From: m.szyprowski@samsung.com (Marek Szyprowski) Date: Mon, 06 Sep 2010 18:10:00 +0900 Subject: [PATCH 3/4] ARM: S5PC110: add common FIMC setup code In-Reply-To: References: <1283745044-15514-1-git-send-email-m.szyprowski@samsung.com> <1283745044-15514-4-git-send-email-m.szyprowski@samsung.com> Message-ID: <4C84AFE8.2010402@samsung.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org Hello, On 2010-09-06 13:52, Jassi Brar wrote: > On Mon, Sep 6, 2010 at 12:50 PM, Marek Szyprowski > wrote: > .... >> + parent = clk_get(NULL, "mout_epll"); >> + if (IS_ERR(parent)) >> + return PTR_ERR(parent); >> + >> + for (i = 0; err == 0&& i< ARRAY_SIZE(fimc_devs); i++) { >> + if (fimc_devs[i]) { >> + clk_fimc = clk_get(fimc_devs[i], "sclk_fimc"); >> + if (IS_ERR(clk_fimc)) { >> + err = PTR_ERR(clk_fimc); >> + break; >> + } >> + clk_set_parent(clk_fimc, parent); >> + clk_put(clk_fimc); >> + } >> + } > > The sclk_fimc could source clock from a number of options out of a mux. > mout_epll is just one of them. Different machines may want to source the > clock differently. Right, I forgot about this case. > So, IMO the parent selection should not be done in platform code, but rather > in machine init code. > Not the best, but a better solution can be found in > arch/arm/mach-s5pc100/dev-spi.c > Give it a thought. I'm thinking of making the parent clock an argument to the s5pv210_fimc_setup_clks(). I really don't like the idea of passing clock name through the platform data and letting driver to mess with clock's parents. Machine startup code is the last place where such things should be changed. Best regards -- Marek Szyprowski Samsung Poland R&D Center