, just because devlink has the ability > > to persist the setting in NVM. Especially that for WoL you have to get > > the link up so you potentially have all link config stuff as well. And > > that n-tuple filters are one of the WoL options, meaning we'd need the > > ability to persist n-tuple filters via devlink. > > As I said before, firmware will automatically set the link to autoneg > up to the speed supported by Vaux if WoL is enabled. No special link > setting is needed as I said before. I don't think n-tuple is suitable > for the default power-up WoL setting. n-tuple requires ip address. > The ip address belongs to the system, not to the card that can move > from system to system. The n-tuple WoL packet should be a transient > setting set by the OS that won't persist a power down. The default > power-up WoL packet types should be the most basic magic packet and > other basic types. So I really don't think there is a need to persist > n-tuple WoL packet types. Sure your firmware sets the link to autoneg today, and your driver does not allow n-tuple filter WoL. How about we step back and think about the API as a whole rather than exposing ad hoc FW knobs via delink params :/ Also what gives you the idea that n-tuple filters require an ip address?