On 21/03/2014 at 13:29:31 +0100, Sebastian Hesselbarth wrote : > >>Hmm, you probably know better than me, but if cpuclk == cpupll > >>is always true we don't need another clk layer here. If you > >>can scale down cpuclk from cpupll and we just have no driver > >>for it, I am fine with it. > >> > > > >You can actually switch CPU clk from CPU pll to smclk. I'm not sure this > >is completely useful yet though, probably for suspend ? > > Then it should be clk mux instead? > > >Also, while I'm not sure this is a good reason, other clocks are derived > >from CPU pll and have another divider. > > I have no strong opinion, but a fixed-factor-clock with 1:1 just to > rename cpupll to cpuclk seems a bit wasty ;) > > If there is a mux, we should add it now - no matter if we are ever > going to make any use of it. For the derived clocks we should be > careful if they actually depend on cpuclk or always cpupll. > > If your (current) knowledge of the berlin clock trees is almost as > bad as mine, we can also ignore cpuclk mux if you prefer. > Yeah, fact is I know there is a mux but I don't know yet how to get/set its state so I will ignore it until we have more info. -- Alexandre Belloni, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com -- To unsubscribe from this list: send the line "unsubscribe linux-doc" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
