From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S264536AbTL0TCS (ORCPT ); Sat, 27 Dec 2003 14:02:18 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S264537AbTL0TCS (ORCPT ); Sat, 27 Dec 2003 14:02:18 -0500 Received: from louise.pinerecords.com ([213.168.176.16]:11180 "EHLO louise.pinerecords.com") by vger.kernel.org with ESMTP id S264536AbTL0TCR (ORCPT ); Sat, 27 Dec 2003 14:02:17 -0500 Date: Sat, 27 Dec 2003 20:01:38 +0100 From: Tomas Szepe To: Dmitry Torokhov Cc: Andrew Morton , GCS , linux-kernel@vger.kernel.org, Peter Osterlund Subject: Re: Synaptics problems in -mm1 Message-ID: <20031227190138.GE10491@louise.pinerecords.com> References: <20031224095921.GA8147@lsc.hu> <200312271323.22252.dtor_core@ameritech.net> <20031227183704.GD10491@louise.pinerecords.com> <200312271345.41679.dtor_core@ameritech.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <200312271345.41679.dtor_core@ameritech.net> User-Agent: Mutt/1.4.1i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org > > Dmitry Torokhov wrote: > > > > mice: PS/2 mouse device common for all mice > > > > input: PC Speaker > > > > synaptics reset failed > > > > synaptics reset failed > > > > synaptics reset failed > > > > > > Ok, are you running with cpufreq? > > > > Yes, but > > 1) I've been on AC all the time. > > 2) 2.6.0 works. > > > > Well, I have a daemon that monitors load and dynamically switches > between high and low frequency... Sure, I've got one, too, but configured it to always go full throttle when on AC. > > > I think it ACPI PM timer if you have it activated - I am having > > > problems with it myself but didn't have time to look closer. Could > > > you try booting with clock=tsc or clock=pit and see if it fixes the > > > touchpad. > > > > clock=tsc appears to fix the problem. > > clock=pit no change. > > Ok, good. So it is the timer funkiness. I would suggest not using ACPI > PM timer for now then. And yes, timer_pit does not have cpufreq hooks > either so it probably not the best timesource with cpufreq either, > so stick with TSC. Stupid me, I absolutely forgot that I had enabled CONFIG_X86_PM_TIMER (nonexistent in 2.6.0 stock) which indeed seems to be the culprit. Please accept my apologies. -mm1 with that config option unset won't lose the stick, either. Thanks! -- Tomas Szepe