From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mark Brown Subject: Re: [PATCH] mmc: Add architecture dependency for Marvell SoC controller Date: Mon, 13 Sep 2010 13:02:48 +0100 Message-ID: <20100913120247.GJ16606@rakim.wolfsonmicro.main> References: <1284366279-17140-1-git-send-email-broonie@opensource.wolfsonmicro.com> <201009131248.38501.arnd@arndb.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from opensource.wolfsonmicro.com ([80.75.67.52]:37799 "EHLO opensource2.wolfsonmicro.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1754534Ab0IMMCu (ORCPT ); Mon, 13 Sep 2010 08:02:50 -0400 Content-Disposition: inline In-Reply-To: <201009131248.38501.arnd@arndb.de> Sender: linux-mmc-owner@vger.kernel.org List-Id: linux-mmc@vger.kernel.org To: Arnd Bergmann Cc: Chris Ball , Andrew Morton , linux-mmc@vger.kernel.org, linux-kernel@vger.kernel.org On Mon, Sep 13, 2010 at 12:48:38PM +0200, Arnd Bergmann wrote: > I think we normally try to do the the other way round: in order to increase > build coverage, try to enable all drivers on as many platforms as possible, > just don't have them in the defconfig. This is not the general pattern for on-CPU devices on the embedded SoCs, for those we normally only offer the config option when building for a system which could physically have the device. Looking at the existing MMC host drivers you'll see this pattern is already heavily used there - all the existing embedded controllers have similar dependencies. > Do you get build errors on other platforms? If yes, there is probably > a bug in the driver that should be fixed instead. No, I didn't even try to build it - I saw that I was being offered a device for a Marvell CPU when building for a non-Marvell CPU.