On 02/17, Joachim Eastwood wrote: > > For creg-clk it would probably work. > > cgu would need to register most of clocks early since it needs to > support whatever setup the bootloader throws at it. The only clocks > that could be taken out is the base clocks (clk_onecell_data > clk_base_data) except for one of them. But splitting this table would > be a bad idea as hw clk numbers are used as indexes for easy look up. > > Come to think of it; isn't splitting the clock table between a early > init and platform driver going become a problem? > At least the lookup in the driver or framework would become more complex(?) > And won't you confuse the lookup? > If the early driver register one clock table with > of_clk_add_provider() and then sometime later the platform driver > register another clock table, both using the same of_node. How is that > suppose to work? I believe that should work, although I haven't tested it. The code looks until it finds a provider that matches the device node and gives us a valid clk pointer. If we don't get a valid pointer we keep looking for another provider until we exhaust all providers. -- Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project -- 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
