Hi Mark Rutland > > -------------------- > > So, ideally. However we have to consider the fact that the clock API > > isn't reliably available makes this harder than it should be. Even > > among the DT using platforms at least PowerPC still uses a custom clock > > API. We could just use this as a carrot to push people to convert > > though. > > --------------------- > > I would be happier if we could unify the common clock infrastructure, it > would make this kind of thing a lot lessy messy. However, I'm not > against the system-clock-frequency property for now. Thank you > > OK, sorry for my English > > Sorry for the confusion, I'll try to be less ambiguous in future :) > > What I was trying to get at here is that if there is neither a clock or > a system-clock-frequency property in the device tree, dai->sysclk will > not have been initialised in this path. Is this a valid case, and will > dai->sysclk have a well-defined, sane value? My understanding, this "dai" itself is created by devm_kzalloc() So, default dai->sysclk is 0. And, if there is no clocks, no system-clock-frequency property, it try of_property_read_u32() side. but it will do nothing to dai->sysclk in such case. so dai->sysclk is still 0, and it is very sane on this driver. Is this good answer ? + clk = of_clk_get(np, 0); + if (IS_ERR(clk)) + of_property_read_u32(np, + "system-clock-frequency", + &dai->sysclk); + else + dai->sysclk = clk_get_rate(clk); Best regards --- Kuninori Morimoto -- 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
