From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter Zijlstra Subject: Re: [PATCH v3 1/4] Add correlated clocksource deriving system time from an auxiliary clocksource Date: Fri, 4 Sep 2015 17:10:21 +0200 Message-ID: <20150904151021.GF18489@twins.programming.kicks-ass.net> References: <1440183128-1384-1-git-send-email-christopher.s.hall@intel.com> <1440183128-1384-2-git-send-email-christopher.s.hall@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: "Hall, Christopher S" , "Kirsher, Jeffrey T" , "hpa@zytor.com" , "mingo@redhat.com" , "john.stultz@linaro.org" , "richardcochran@gmail.com" , "x86@kernel.org" , "linux-kernel@vger.kernel.org" , "netdev@vger.kernel.org" , "intel-wired-lan@lists.osuosl.org" To: Thomas Gleixner Return-path: Received: from casper.infradead.org ([85.118.1.10]:35769 "EHLO casper.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751331AbbIDPKa (ORCPT ); Fri, 4 Sep 2015 11:10:30 -0400 Content-Disposition: inline In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Fri, Sep 04, 2015 at 03:02:19PM +0200, Thomas Gleixner wrote: > > For example, supply the ART value as an argument and, in the case of > > the realtime clock, keep a short history of clock changes. It would > > It's not only clock realtime which is affected by those. > > > fail in cases where there are a lot of calls to adjtimex(), > > That has nothing to do with lots of adjtimex calls. The kernel does a > slow correction of the conversion values itself to avoid time jumping > around. I think what they're getting at is asking if there's a rate limit to time adjustments, without that, saving the last n transition points will still not cover any given length of history. So what I think they're looking for; is given an upper bound on the DSP delaying its data, come up with a fixed minimal amount of transitions points we must store to cover the history.