From mboxrd@z Thu Jan 1 00:00:00 1970 From: kbeldan@baylibre.com (Karl Beldan) Date: Tue, 23 Aug 2016 09:52:37 +0000 Subject: [PATCH v2] ARM: davinci: da8xx-dt: Add ti-aemif lookup for clock matching In-Reply-To: <2ef9651f-9a1e-3fb6-cc81-b1b714c9e043@ti.com> References: <20160819164736.10199-1-kbeldan@baylibre.com> <20160819164736.10199-2-kbeldan@baylibre.com> <2ef9651f-9a1e-3fb6-cc81-b1b714c9e043@ti.com> Message-ID: <20160823095237.GA8276@gobelin> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Tue, Aug 23, 2016 at 02:43:36PM +0530, Sekhar Nori wrote: > On Friday 19 August 2016 10:17 PM, Karl Beldan wrote: > > The davinci boards don't have their clocks in DT yet and getting a clock > > with a null con_id will fail, unless registering them as clk_lookups. > > This statement is true, but I am not sure why you limit it to those > clocks with null con_id. On DaVinci, all clocks need to be registered > via a clk_lookup entry. Else clk_get() will fail. > > So can I just edit this to: > > " > The davinci boards don't have their clocks in DT yet and getting a clock > will fail, unless registering them as clk_lookups. > " > > while applying? > Sure, as well as a 'reworded commit message' tag. Thanks, Karl > > This registers the aemif clock for the ti-aemif memory driver. > > > > The current aemif lookup entry resolving to the same clock: > > 'CLK(NULL, "aemif", &aemif_clk)' > > is currently used by davinci_nand and remains for non-DT and backward > > compatibility. > > > > Currently the davinci boards only configure the AEMIF in case of non-DT > > boot, via some code in mach-davinci. > > This change will allow DT-based davinci platforms to do the same, via > > the ti-aemif memory driver code. > > > > Signed-off-by: Karl Beldan > > --- > > This patch was originally patch 1/4 of the series > > "Add DT support for NAND to LCDK via ti-aemif" from which two other > > patches have already applied in Sekhar's tree. > > Sekhar, please note that commit 31e3a881 ("ARM: dts: da850,da850-evm: > > Add an aemif node and use it for the NAND") of your tree depends on > > this change. > > yes, noted. While sending the pull request to ARM-SoC, I send the SoC > changes first. > > Thanks, > Sekhar