From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1161107AbWJNJ6U (ORCPT ); Sat, 14 Oct 2006 05:58:20 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1161103AbWJNJ6T (ORCPT ); Sat, 14 Oct 2006 05:58:19 -0400 Received: from poczta.o2.pl ([193.17.41.142]:42136 "EHLO poczta.o2.pl") by vger.kernel.org with ESMTP id S1161107AbWJNJ6S (ORCPT ); Sat, 14 Oct 2006 05:58:18 -0400 From: =?iso-8859-2?q?=A3ukasz_Fibinger?= Reply-To: lucke@o2.pl To: linux-kernel@vger.kernel.org Date: Sat, 14 Oct 2006 11:58:14 +0200 User-Agent: KMail/1.9.5 MIME-Version: 1.0 Content-Disposition: inline Subject: cpufreq not working after putting core down and up Content-Type: text/plain; charset="iso-8859-2" Content-Transfer-Encoding: 7bit Message-Id: <200610141158.15006.lucke@o2.pl> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Dear kernel hackers, I'm having this peculiar issue on my Athlon 64 X2 3800+ (Asus A8N-E motherboard, latest BIOS) box. Namely, after I put one core offline and then online, its cpufreq driver doesn't work anymore. # cd /sys/devices/system/cpu/cpu1 # ls cache cpufreq crash_notes online topology # echo 0 >> online # echo 1 >> online # ls cache crash_notes online topology Note the lack of "cpufreq" symlink to cpu0/cpufreq in cpu1 dir. This behaviour prohibits me from using cpufreq after resuming a suspended system. I have been trying to track down the issue for ca. two months. Yet, I failed to pinpoint the culprit. Mark Langsdorf (AMD) provided me with his kind assistance on cpufreq mailinglist, but we failed to come to any conclusions/solutions (for one, Mark couldn't reproduce it himself). I only managed to find a report of a similar problem by a Turion X2 user, yet it allegedly disappeared after him moving from ArchLinux to Ubuntu. However, no other X2 user of ArchLinux (my distro of choice) seems to suffer from the same problem (except that Turion guy), not even on a A8N-E MB. My tests weren't full-blown (due to the lack of resources), but it seems that the problem persists on different distros on my box as well. I tried quite a bunch of different kernel/kernel config combinations too, without effect. I fail to find any relationship (reminds me of a criptid hunt) - thus, I'm totally clueless whether it's a hardware/kernel/userspace problem. Any suggestions would be greatly appreciated. Of course, I'm able to provide dmesg with cpufreq debugging or what else is needed upon request. As I'm not suscribed to lkml, please kindly CC all replies to my address. Regards, Luke