From mboxrd@z Thu Jan 1 00:00:00 1970 From: Anthony Liguori Subject: Re: [Qemu-devel] Re: [PATCH] virtio: Use ioeventfd for virtqueue notify Date: Mon, 04 Oct 2010 09:01:14 -0500 Message-ID: <4CA9DE2A.5080309@linux.vnet.ibm.com> References: <1285855312-11739-1-git-send-email-stefanha@linux.vnet.ibm.com> <4CA862A7.2080302@redhat.com> <20101003135138.GA19775@redhat.com> <4CA89185.6070302@redhat.com> <20101003142828.GB19775@redhat.com> <4CA92B79.2030805@linux.vnet.ibm.com> <4CA98A8D.4000500@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: "Michael S. Tsirkin" , Steve Dobbelstein , Stefan Hajnoczi , kvm@vger.kernel.org, qemu-devel@nongnu.org, Khoa Huynh , sri@linux.vnet.ibm.com To: Avi Kivity Return-path: Received: from e8.ny.us.ibm.com ([32.97.182.138]:57402 "EHLO e8.ny.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755052Ab0JDOBr (ORCPT ); Mon, 4 Oct 2010 10:01:47 -0400 Received: from d01relay03.pok.ibm.com (d01relay03.pok.ibm.com [9.56.227.235]) by e8.ny.us.ibm.com (8.14.4/8.13.1) with ESMTP id o94DgVLv031906 for ; Mon, 4 Oct 2010 09:42:31 -0400 Received: from d01av03.pok.ibm.com (d01av03.pok.ibm.com [9.56.224.217]) by d01relay03.pok.ibm.com (8.13.8/8.13.8/NCO v10.0) with ESMTP id o94E1khb279782 for ; Mon, 4 Oct 2010 10:01:46 -0400 Received: from d01av03.pok.ibm.com (loopback [127.0.0.1]) by d01av03.pok.ibm.com (8.14.4/8.13.1/NCO v10.0 AVout) with ESMTP id o94E1iUJ004990 for ; Mon, 4 Oct 2010 11:01:46 -0300 In-Reply-To: <4CA98A8D.4000500@redhat.com> Sender: kvm-owner@vger.kernel.org List-ID: On 10/04/2010 03:04 AM, Avi Kivity wrote: > On 10/04/2010 03:18 AM, Anthony Liguori wrote: >> On 10/03/2010 09:28 AM, Michael S. Tsirkin wrote: >>> >>>> This is using eventfd as well. >>> Sorry, I meant irqfd. >> >> I've tried using irqfd in userspace. It hurts performance quite a >> bit compared to doing an ioctl so I would suspect this too. >> >> A last_used_idx or similar mechanism should help performance quite a >> bit on top of ioeventfd too. >> > > Any idea why? While irqfd does quite a bit of extra locking, it > shouldn't be that bad. Not really. It was somewhat counter intuitive. A worthwhile experiment might be to do some layering violations and have vhost do an irq injection via an ioctl and see what the performance delta is. I suspect it could give vhost a nice boost. Regards, Anthony Liguori