From mboxrd@z Thu Jan 1 00:00:00 1970 From: John Fastabend Subject: Re: [PATCH net-next 3/3] net/mlx4_en: Enable open-lldp DCB support Date: Fri, 05 Apr 2013 00:52:38 -0700 Message-ID: <515E82C6.9080406@gmail.com> References: <1365085574-12057-1-git-send-email-ogerlitz@mellanox.com> <1365085574-12057-4-git-send-email-ogerlitz@mellanox.com> <515DA342.5020503@intel.com> <515DA579.3050006@mellanox.com> <515DCDF2.8070304@intel.com> <515E129D.2020009@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: John Fastabend , Sagi Grimberg , Or Gerlitz , davem@davemloft.net, netdev@vger.kernel.org, amirv@mellanox.com To: Or Gerlitz Return-path: Received: from mail-ob0-f170.google.com ([209.85.214.170]:58142 "EHLO mail-ob0-f170.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932563Ab3DEHyh (ORCPT ); Fri, 5 Apr 2013 03:54:37 -0400 Received: by mail-ob0-f170.google.com with SMTP id uy19so1266682obc.15 for ; Fri, 05 Apr 2013 00:54:37 -0700 (PDT) In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On 04/05/2013 12:15 AM, Or Gerlitz wrote: > John Fastabend wrote: > >> Agree it doesn't work without this patch but we can fix it in user >> space. This has the nice benefit of letting lldpad work with mlx4 on >> older kernels. In general I think its easier to update user space then >> the kernel for most users. I'll send you a lldpad patch in a few moments. > > So to be sure we're on the same page, what we should do is enhance > mlx4_en_dcbnl_getdcbx() routine to also return DCB_CAP_DCBX_HOST and Right I think this is a better approach in the long run. > work with your user space patch? sounds nice, except for what happens > if people run with older versions of lldpad.. > Well it won't work. But on the flip side all your older drivers _will_ work with the new daemon. My working assumption is for most users its easier to upgrade a user space pkg than a kernel. And the simpler we can keep the kernel code the better. Oh and also if there is a bug in user space we shouldn't fix it with kernel code. > Or. > -- John Fastabend Intel Corporation