[PATCH 4/4] clk: wm831x: Add initial WM831x clock driver

Turquette, Mike mturquette at ti.com
Thu Mar 22 14:35:29 EDT 2012

On Thu, Mar 22, 2012 at 4:33 AM, Mark Brown
<broonie at opensource.wolfsonmicro.com> wrote:
> On Thu, Mar 22, 2012 at 12:26:13PM +0100, Sascha Hauer wrote:
>> On Thu, Mar 22, 2012 at 11:15:31AM +0000, Mark Brown wrote:
>> > No, that's not sensible.  We shouldn't be open coding this into each
>> > individual driver that provides clocks, and we shouldn't have clock
>> > users having to guess at what scheme the driver author used to dedupe
>> > the clocks.  As a driver author you would assume that the reason we're
>> It seems they got dropped. Currently the clock framework does nothing
>> with the device argument. But right, it could use this argument to
>> generate a suitable name.
> Indeed, and as with adding the stub clk_unregister() it seems like it's
> better at this point to just code the drivers as we'd expect them to
> work and fill out the framework to handle this rather than introducing
> workarounds in the drivers.

Agreed.  I put this off some time back, but I figured the WM831X
driver would bring the issue back up.

I'm happy for the core to concatenate the strings when a struct device
*dev is passed in.  However the reason I dropped this in the first
place is that I have some ideas on providing something like clk_get
directly from the common clk core and I hadn't figured out all of the
details yet.  For now we can handle the string concatenation in the
core and figure out those tricky details if/when the time comes to
provide a clk_get which is more closely linked to the clock framework

I'll roll in a patch for my fixes series.


More information about the linux-arm-kernel mailing list