On Fri, Jun 20, 2014 at 02:13:58PM +0800, Chen-Yu Tsai wrote:

quoted

If these two clocks are these different though, maybe it would just be
easier to add a new driver. These are trivial enough anyway.

Yes it is. But it adds more of the same boilerplate code than actual
differences of the hardware.

I'm inclined to keep it as it is for now, unless someone strongly
objects.

I'm not strongly objecting, but I'd strongly prefer to go with two
drivers.

I don't really want to end up in the same case than with clk-sunxi.c,
which handles a lot of weird corner case for one particular clock,
while most of them are just a single-bit gate.

Especially when they have nothing in common, which is pretty much the
case here.

Maxime

-- 
Maxime Ripard, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com