From mboxrd@z Thu Jan 1 00:00:00 1970 From: Miroslav Lichvar Subject: Re: [net 1/8] igb: shorten maximum PHC timecounter update interval Date: Thu, 1 Nov 2018 15:12:39 +0100 Message-ID: <20181101141239.GF19434@localhost> References: <20181031194254.16417-1-jeffrey.t.kirsher@intel.com> <20181031194254.16417-2-jeffrey.t.kirsher@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: davem@davemloft.net, netdev@vger.kernel.org, nhorman@redhat.com, sassmann@redhat.com, Jacob Keller , Richard Cochran To: Jeff Kirsher Return-path: Received: from mx1.redhat.com ([209.132.183.28]:40114 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1728245AbeKAXPv (ORCPT ); Thu, 1 Nov 2018 19:15:51 -0400 Content-Disposition: inline In-Reply-To: <20181031194254.16417-2-jeffrey.t.kirsher@intel.com> Sender: netdev-owner@vger.kernel.org List-ID: On Wed, Oct 31, 2018 at 12:42:47PM -0700, Jeff Kirsher wrote: > From: Miroslav Lichvar > > The timecounter needs to be updated at least once per ~550 seconds in > order to avoid a 40-bit SYSTIM timestamp to be misinterpreted as an old > timestamp. > > Since commit 500462a9d ("timers: Switch to a non-cascading wheel"), > scheduling of delayed work seems to be less accurate and a requested > delay of 540 seconds may actually be longer than 550 seconds. Shorten > the delay to 480 seconds to be sure the timecounter is updated in time. It looks like this is the v1 of the patch. There was a v2 I sent on Oct 26, which made the interval even shorter. I can send a separate patch for that change. -- Miroslav Lichvar