From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752733AbYIJOQS (ORCPT ); Wed, 10 Sep 2008 10:16:18 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751874AbYIJOQH (ORCPT ); Wed, 10 Sep 2008 10:16:07 -0400 Received: from mx2.redhat.com ([66.187.237.31]:45216 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751574AbYIJOQG (ORCPT ); Wed, 10 Sep 2008 10:16:06 -0400 From: Jarod Wilson Organization: Red Hat, Inc. To: Janne Grunau Subject: Re: [PATCH 06/18] lirc driver for the ATI USB RF remote receiver Date: Wed, 10 Sep 2008 10:13:18 -0400 User-Agent: KMail/1.10.1 (Linux/2.6.25.16-1.fc10.x86_64; KDE/4.1.1; x86_64; ; ) Cc: Christoph Hellwig , Ville =?iso-8859-1?q?Syrj=E4l=E4?= , linux-kernel@vger.kernel.org, Christoph Bartelmus References: <1220933164-10160-1-git-send-email-jwilson@redhat.com> <20080910131432.GA2161@infradead.org> <200809101544.17704.j@jannau.net> In-Reply-To: <200809101544.17704.j@jannau.net> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200809101013.19205.jwilson@redhat.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wednesday 10 September 2008 09:44:17 Janne Grunau wrote: > On Wednesday 10 September 2008 15:14:32 Christoph Hellwig wrote: > > On Wed, Sep 10, 2008 at 09:05:58AM -0400, Jarod Wilson wrote: > > > True, though I think some users still prefer using them with the > > > lirc drivers for assorted reasons. Configuring something like > > > mythtv to work with the ati_remote{,2} driver appears to be a bit > > > more complex (or at least non- standard vs. several other popular > > > remotes) and not as functional vs. configuring mythtv > > > w/lirc_atiusb. > > > > Bad idea to have two drivers for the same piece of hardware. And > > this gets straight back into the why should lirc be different from > > the input layer point raised earlies. I think we really shouldn't > > keep lirc as a separate subsystem, but make sure all the drivers are > > written to the input layer. > > Some drivers don't report events but deliver the signal to the lirc > daemon. The daemon decodes the signal. Decoding of several IR protocols > and mapping of each possible remote to events are things much easier > done in userspace. > Aslo LIRC is not only about IR input but also output. > > See following mail from Christoph Bertelmus for more details. > http://sourceforge.net/mailarchive/message.php?msg_id=AAryKgq9UOB%40bartelm >us.de Yeah, I think that pretty well covers the issues. > > To make the migration easier for exiting > > users we could add a /dev/lirc driver ontop of the input layer, > > This shouldn't be necessary since the lirc daemon can already read from > input devices. Okay, so what I'm hearing and thinking is that we should at drop lirc_atiusb from consideration for inclusion, and instead extend ati_remote{,2} as/if necessary to be able to let the lirc daemon use it via the provided input device, so people can get the best of both worlds (it behaves like a keyboard if so configured, talks to lircd if so configured). -- Jarod Wilson jarod@redhat.com