From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751491AbWGaFYW (ORCPT ); Mon, 31 Jul 2006 01:24:22 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751492AbWGaFYW (ORCPT ); Mon, 31 Jul 2006 01:24:22 -0400 Received: from [210.76.114.181] ([210.76.114.181]:42934 "EHLO ccoss.com.cn") by vger.kernel.org with ESMTP id S1751491AbWGaFYV (ORCPT ); Mon, 31 Jul 2006 01:24:21 -0400 Message-ID: <44CD9408.1020405@ccoss.com.cn> Date: Mon, 31 Jul 2006 13:24:25 +0800 From: liyu User-Agent: Thunderbird 1.5 (X11/20051201) MIME-Version: 1.0 To: Dmitry Torokhov CC: Pavel Machek , LKML , Greg KH , Peter , The Doctor Subject: Re: [PATCH 1/3] usbhid: Driver for microsoft natural ergonomic keyboard 4000 References: <44C74708.6090907@ccoss.com.cn> <20060728135428.GC4623@ucw.cz> <44CD536C.3050703@ccoss.com.cn> <200607302130.56881.dtor@insightbb.com> In-Reply-To: <200607302130.56881.dtor@insightbb.com> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Well, I think you are right. I reviewed the "Linux Device Driver", there are some follow words in that book: "the role of a device driver is providing mechanism, not policy." It seem my hack break this rule obviously. I will "kcah" it back soon. Thanks. Dmitry Torokhov wrote: > On Sunday 30 July 2006 20:48, liyu wrote: > >> Pavel Machek Wrote: >> >>> Hi! >>> >>> >>> >>>> This new version get some improvements: >>>> >>>> 2. Support left paren key "(", right paren key ")", equal key "=" on >>>> right-top keypad. In fact, this keyboard generate KEYPAD_XXX usage code >>>> for them, but I find many applications can not handle them on default >>>> configuration, especially X.org. To get the most best usability, I use a >>>> bit magic here: map them to "Shift+9" and "Shift+0". >>>> >>>> >>> That is hardly 'improvement'. 'X is broken, so lets break input, too'. >>> >>> >>> >>> >>> >> Well, however, this can work truly. If we do not hack as this way. >> Many applications can not get its input. I think the usability for >> most people should be first, but not follow rules. >> >> > > I do not quite understand why X would have issues with it. KEY_KPEQUAL, > KEY_KPLEFTPAREN and KEY_KPRIGHTPAREN should work fine even with legacy > X keyboard driver (one that is using PS/2 protocol instead of evdev). > You might want to adjust your XKB map or use xmodmap, but kernel should > report true keycodes. > > >> I think we can add one module parameter like "shift_hack" to switch it ?! >> >> > > No please don't. > >