From mboxrd@z Thu Jan  1 00:00:00 1970
From: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Subject: Re: [PATCH] vhost: change mbuf allocation logs to debug
Date: Wed, 29 Mar 2017 15:01:50 +0800
Message-ID: <20170329070150.GF18844@yliu-dev.sh.intel.com>
References: <20170327162930.54281-1-allain.legacy@windriver.com>
 <20170328064841.GZ18844@yliu-dev.sh.intel.com>
 <70A7408C6E1BFB41B192A929744D8523968F8CC8@ALA-MBC.corp.ad.wrs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: "mtetsuyah@gmail.com" <mtetsuyah@gmail.com>,
 "maxime.coquelin@redhat.com" <maxime.coquelin@redhat.com>,
 "dev@dpdk.org" <dev@dpdk.org>, "Peters, Matt" <Matt.Peters@windriver.com>,
 Thomas Monjalon <thomas.monjalon@6wind.com>
To: "Legacy, Allain" <Allain.Legacy@windriver.com>
Return-path: <dev-bounces@dpdk.org>
Received: from mga02.intel.com (mga02.intel.com [134.134.136.20])
 by dpdk.org (Postfix) with ESMTP id 4E7F11396
 for <dev@dpdk.org>; Wed, 29 Mar 2017 09:04:07 +0200 (CEST)
Content-Disposition: inline
In-Reply-To: <70A7408C6E1BFB41B192A929744D8523968F8CC8@ALA-MBC.corp.ad.wrs.com>
List-Id: DPDK patches and discussions <dev.dpdk.org>
List-Unsubscribe: <http://dpdk.org/ml/options/dev>,
 <mailto:dev-request@dpdk.org?subject=unsubscribe>
List-Archive: <http://dpdk.org/ml/archives/dev/>
List-Post: <mailto:dev@dpdk.org>
List-Help: <mailto:dev-request@dpdk.org?subject=help>
List-Subscribe: <http://dpdk.org/ml/listinfo/dev>,
 <mailto:dev-request@dpdk.org?subject=subscribe>
Errors-To: dev-bounces@dpdk.org
Sender: "dev" <dev-bounces@dpdk.org>

On Tue, Mar 28, 2017 at 11:39:37AM +0000, Legacy, Allain wrote:
> > -----Original Message-----
> > From: Yuanhan Liu [mailto:yuanhan.liu@linux.intel.com]
> > Sent: Tuesday, March 28, 2017 2:49 AM
> <...>
> > > In order to prevent this condition, but
> > > still enable debugging, the logs are being changed to debug logs to ensure
> > > they are not emitted unless the CONFIG_RTE_LIBRTE_VHOST_DEBUG
> > option is
> > > enabled.
> > 
> > That basically means you will never see them in real life. I'm suggesting
> > to add a new log API, RTE_LOG_ONCE. Sounds good?
> 
> 
> Rather than a log what is really needed here is a stat/counter, but this library does not have the infrastructure in place to support reporting datapath errors via stats back to the application.

Right. It's supposed to be done in vhost-pmd, while the real Rx implementation
is at vhost lib.

>  Are there any plans to add a stats infrastructure or a desired design for what that would look like?

We could have such counter support once we move virtio_net.c to vhost-pmd.
Unluckily, I don't think it will happen soon.

OTOH, it's still good to have RTE_LOG_ONCE. It explicitly tells you something
is wrong, then you could dump those stats for more info.

	--yliu