From mboxrd@z Thu Jan  1 00:00:00 1970
From: "Michael S. Tsirkin" <mst@redhat.com>
Subject: Re: [PATCH net-next] virtio_net: force_napi_tx module param.
Date: Thu, 2 Aug 2018 01:48:42 +0300
Message-ID: <20180802014755-mutt-send-email-mst@kernel.org>
References: <20180723231119.142904-1-caleb.raitto@gmail.com>
 <20180729.090027.1373538625446665385.davem@davemloft.net>
 <CAF=yD-+XnL=TowYwdaww7hBkPQek4_cDMSiqo02UaJhKy5PrTA@mail.gmail.com>
 <20180731153204-mutt-send-email-mst@kernel.org>
 <CAF=yD-JVj9SLNBftB=ut2-duhi4HQdLBByOHJaZDdavLabD+OA@mail.gmail.com>
 <20180802012405-mutt-send-email-mst@kernel.org>
 <CAF=yD-+Wvy_93h2=OV=hkrj7SRoR_NMpOyc8qMStUy74+6-t+Q@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: David Miller <davem@davemloft.net>, caleb.raitto@gmail.com,
        Jason Wang <jasowang@redhat.com>,
        Network Development <netdev@vger.kernel.org>,
        Caleb Raitto <caraitto@google.com>
To: Willem de Bruijn <willemdebruijn.kernel@gmail.com>
Return-path: <netdev-owner@vger.kernel.org>
Received: from mx3-rdu2.redhat.com ([66.187.233.73]:33628 "EHLO mx1.redhat.com"
        rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP
        id S1726688AbeHBAgt (ORCPT <rfc822;netdev@vger.kernel.org>);
        Wed, 1 Aug 2018 20:36:49 -0400
Content-Disposition: inline
In-Reply-To: <CAF=yD-+Wvy_93h2=OV=hkrj7SRoR_NMpOyc8qMStUy74+6-t+Q@mail.gmail.com>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

On Wed, Aug 01, 2018 at 06:43:53PM -0400, Willem de Bruijn wrote:
>  > > So e.g. we could set an affinity hint to a group of CPUs that
> > > > might transmit to this queue.
> > >
> > > We also want to set the xps mask for all cpus in the group to this queue.
> > >
> > > Is there a benefit over explicitly choosing one cpu from the set, btw?
> >
> > If only one CPU actually transmits on this queue then probably yes.
> > And virtio doesn't know whether that's the case.
> 
> Oh, yes, good example.
> 
> To plumb that through, vp_set_vq_affinity would have to take a cpu
> mask like irq_set_affinity_hint, instead of a single value.
> 
> Though I don't suggest doing that as part of the core
> virtnet_set_affinity changes.

Why not? It seems a reasonable thing to do.

-- 
MST