From mboxrd@z Thu Jan 1 00:00:00 1970 From: Rusty Russell Subject: Re: virtio-blk performance regression and qemu-kvm Date: Mon, 13 Feb 2012 10:25:59 +1030 Message-ID: <87sjifd2sw.fsf@rustcorp.com.au> References: <20120210143639.GA17883@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: kvm@vger.kernel.org To: Dongsu Park , qemu-devel@nongnu.org Return-path: Received: from ozlabs.org ([203.10.76.45]:58045 "EHLO ozlabs.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756007Ab2BLX7l (ORCPT ); Sun, 12 Feb 2012 18:59:41 -0500 In-Reply-To: <20120210143639.GA17883@gmail.com> Sender: kvm-owner@vger.kernel.org List-ID: On Fri, 10 Feb 2012 15:36:39 +0100, Dongsu Park wrote: > Hi, > > Recently I observed performance regression regarding virtio-blk, > especially different IO bandwidths between qemu-kvm 0.14.1 and 1.0. > So I want to share the benchmark results, and ask you what the reason > would be. Interesting. There are two obvious possibilities here. One is that qemu has regressed, the other is that virtio_blk has regressed; the new qemu may negotiate new features. Please do the following in the guest with old and new qemus: cat /sys/class/block/vdb/device/features (eg, here that gives: 0010101101100000000000000000100e0). Thanks, Rusty. From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([140.186.70.92]:53309) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1RwjKb-0007Za-GW for qemu-devel@nongnu.org; Sun, 12 Feb 2012 18:59:46 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1RwjKa-0005Is-7E for qemu-devel@nongnu.org; Sun, 12 Feb 2012 18:59:45 -0500 Received: from ozlabs.org ([203.10.76.45]:37484) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1RwjKZ-0005Ib-Tj for qemu-devel@nongnu.org; Sun, 12 Feb 2012 18:59:44 -0500 From: Rusty Russell In-Reply-To: <20120210143639.GA17883@gmail.com> References: <20120210143639.GA17883@gmail.com> Date: Mon, 13 Feb 2012 10:25:59 +1030 Message-ID: <87sjifd2sw.fsf@rustcorp.com.au> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Subject: Re: [Qemu-devel] virtio-blk performance regression and qemu-kvm List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Dongsu Park , qemu-devel@nongnu.org Cc: kvm@vger.kernel.org On Fri, 10 Feb 2012 15:36:39 +0100, Dongsu Park wrote: > Hi, > > Recently I observed performance regression regarding virtio-blk, > especially different IO bandwidths between qemu-kvm 0.14.1 and 1.0. > So I want to share the benchmark results, and ask you what the reason > would be. Interesting. There are two obvious possibilities here. One is that qemu has regressed, the other is that virtio_blk has regressed; the new qemu may negotiate new features. Please do the following in the guest with old and new qemus: cat /sys/class/block/vdb/device/features (eg, here that gives: 0010101101100000000000000000100e0). Thanks, Rusty.