From mboxrd@z Thu Jan  1 00:00:00 1970
From: Stephen Hemminger <stephen@networkplumber.org>
Subject: Re: [RFC PATCH] virtio_net: Extend virtio to use VF datapath when
 available
Date: Tue, 19 Dec 2017 10:13:45 -0800
Message-ID: <20171219101345.7bf43a8d@xeon-e3>
References: <1513644036-45230-1-git-send-email-sridhar.samudrala@intel.com>
        <20171219174259-mutt-send-email-mst@kernel.org>
        <ca709329-a47a-2e76-b1fc-38fc665d7c0b@intel.com>
        <20171219095548.61d70d10@xeon-e3>
        <20171219200611-mutt-send-email-mst@kernel.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Cc: "Samudrala, Sridhar" <sridhar.samudrala@intel.com>,
        netdev@vger.kernel.org, virtualization@lists.linux-foundation.org,
        alexander.duyck@gmail.com,
        "Brandeburg, Jesse" <jesse.brandeburg@intel.com>
To: "Michael S. Tsirkin" <mst@redhat.com>
Return-path: <netdev-owner@vger.kernel.org>
Received: from mail-pf0-f195.google.com ([209.85.192.195]:46972 "EHLO
        mail-pf0-f195.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S1750918AbdLSSNt (ORCPT
        <rfc822;netdev@vger.kernel.org>); Tue, 19 Dec 2017 13:13:49 -0500
Received: by mail-pf0-f195.google.com with SMTP id c204so11445824pfc.13
        for <netdev@vger.kernel.org>; Tue, 19 Dec 2017 10:13:48 -0800 (PST)
In-Reply-To: <20171219200611-mutt-send-email-mst@kernel.org>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

On Tue, 19 Dec 2017 20:07:01 +0200
"Michael S. Tsirkin" <mst@redhat.com> wrote:

> On Tue, Dec 19, 2017 at 09:55:48AM -0800, Stephen Hemminger wrote:
> > On Tue, 19 Dec 2017 09:41:39 -0800
> > "Samudrala, Sridhar" <sridhar.samudrala@intel.com> wrote:
> >  =20
> > > On 12/19/2017 7:47 AM, Michael S. Tsirkin wrote: =20
> > > > I'll need to look at this more, in particular the feature
> > > > bit is missing here. For now one question:
> > > >
> > > > On Mon, Dec 18, 2017 at 04:40:36PM -0800, Sridhar Samudrala wrote: =
  =20
> > > >> @@ -56,6 +58,8 @@ module_param(napi_tx, bool, 0644);
> > > >>    */
> > > >>   DECLARE_EWMA(pkt_len, 0, 64)
> > > >>  =20
> > > >> +#define VF_TAKEOVER_INT	(HZ / 10)
> > > >> +
> > > >>   #define VIRTNET_DRIVER_VERSION "1.0.0"
> > > >>  =20
> > > >>   static const unsigned long guest_offloads[] =3D {   =20
> > > > Why is this delay necessary? And why by 100ms?   =20
> > >=20
> > > This is based on netvsc implementation and here is the commit that
> > > added this delay.=C2=A0 Not sure if this needs to be 100ms.
> > >=20
> > > commit 6123c66854c174e4982f98195100c1d990f9e5e6
> > > Author: stephen hemminger <stephen@networkplumber.org>
> > > Date:=C2=A0=C2=A0 Wed Aug 9 17:46:03 2017 -0700
> > >=20
> > >  =C2=A0=C2=A0=C2=A0 netvsc: delay setup of VF device
> > >=20
> > >  =C2=A0=C2=A0=C2=A0 When VF device is discovered, delay bring it auto=
matically up in
> > >  =C2=A0=C2=A0=C2=A0 order to allow userspace to some simple changes (=
like renaming).
> > >=20
> > >=20
> > >  =20
> >=20
> > could be 10ms, just enough to let udev do its renaming =20
>=20
> Isn't there a way not to depend on udev completing its thing within a giv=
en timeframe?

Not that I know. the path is quite indirect.