From mboxrd@z Thu Jan 1 00:00:00 1970 From: Wolfgang Grandegger Subject: Re: High latencies with linux-2.6.25.4-rt6 on MPC5200 Date: Mon, 30 Jun 2008 19:34:57 +0200 Message-ID: <48691941.5040100@grandegger.com> References: <4858C229.2060509@grandegger.com> <20080619131406.GK5724@pengutronix.de> <485A675A.9070509@grandegger.com> <20080627143213.GG5666@pengutronix.de> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit To: Wolfgang Grandegger , RT Return-path: Received: from mail-out.m-online.net ([212.18.0.9]:52865 "EHLO mail-out.m-online.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751786AbYF3Rea (ORCPT ); Mon, 30 Jun 2008 13:34:30 -0400 In-Reply-To: <20080627143213.GG5666@pengutronix.de> Sender: linux-rt-users-owner@vger.kernel.org List-ID: Hi Fu, Luotao Fu wrote: > Hi Wolfgang, > I finally found some time and did some tests with your config file and the shiny > new 2.6.25.8-rt7 > > On Thu, Jun 19, 2008 at 04:04:10PM +0200, Wolfgang Grandegger wrote: >> Hi Fu, >> >> Luotao Fu wrote: >> Keep in mind that I'm running Linux on a rootfs mounted via NFS. > > I have exactly the same test configuration here. > > Again everything looks just fine to me. Depending on which debugging Option are > turned on my max. results pended between 70 and 350 us. However I could not > produce any peak worst case values, which are so far away from the average like > in your case. The .config you send me correspond to what latency (debugging options turned on). > Have you tried the new -rt patch on your board? Still any strange results? I just re-tested 2.6.25.8-rt7 with a .config as close as possible to yours. Still, the latency measured with "cyclictest -p 98 -n -t1 -i1000" goes up to 600us within a few minutes. I also re-measured latencies with older kernel version: RT kernel Max. latency 2.6.25.8-rt7 605 us 2.6.24.7-rt13 230 us 2.6.24.4-rt4 130 us I will now try to narrow down the latency evils. Wolfgang.