From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:46683) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1WiJwK-0001wU-Qo for qemu-devel@nongnu.org; Thu, 08 May 2014 04:44:33 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1WiJwG-0003fX-EU for qemu-devel@nongnu.org; Thu, 08 May 2014 04:44:28 -0400 Received: from mx1.redhat.com ([209.132.183.28]:17203) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1WiJwG-0003fN-6a for qemu-devel@nongnu.org; Thu, 08 May 2014 04:44:24 -0400 Message-ID: <536B43E0.4050208@redhat.com> Date: Thu, 08 May 2014 10:44:16 +0200 From: Paolo Bonzini MIME-Version: 1.0 References: <20140506231350.GA3238@gmail.com> <20140507080808.GD27925@stefanha-thinkpad.muc.redhat.com> <20140508072422.GA10416@gmail.com> In-Reply-To: <20140508072422.GA10416@gmail.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] Configure virtio-scsi options via libvirt List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Mike Perez , Stefan Hajnoczi Cc: libvir-list@redhat.com, nab@daterainc.com, qemu-devel@nongnu.org Il 08/05/2014 09:24, Mike Perez ha scritto: > On 10:08 Wed 07 May , Stefan Hajnoczi wrote: >> On Tue, May 06, 2014 at 04:13:50PM -0700, Mike Perez wrote: >>> I would like be able to configure virtio-scsi options num_queues, max_sectors, >>> and cmd_per_lun via libvirt. Are there any plans to have this support? >> >> Hi Mike, >> I'm not sure about the status of libvirt support for virtio-scsi options >> but in the meantime you can use passthrough: >> >> http://blog.vmsplice.net/2011/04/how-to-pass-qemu-command-line-options.html >> >> Stefan > > Thanks Stefan, > > My plan is to write the support in OpenStack, which unfortunately does not > accept patches doing qemu passthrough: > > http://lists.openstack.org/pipermail/openstack-dev/2013-November/018551.html Why do you need such fine-grained control? The higher you go in the stack, the simpler configuration should be in my opinion. For openstack, I think all you need is a singlequeue vs. multiqueue knob (setting num_queues = # VCPUs for multiqueue). Paolo