Quoting Tony Lindgren (2019-09-18 13:53:44) > Hi, > > * Stephen Boyd <sboyd@xxxxxxxxxx> [190918 18:08]: > > Quoting Tony Lindgren (2019-09-08 12:42:41) > > > Or do you have some better ideas on how to name a clock controller > > > in the device tree? > > > > > > > Why does the name of the clock controller or clkdm_name matter? Using a > > string from reg-names smells like a workaround to describe some sort of > > linkage between things that isn't being described in DT today. > > Correct. This problem will eventually disappear with genpd > handling the clockdomains. > > But currently the clockdomain name is parsed from the dt node > name, which is not standard practise. Using reg-names > is a standard binding, and it's usage follows the standand > here to describe the reg range. > > Then eventually with genpd, the reg-names will just become > optinoal. But until that happens the $subject patch fixes > issues as described in the patch. > Is anything broken? It looks like the hidden dependency on the node name is being changed to be a slightly less hidden dependency on reg-names. reg-names is supposed to be an optional property, so we're trading one thing for another. I still don't understand the reasoning here, but if Tero is happy to ack/review this change then I'm not too worried about it assuming the reg-names property eventually becomes optional. Just seems like more work and DT churn for no to little gain?
