Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263005AbTIVGJu (ORCPT ); Mon, 22 Sep 2003 02:09:50 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263007AbTIVGJu (ORCPT ); Mon, 22 Sep 2003 02:09:50 -0400 Received: from twilight.ucw.cz ([81.30.235.3]:57528 "EHLO twilight.ucw.cz") by vger.kernel.org with ESMTP id S263005AbTIVGJs (ORCPT ); Mon, 22 Sep 2003 02:09:48 -0400 Date: Mon, 22 Sep 2003 08:09:38 +0200 From: Vojtech Pavlik To: Dmitry Torokhov Cc: Vojtech Pavlik , Peter Osterlund , Linus Torvalds , linux-kernel@vger.kernel.org Subject: Re: Broken synaptics mouse.. Message-ID: <20030922060938.GA27702@ucw.cz> References: <200309211816.36783.dtor_core@ameritech.net> <20030922053103.GA27045@ucw.cz> <200309220058.48015.dtor_core@ameritech.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <200309220058.48015.dtor_core@ameritech.net> User-Agent: Mutt/1.5.4i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1705 Lines: 41 On Mon, Sep 22, 2003 at 12:58:48AM -0500, Dmitry Torokhov wrote: > But in this case not only will I have to specify the event device Synaptics > is connected to but also explicitly specify _every other_ input device I use > (besides the touchpad I have a track-stick as a separate device and an USB > mouse in my docking station). I will also loose hot-plug capabilities I have > now for free. The problem is that nothing comes for free. The hotplug capability you have now comes at the cost of utter nonconfigurability, that is - it gets input from every pointing devices. The bright-future solution is to have GPM and XFree configured at runtime by the /sbin/hotplug agent to open/close devices as they come and go. This again comes at a cost - more complex setup - but one that is completely configurable as to which device goes where, stable in regard of plug-in order, etc. > All in all it just doesn't fly... I wonder if we could declare evdev > the master handler and do not propagate events to the secondary > handlers if some process has appropriate event device opened. We could. Don't do a 'cat /dev/input/event0' if event0 is your keyboard then, though. ;) We also could change the EVIOCGRAB semantics to only grab in respect to the handler, not just the single handle ... But I still hope we'll be able to get to the /sbin/hotplug solution without having to create too many intermediate solutions. -- Vojtech Pavlik SuSE Labs, SuSE CR - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/