From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:35305) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1c57Bq-0001mJ-Es for qemu-devel@nongnu.org; Fri, 11 Nov 2016 03:28:03 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1c57Bn-0001Mn-9T for qemu-devel@nongnu.org; Fri, 11 Nov 2016 03:28:02 -0500 Received: from mga02.intel.com ([134.134.136.20]:45878) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1c57Bn-0001MT-0E for qemu-devel@nongnu.org; Fri, 11 Nov 2016 03:27:59 -0500 Message-ID: <58258125.3060304@intel.com> Date: Fri, 11 Nov 2016 16:28:21 +0800 From: Wei Wang MIME-Version: 1.0 References: <1478746069-79574-1-git-send-email-wei.w.wang@intel.com> <1478746069-79574-3-git-send-email-wei.w.wang@intel.com> In-Reply-To: Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 8bit Subject: Re: [Qemu-devel] [RFC PATCH 2/4] vhost-user: add the vhost-user extension to support the vhost-pci based inter-vm communication List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: =?UTF-8?B?TWFyYy1BbmRyw6kgTHVyZWF1?= Cc: mst@redhat.com, stefanha@redhat.com, pbonzini@redhat.com, qemu-devel@nongnu.org, virtio-dev@lists.oasis-open.org On 11/10/2016 07:36 PM, Marc-André Lureau wrote: > Hi > > On Thu, Nov 10, 2016 at 6:47 AM Wei Wang > wrote: > > This is the slave part of vhost-user implemented in QEMU, with an > extension > to support vhost-pci. > > > Instead of implementing "another vhost-user slave", it would be worth > investigating using libvhost-user instead > (https://lists.gnu.org/archive/html/qemu-devel/2016-10/msg03990.html). > This is just a suggestion, it is quite fine for vhost-pci to have its > own smaller/specific vhost-user slave implementation (without virtio > rings handlings etc). (libvhost-user is also very young, not yet in > qemu, we should be able shape it for vhost-pci needs) Thanks for the suggestion. It looks possible. I will need to investigate more about it. libvhost-user is still at an initial stage and not in QEMU yet, I think we can probably have vhost-pci as an independent slave for now - this will not complicate libvhost-user when it's in the initial stage (it might not be good to make it a quite complex design at the beginning), and on the other side, the vhost-pci based slave implementation draft is already there. We can merge them if it is necessary when they both are stable in QEMU. What do you think? Best, Wei