From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnd Bergmann Subject: Re: [PATCH v2 29/38] thermal: exynos: Support both Exynos4x12 SoCs Date: Tue, 18 Jun 2013 15:02:52 +0200 Message-ID: <201306181502.52589.arnd@arndb.de> References: <1371486863-12398-1-git-send-email-t.figa@samsung.com> <201306172159.15352.arnd@arndb.de> <51BF9D0A.6050706@ti.com> Mime-Version: 1.0 Content-Type: Text/Plain; charset="iso-8859-15" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <51BF9D0A.6050706@ti.com> Sender: linux-samsung-soc-owner@vger.kernel.org To: Eduardo Valentin Cc: Tomasz Figa , linux-samsung-soc@vger.kernel.org, linux-arm-kernel@lists.infradead.org, Kukjin Kim , Olof Johansson , Marek Szyprowski , Sylwester Nawrocki , Thomas Abraham , linux-pm@vger.kernel.org, Zhang Rui , Kyungmin Park , Amit Daniel Kachhap List-Id: linux-pm@vger.kernel.org On Tuesday 18 June 2013, Eduardo Valentin wrote: > The driver will work on supported exynos variants. Those that have the > need for thermal sensing. And each of them have specific thermal needs > (trigger points, thresholds, etc). That is what this file tries to > isolate. And there is specific data structures for each soc version. Correct. My point is that the driver itself is much larger than the SoC-specific data sets in it. There is no reason to conditionally build a 108 byte data structure, making it more maintainably by removing all the #ifdef far outweighs the cost. You can also change the driver to work only for DT based booting, that will simplify it further and save you more in terms of object code size than the exynos_tmu_platform_data instances. Arnd