From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ingo Molnar Subject: Re: [PATCH] raise tsc clocksource rating Date: Mon, 29 Oct 2007 23:55:28 +0100 Message-ID: <20071029225528.GA31293@elte.hu> References: <11936994092607-git-send-email-gcosta@redhat.com> <1193697734.9793.86.camel@bodhitayantram.eng.vmware.com> <20071029224852.GA27547@elte.hu> <47266431.1010500@goop.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Zachary Amsden , Glauber de Oliveira Costa , linux-kernel@vger.kernel.org, tglx@linutronix.de, rusty@rustcorp.com.au, --cc@redhat.com, avi@quramnet.com, kvm-devel@lists.sourceforge.net, Glauber de Oliveira Costa , Dan Hecht , Garrett Smith To: Jeremy Fitzhardinge Return-path: Content-Disposition: inline In-Reply-To: <47266431.1010500@goop.org> Sender: linux-kernel-owner@vger.kernel.org List-Id: kvm.vger.kernel.org * Jeremy Fitzhardinge wrote: > > if it's inaccurate why are you exposing it to the guest then? Native > > only uses the TSC if it's safe and accurate to do so. > > It is used as part of the Xen clocksource as a short term > extrapolator, with correction parameters supplied by the hypervisor. > It should never be used directly. that's totally broken then. You cannot create an SMP-safe monotonic clocksource via interpolation - native does not do it either. Good thing this problem got exposed, it needs to be fixed. Ingo