2016-02-25 7:08 GMT+09:00 Stephen Boyd [off-list ref]:

On 02/23, Rob Herring wrote:

quoted

On Mon, Feb 22, 2016 at 08:48:39PM +0900, Akinobu Mita wrote:

quoted

This adds userspace consumer for common clock.

This driver is inspired from Userspace regulator consumer
(REGULATOR_USERSPACE_CONSUMER) and it is useful for test purposes and
some classes of devices that are controlled entirely from user space.

Whether the kernel handles a clock or userspace does should have nothing
to do with DT. Plus DT does not expose many of the clocks, so any
debug feature needs to expose all clocks and be independent of DT.

Beyond debug, I don't think we should be exposing clock control to
userspace.

Agreed. Plus, the userspace regulator consumer doesn't have any
DT binding for it, so I don't see why we would have a DT binding
for the clk one.

OK.  I'll drop a DT binding part of this driver.

I'll still try to expose this feature by platform device instead of
exposing under clk debugfs.  Because we don't want to allow controlling
all clock devices (which may result abuse of the feature).  Instead,
this userspace consumer only allows explicitly specified devices to
control.