From mboxrd@z Thu Jan  1 00:00:00 1970
From: Stephen Hemminger <stephen@networkplumber.org>
Subject: Re: [PATCH 4/5] net/netvsc: implement link state change
 callback
Date: Fri, 31 Aug 2018 08:13:23 -0700
Message-ID: <20180831081323.3925e3ad@xeon-e3>
References: <20180830223512.21297-1-stephen@networkplumber.org>
 <20180830223512.21297-5-stephen@networkplumber.org>
 <20180831082547.s3xm3tb7keruam6m@bidouze.vm.6wind.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Cc: dev@dpdk.org, Stephen Hemminger <sthemmin@microsoft.com>
To: =?UTF-8?B?R2HDq3Rhbg==?= Rivet <gaetan.rivet@6wind.com>
Return-path: <dev-bounces@dpdk.org>
Received: from mail-pg1-f193.google.com (mail-pg1-f193.google.com
 [209.85.215.193]) by dpdk.org (Postfix) with ESMTP id 549041AEF0
 for <dev@dpdk.org>; Fri, 31 Aug 2018 17:13:26 +0200 (CEST)
Received: by mail-pg1-f193.google.com with SMTP id 7-v6so642985pgf.2
 for <dev@dpdk.org>; Fri, 31 Aug 2018 08:13:26 -0700 (PDT)
In-Reply-To: <20180831082547.s3xm3tb7keruam6m@bidouze.vm.6wind.com>
List-Id: DPDK patches and discussions <dev.dpdk.org>
List-Unsubscribe: <https://mails.dpdk.org/options/dev>,
 <mailto:dev-request@dpdk.org?subject=unsubscribe>
List-Archive: <http://mails.dpdk.org/archives/dev/>
List-Post: <mailto:dev@dpdk.org>
List-Help: <mailto:dev-request@dpdk.org?subject=help>
List-Subscribe: <https://mails.dpdk.org/listinfo/dev>,
 <mailto:dev-request@dpdk.org?subject=subscribe>
Errors-To: dev-bounces@dpdk.org
Sender: "dev" <dev-bounces@dpdk.org>

On Fri, 31 Aug 2018 10:25:47 +0200
Ga=C3=ABtan Rivet <gaetan.rivet@6wind.com> wrote:

> Hi Stephen,
>=20
> On Thu, Aug 30, 2018 at 03:35:11PM -0700, Stephen Hemminger wrote:
> > From: Stephen Hemminger <sthemmin@microsoft.com>
> >=20
> > Implement callback functionality on link state changes.
> > This is not really driven off of interrupt file descriptor like most ot=
her
> > PMD's. Instead, it happens when a link state change message arrives
> > in the common ring buffer.
> >  =20
>=20
> Does this mean that the lsc event will be processed in a dataplane
> thread? Looking at the _rte_eth_dev_callback_process() call, it seems
> so.
>=20
> Shouldn't this be executed in the context of the eal-intr-thread
> instead? This thread is marked control and should be configured with the
> proper afinity, unless dataplane threads.
>=20
> Maybe I missed something, it's just to double-check that this will
> behave nicely with applications relying on the eal-intr-thread afinity.
>=20
> --=20
> Ga=C3=ABtan Rivet
> 6WIND

There is no EAL API to take the event and propogate it over to the eal inte=
rrupt
thread. The interrupt thread is buried in the internals of EAL.