From: Robert Schwebel Subject: Re: High latencies on MPC5200 with patch-2.6.24-rc5-rt1 Date: Thu, 10 Jan 2008 17:23:51 +0100 Message-ID: <20080110162351.GN3177@pengutronix.de> References: <47834CB7.1030102@grandegger.com> <20080108154947.GB3177@pengutronix.de> <20080109090345.GG3177@pengutronix.de> <20080109225107.GV3177@pengutronix.de> <20080110000124.GW3177@pengutronix.de> <47861ACE.9060301@grandegger.com> <20080110143527.GI3177@pengutronix.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-15 Cc: Thomas Gleixner , linux-rt-users@vger.kernel.org, John Stultz To: Wolfgang Grandegger Return-path: Received: from metis.extern.pengutronix.de ([83.236.181.26]:59657 "EHLO metis.extern.pengutronix.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756075AbYAJQX5 (ORCPT ); Thu, 10 Jan 2008 11:23:57 -0500 Content-Disposition: inline In-Reply-To: <20080110143527.GI3177@pengutronix.de> Sender: linux-rt-users-owner@vger.kernel.org List-ID: On Thu, Jan 10, 2008 at 03:35:27PM +0100, Robert Schwebel wrote: > > Could you run your cyclictest much longer? > > I'll re-test with your parameters. Done, here are the results (data and histogram): http://www.pengutronix.de/software/realtime-preempt/20080110-1-cyclictest/20080110-1-phyCORE-MPC5200B-tiny-data.png http://www.pengutronix.de/software/realtime-preempt/20080110-1-cyclictest/20080110-1-phyCORE-MPC5200B-tiny-histo.png Especially the data plot is interesting: it shows that most of the shots are below 80 us, but there are regular peaks. If you look carefully, you also see that the peaks are "growing", which usually means that there is some interference. Time to activate latency hunting... Robert -- Dipl.-Ing. Robert Schwebel | http://www.pengutronix.de Pengutronix - Linux Solutions for Science and Industry Handelsregister: Amtsgericht Hildesheim, HRA 2686 Hannoversche Str. 2, 31134 Hildesheim, Germany Phone: +49-5121-206917-0 | Fax: +49-5121-206917-9