From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jason Wang Subject: Re: [PATCH kvm-unit-test 6/6] Add a test for kvm-clock Date: Sun, 29 Aug 2010 23:07:48 -0400 (EDT) Message-ID: <340473099.825401283137668728.JavaMail.root@zmail05.collab.prod.int.phx2.redhat.com> References: <20100827112702.GK2985@mothafucka.localdomain> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit Cc: mtosatti@redhat.com, avi@redhat.com, kvm@vger.kernel.org To: Glauber Costa Return-path: Received: from mx4-phx2.redhat.com ([209.132.183.25]:50006 "EHLO mx02.colomx.prod.int.phx2.redhat.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1754390Ab0H3DHt (ORCPT ); Sun, 29 Aug 2010 23:07:49 -0400 In-Reply-To: <20100827112702.GK2985@mothafucka.localdomain> Sender: kvm-owner@vger.kernel.org List-ID: ----- "Glauber Costa" wrote: > On Fri, Aug 27, 2010 at 01:49:53PM +0800, Jason Wang wrote: > > This patch implements two tests for kvmclock. First one check > whether > > the date of time returned by kvmclock matches the value got from > > host. Second one check whether the cycle of kvmclock grows > > monotonically in smp guest. > > > > Three parameters were accepted by the test: test loops, seconds > > since 1970-01-01 00:00:00 UTC which could be easily get through > date > > +%s and the max accepted offset value between the tod of guest and > > host. > Good. > > I liked the flag usage. Might help us in the future, when we apply > zach's series plus a couple of ideas we have, to see if it the > problem indeed goes away. > > A minor tip, not strong feelings towards this, would be to extract > some information from host cpu, and print it too. > It would be easier when we're analyzing this reports in the future. > > tsc-based mechanisms are very sensible to: > - vendor > - # of cpus > - # of sockets > - tsc flags > > Sure we can get all this information from /proc/cpuinfo, but having > it > in your final report automatically would be convenient, I think. Yes, they are useful. But since unit tests are running as guest, it would be a hard to gather host information by it self. So maybe we could do this through autotest or other kind of test launcher.