On Mon, Aug 04, 2014 at 04:13:41AM +0200, Luc Verhaegen wrote: > > How would you represent a single register controling 32 gates then? > > with 32 nodes? Be reasonable. > > > > We have to have an ID. We chose for that ID to have the one described > > in the datasheet. What fools we make.... > > > And I'm supposed to catch all your typos then? > > > Again, nothing to do with the clocks property. Could we *please* get > > to *a* point? Any point would do, but really, get your facts together. > > > Still nothing to do with the clocks property. > > > Yes, and? > > > We are neither name driver nor bit driven, we are phandle driven. Like > > I said above, exposing a register controlling 32 gates would then > > require to create 32 nodes. For sanity reasons, we chose to add an > > extra argument. Just like it's already used for interrupts, reset > > lines, gpios, pretty much every thing. > > > The bitmask here is to make sure we don't get an invalid clock ID, and > > be able to deal with "holes" in our register. > > In time.
? And I was expecting some answers... That seems hopeless, may I ask for suggestions then? -- Maxime Ripard, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com
signature.asc
Description: Digital signature
