From mboxrd@z Thu Jan 1 00:00:00 1970 From: Yuanhan Liu Subject: Re: [PATCH] vhost: remove internal lockless enqueue Date: Tue, 14 Jun 2016 22:07:34 +0800 Message-ID: <20160614140734.GC10038@yliu-dev.sh.intel.com> References: <1465818732-116302-1-git-send-email-huawei.xie@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: dev@dpdk.org, yuanhan.liu@intel.com, jianfeng.tan@intel.com, mukawa@igel.co.jp, kevin.traynor@intel.com, haifeng.lin@huawei.com To: Huawei Xie Return-path: Received: from mga09.intel.com (mga09.intel.com [134.134.136.24]) by dpdk.org (Postfix) with ESMTP id AA5356A1C for ; Tue, 14 Jun 2016 16:06:33 +0200 (CEST) Content-Disposition: inline In-Reply-To: <1465818732-116302-1-git-send-email-huawei.xie@intel.com> List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" Firstly, it's V2. So, don't forget to add version number and link it to the previous version next time. On Mon, Jun 13, 2016 at 07:52:12PM +0800, Huawei Xie wrote: > All other DPDK PMDs doesn't support concurrent receiving or sending > packets to the same queue. The upper application should deal with > this, normally through queue and core bindings. > > Due to historical reason, vhost internally supports concurrent lockless > enqueuing packets to the same virtio queue through costly cmpset operation. > This patch removes this internal lockless implementation and should improve > performance a bit. > > Luckily DPDK OVS doesn't rely on this behavior. > > Signed-off-by: Huawei Xie Applied to dpdk-next-virtio, with the rebase on top of my vhost ABI/API changes. FYI, I also renamed the title a bit to "remove concurrent enqueue" as Thomas mentioned in the last version, that it's confusing to say "remove lockless" here, since we are actually removing the lock. Thanks. --yliu