On Wed, Feb 08, 2012 at 09:54:37AM +0000, Liam Girdwood wrote:

> However, I think one of the intentions of the clk framework rework is to
> allow non CPU clocks like CODEC PLLs to register as clocks and have
> their dependencies worked out. I've no idea of the status of this atm,
> but maybe Mark knows or you could ask on lkml.

It's going nowhere slowly.  Seems unlikely for 3.4 at this point,
there's been no activity for a month or two except for rmk getting
annoyed at the lack of activity.  The current block is that very few
people are converting the new clk_prepare() APIs, though there's other
problems we seem unable to get any traction on beyond that.

I rather suspect that the general lack of any end in sight is not
helpful for motivation of random maintainers and none of the folks
working on the core API are doing any of that legwork for them.