> > Another artefact of not have a child-parent relationship. I wounder if
> > it makes sense to change the architecture. Have the PCS driver
> > instantiate the PCS devices as its children. They then have a device
> > structure for calls like clk_bulk_get(), and a more normal
> > consumer/provider setup.
> >
>
> I think you may be suggesting to drop the child node usage in the DTS, so
> that we can attach all the MII clocks to the single PCS node, to facilitate
> usage of bulk get() API to retrieve the MII clocks for that PCS.
I would keep the child nodes. They describe the cookie-cutter nature
of the hardware. The problem is with the clk_bulk API, not allowing
you to pass a device_node. of_clk_bulk_get() appears to do what you
want, but it is not exported. What we do have is:
/**
* devm_get_clk_from_child - lookup and obtain a managed reference to a
* clock producer from child node.
* @dev: device for clock "consumer"
* @np: pointer to clock consumer node
* @con_id: clock consumer ID
*
* This function parses the clocks, and uses them to look up the
* struct clk from the registered list of clock providers by using
* @np and @con_id
*
* The clock will automatically be freed when the device is unbound
* from the bus.
*/
struct clk *devm_get_clk_from_child(struct device *dev,
struct device_node *np, const char *con_id);
So maybe a devm_get_clk_bulk_from_child() would be accepted?
However, it might not be worth the effort. Using the bulk API was just
a suggestion to make the code simpler, not a strong requirement.
Andrew
