From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jean Delvare Subject: Re: [PATCH] clk: mediatek: Fix MT2701 dependencies Date: Tue, 24 Jan 2017 13:58:36 +0100 Message-ID: <20170124135836.17f32ade@endymion> References: <20170109113621.31d384c9@endymion> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8BIT Return-path: In-Reply-To: Sender: linux-clk-owner@vger.kernel.org To: Andreas =?UTF-8?B?RsOkcmJlcg==?= Cc: linux-clk@vger.kernel.org, linux-mediatek@lists.infradead.org, Matthias Brugger , James Liao , Erin Lo , Stephen Boyd , Shunli Wang , Michael Turquette List-Id: linux-mediatek@lists.infradead.org Hallo Andreas, On Mon, 9 Jan 2017 21:08:50 +0100, Andreas Färber wrote: > Another aspect here is that this is a 32-bit SoC but it propagates into > the arm64 configs, so maybe (ARCH_MEDIATEK && !ARM64) || COMPILE_TEST? > > Same for mt2701 pinctrl. I took a look and the pinctrl situation is different: config PINCTRL_MT2701 bool "Mediatek MT2701 pin control" if COMPILE_TEST && !MACH_MT2701 (...) default MACH_MT2701 So the user will not be prompt about it on ARM64 (unless build-testing) because MACH_MT2701 isn't set on ARM64. The only issue with this construct is that you end up with useless symbols defined in the configuration file (they can only be "n".) So I would argue that the following is preferred: config PINCTRL_MT2701 bool "Mediatek MT2701 pin control" (...) depends on MACH_MT2701 || COMPILE_TEST default MACH_MT2701 And same for the other 3 PINCTRL_MT* options. Yes, that means the user will be asked about the options on Mediatek kernels, but actually I believe this is desirable, as advanced users should be allowed to disable specific drivers if they know what they are doing. Other users can just stick to the default. -- Jean Delvare SUSE L3 Support