RFC: representing sdio devices oob interrupt, clks, etc. in device tree

[Ulf Hansson]

[snip]

>>
>> Why do we need to put the sdio functions devices in DT?
>
> To define sdio function specific non probable info, such as oob irqs,
> also see the "mmc: Add SDIO function devicetree subnode parsing" patch-set
> of which I send v3 this morning.

Yes, of course - makes sense.

[snip]

>>>> I see these important things that follow if we decide to use runtime
>>>> PM to trigger the power up/off sequence.
>>>> 1) In cases of !CONFIG_PM_RUNTIME, it means the "powerup driver" once
>>>> probed, will keep it's resources enabled forever.
>>>
>>> Ack.
>>
>> So, the consequence is that for CONFIG_PM_SLEEP systems not using
>> CONFIG_PM_RUNTIME - we don't have a good solution.
>>
>> Is that acceptable?
>
> IMHO yes, if people want maximum power savings they should use
> CONFIG_PM_RUNTIME. And since this is all for yet to be added
> systems / configs I would expect CONFIG_PM_RUNTIME to be supported
> there just fine.

That's a valid point, but on the other hand, "old" systems could
benefit from this "feature" as well, were runtime PM might not be that
interesting.

Kind regards
Uffe