On 06/16/2016 06:39 PM, Sylwester Nawrocki wrote: > On 06/15/2016 11:40 AM, Lars-Peter Clausen wrote: >> On 06/13/2016 12:52 PM, Sylwester Nawrocki wrote: >> [...] >>>> + - compatible : "samsung,tm2-audio" >>>> + - samsung,model : the user-visible name of this sound complex >>>> + - clocks : must contain an entry for each entry in clock-names, >>>> + see ../clocks/clock-bindings.txt for details >>>> + - clock-names : must include the following entries: >>>> + "mclk1", "mclk2" >> >> If these are the MCLKs that go to the CODEC they should be properties of the >> CODEC node. > > Yes, these are both CODEC clocks. I couldn't find any related clk code in > the wm5110 arizona codec driver and indeed the clocks stay disabled when > I specify them in the wm5110 DT node. AFAICS from other thread in this > mailing list it is going to take some time until the CODEC drivers are > updated. How about minimizing the requirement of those clocks, making them > optional and then when the CODEC driver gets updated specifying the clocks > in the codec node instead of the sound card one? > The MCLK1 rate could be just hard coded instead of using clk_get_rate(), > but without clk enable calls nothing works. You can still handle the clocks in the machine driver, but they should be specified in the node of the CODEC to accurately represent the hardware. The simple-card driver e.g. does this, requests the clock of the CODEC and uses it to configure the CODEC input frequency based on it with snd_soc_dai_set_sysclk(). - Lars -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html