From mboxrd@z Thu Jan 1 00:00:00 1970 From: Richard Cochran Subject: Re: [patch 4/5] PTP: add PTP_SYS_OFFSET emulation via cross timestamps infrastructure Date: Tue, 24 Jan 2017 06:43:03 +0100 Message-ID: <20170124054302.GB1906@localhost.localdomain> References: <20170120122025.665985919@redhat.com> <20170120122503.746158230@redhat.com> <20170120202502.GA10368@localhost.localdomain> <20170123131913.GA28104@amt.cnet> <20170123184415.GA2158@localhost.localdomain> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Marcelo Tosatti , kvm@vger.kernel.org, linux-kernel@vger.kernel.org, Radim Krcmar , Miroslav Lichvar To: Paolo Bonzini Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org List-Id: kvm.vger.kernel.org On Mon, Jan 23, 2017 at 08:44:53PM +0100, Paolo Bonzini wrote: > If you just implement getclock64 the PTP_SYS_OFFSET output: > > device clock | |sample2| |sample4| |sample6| ... > ------------------------------------------------------------- > realtime clock |sample1| |sample3| |sample5| > > has a very large distance between samples on the same line (about 1 us), > and I think it is too noisy for userspace to make sense of the output. One microsecond is not too bad at all. The PCIe devices have intervals of 5-6 usec, and the result is quite usable, certainly better than generic NTP. > Marcelo's patch then produces fake realtime clock samples that, however, > let chrony derive the cross timestamps that KVM produced in the first > place. The outcome is really great accuracy compared to previous > versions of the patch, often just +/- 2 or 3 nanoseconds. Why can't chrony learn about PTP_SYS_OFFSET_PRECISE? Thanks, Richard