From mboxrd@z Thu Jan 1 00:00:00 1970 From: arnd@arndb.de (Arnd Bergmann) Date: Mon, 9 Apr 2012 14:14:28 +0000 Subject: [PATCH V3 1/5] ARM: kirkwood: Basic support for DNS-320 and DNS-325 In-Reply-To: References: <1332885255-8304-1-git-send-email-jm@lentin.co.uk> <20120406234955.840F73E1323@localhost> Message-ID: <201204091414.28918.arnd@arndb.de> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Monday 09 April 2012, Jamie Lentin wrote: > On Fri, 6 Apr 2012, Grant Likely wrote: > > These two lines are dangerous. It is not safe to 'select' Kconfig > > symbols that have 'depends' constraints. Otherwise, the symbol will > > get forced on without it's dependencies. > > > > Typically other code handles this by creating blank "HAVE_*" symbols > > that the needed symbol can do something like "default y if HAVE_*" > > > > Okay I didn't realise this, thanks. The options in question aren't > hard dependencies per-se, but no serial or NAND support is probably a > mistake, so added them here to avoid a certain amount of shooting in foot. > > What's the etiquette in this situation? Adding HAVE_* seems excessive, > would selecting them in kirkwood_defconfig make more sense? Or would it > simply be covered by having a reasonable config on my website or suchlike? Either way works. We don't prevent people from shooting themselves in the foot in Kconfig, there are plenty of ways to do that already (e.g. disabling network support on a headless system). There should be a defconfig for each platform, so just make sure 'make kirkwood_defconfig' enables everything you need and works on your system, it should be the superset of all kirkwood based systems and users can disable the stuff they don't need. Arnd From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnd Bergmann Subject: Re: [PATCH V3 1/5] ARM: kirkwood: Basic support for DNS-320 and DNS-325 Date: Mon, 9 Apr 2012 14:14:28 +0000 Message-ID: <201204091414.28918.arnd@arndb.de> References: <1332885255-8304-1-git-send-email-jm@lentin.co.uk> <20120406234955.840F73E1323@localhost> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: devicetree-discuss-bounces+gldd-devicetree-discuss=m.gmane.org-uLR06cmDAlY/bJ5BZ2RsiQ@public.gmane.org Sender: devicetree-discuss-bounces+gldd-devicetree-discuss=m.gmane.org-uLR06cmDAlY/bJ5BZ2RsiQ@public.gmane.org To: Jamie Lentin Cc: devicetree-discuss-uLR06cmDAlY/bJ5BZ2RsiQ@public.gmane.org, Jason , linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org List-Id: devicetree@vger.kernel.org On Monday 09 April 2012, Jamie Lentin wrote: > On Fri, 6 Apr 2012, Grant Likely wrote: > > These two lines are dangerous. It is not safe to 'select' Kconfig > > symbols that have 'depends' constraints. Otherwise, the symbol will > > get forced on without it's dependencies. > > > > Typically other code handles this by creating blank "HAVE_*" symbols > > that the needed symbol can do something like "default y if HAVE_*" > > > > Okay I didn't realise this, thanks. The options in question aren't > hard dependencies per-se, but no serial or NAND support is probably a > mistake, so added them here to avoid a certain amount of shooting in foot. > > What's the etiquette in this situation? Adding HAVE_* seems excessive, > would selecting them in kirkwood_defconfig make more sense? Or would it > simply be covered by having a reasonable config on my website or suchlike? Either way works. We don't prevent people from shooting themselves in the foot in Kconfig, there are plenty of ways to do that already (e.g. disabling network support on a headless system). There should be a defconfig for each platform, so just make sure 'make kirkwood_defconfig' enables everything you need and works on your system, it should be the superset of all kirkwood based systems and users can disable the stuff they don't need. Arnd