From mboxrd@z Thu Jan  1 00:00:00 1970
From: Yuanhan Liu <yliu@fridaylinux.org>
Subject: Re: [RFC v2] vhost: new rte_vhost API proposal
Date: Sun, 20 May 2018 15:07:25 +0800
Message-ID: <20180520070725.4ccoumdtpq3utcqr@yuanhanliu-NB0.tencent.com>
References: <1525958573-184361-1-git-send-email-dariuszx.stojaczyk@intel.com>
 <1526648465-62579-1-git-send-email-dariuszx.stojaczyk@intel.com>
 <27ce772e-9f01-dff9-1f82-b99924efa950@redhat.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>, dev@dpdk.org,
 Tiwei Bie <tiwei.bie@intel.com>, Tetsuya Mukawa <mtetsuyah@gmail.com>,
 Thomas Monjalon <thomas@monjalon.net>,
 Stefan Hajnoczi <stefanha@redhat.com>,
 James Harris <james.r.harris@intel.com>
To: Maxime Coquelin <maxime.coquelin@redhat.com>
Return-path: <dev-bounces@dpdk.org>
Received: from out1-smtp.messagingengine.com (out1-smtp.messagingengine.com
 [66.111.4.25]) by dpdk.org (Postfix) with ESMTP id 4084449E0
 for <dev@dpdk.org>; Sun, 20 May 2018 12:15:57 +0200 (CEST)
Content-Disposition: inline
In-Reply-To: <27ce772e-9f01-dff9-1f82-b99924efa950@redhat.com>
List-Id: DPDK patches and discussions <dev.dpdk.org>
List-Unsubscribe: <https://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: <https://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 Fri, May 18, 2018 at 03:50:45PM +0200, Maxime Coquelin wrote:
> Hi Dariusz,
> 
> On 05/18/2018 03:01 PM, Dariusz Stojaczyk wrote:
> > rte_vhost is not vhost-user spec compliant. Some Vhost drivers have
> > been already confirmed not to work with rte_vhost. virtio-user-scsi-pci
> > in QEMU 2.12 doesn't fully initialize its management queues at SeaBIOS
> > stage. This is perfectly fine from the Vhost-user spec perspective, but
> > doesn't meet rte_vhost expectations. rte_vhost waits for all queues
> > to be fully initialized before it allows the entire device to be
> > processed. qFixing rte_vhost directly would require quite a big amount
> > of changes, which would completely break backwards compatibility.
> > 
> > This rte_vhost2 library is intended to smooth out the transition.
> > It exposes a low-level API for implementing new Vhost-user slaves.
> > The existing rte_vhost is about to be refactored to use rte_vhost2
> > library underneath, and demanding backends could now use rte_vhost2
> > directly.
> 
> I like the idea, and the proposed way to smooth the transition.

Please be aware of that I just had a quick glimpse of this patch and it's
likely I don't have too much time to follow this.

However, I also like this idea. And thank you for working on it.

	--yliu