From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chris Ball Subject: Re: [PATCH 3/3] mmc: esdhc i.MX: Support 8bit mode Date: Tue, 25 Sep 2012 03:52:09 -0400 Message-ID: References: <1348471345-30785-1-git-send-email-s.hauer@pengutronix.de> <1348471345-30785-4-git-send-email-s.hauer@pengutronix.de> <20120925071505.GG27291@S2101-09.ap.freescale.net> <20120925072714.GY1322@pengutronix.de> <20120925073317.GI27291@S2101-09.ap.freescale.net> <20120925073808.GZ1322@pengutronix.de> Mime-Version: 1.0 Content-Type: text/plain Return-path: Received: from void.printf.net ([89.145.121.20]:49356 "EHLO void.printf.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753607Ab2IYHvT (ORCPT ); Tue, 25 Sep 2012 03:51:19 -0400 In-Reply-To: <20120925073808.GZ1322@pengutronix.de> (Sascha Hauer's message of "Tue, 25 Sep 2012 09:38:08 +0200") Sender: linux-mmc-owner@vger.kernel.org List-Id: linux-mmc@vger.kernel.org To: Sascha Hauer Cc: Shawn Guo , linux-mmc@vger.kernel.org, linux-arm-kernel@lists.infradead.org, Eric =?utf-8?Q?B=C3=A9nard?= , kernel@pengutronix.de Hi, On Tue, Sep 25 2012, Sascha Hauer wrote: >> > > > +- bus-width : Maximum supported bus width. Defaults to 4 if omitted >> > > > >> > > This is a common mmc property documented in bindings/mmc/mmc.txt. >> > > Instead of duplicating the documentation, we should try to make our >> > > implementation conform to the common definition of the property. >> > >> > It is conform to the common definition, so all I have to do is drop the >> > line duplicating the docs. That's easy ;) >> > >> Maybe not. The common binding defines it as a required property while >> the patch implements it as an optional one. > > So you want to make the code conform to the common spec which has > bus-width as a required property. Is this really worth it to add an > incompatible change to our devicetrees? We could also break the stalemate by just changing the common spec to have bus-width become optional, default 4. It's not going to break the code of anyone who's been treating it as required. I don't think there there was a principled reason behind making it required. Thanks, - Chris. -- Chris Ball One Laptop Per Child From mboxrd@z Thu Jan 1 00:00:00 1970 From: cjb@laptop.org (Chris Ball) Date: Tue, 25 Sep 2012 03:52:09 -0400 Subject: [PATCH 3/3] mmc: esdhc i.MX: Support 8bit mode In-Reply-To: <20120925073808.GZ1322@pengutronix.de> (Sascha Hauer's message of "Tue, 25 Sep 2012 09:38:08 +0200") References: <1348471345-30785-1-git-send-email-s.hauer@pengutronix.de> <1348471345-30785-4-git-send-email-s.hauer@pengutronix.de> <20120925071505.GG27291@S2101-09.ap.freescale.net> <20120925072714.GY1322@pengutronix.de> <20120925073317.GI27291@S2101-09.ap.freescale.net> <20120925073808.GZ1322@pengutronix.de> Message-ID: To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org Hi, On Tue, Sep 25 2012, Sascha Hauer wrote: >> > > > +- bus-width : Maximum supported bus width. Defaults to 4 if omitted >> > > > >> > > This is a common mmc property documented in bindings/mmc/mmc.txt. >> > > Instead of duplicating the documentation, we should try to make our >> > > implementation conform to the common definition of the property. >> > >> > It is conform to the common definition, so all I have to do is drop the >> > line duplicating the docs. That's easy ;) >> > >> Maybe not. The common binding defines it as a required property while >> the patch implements it as an optional one. > > So you want to make the code conform to the common spec which has > bus-width as a required property. Is this really worth it to add an > incompatible change to our devicetrees? We could also break the stalemate by just changing the common spec to have bus-width become optional, default 4. It's not going to break the code of anyone who's been treating it as required. I don't think there there was a principled reason behind making it required. Thanks, - Chris. -- Chris Ball One Laptop Per Child