From mboxrd@z Thu Jan 1 00:00:00 1970 From: Paolo Bonzini Subject: Re: [PATCH v3 0/7] RTC: New logic to emulate RTC Date: Wed, 14 Mar 2012 09:54:16 +0100 Message-ID: <4F605CB8.20200@redhat.com> References: <4F60581A.7000800@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: "qemu-devel@nongnu.org" , Jan Kiszka , "kvm@vger.kernel.org" , "aliguori@us.ibm.com" , Marcelo Tosatti To: "Zhang, Yang Z" Return-path: Received: from mx1.redhat.com ([209.132.183.28]:46986 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1759332Ab2CNIy1 (ORCPT ); Wed, 14 Mar 2012 04:54:27 -0400 In-Reply-To: Sender: kvm-owner@vger.kernel.org List-ID: Il 14/03/2012 09:52, Zhang, Yang Z ha scritto: >>>>> Is there any comments with the version 3? >>> >>> Can you explain why you dropped the logic to set the timer to the >>> next event? > Do you mean why I change the rtc logic? The reason is that: When a > guest is idle, the main activity inside qemu is the rtc update > timer(2 per second). In our experience(running 64 rhel6u1 guests), it > will decrease pkg C6 residency about 6%(6% means 2 watts in my box). > And normally, the guest will not use the update-end interrupt and > alarm. So there has no need to run a periodic timer when guest isn't > using it. No, why you're keeping roughly the same logic as current QEMU, instead of the more radical changes that were in v2. Paolo