[RFC 1/2] pwrseq: Add subsystem to handle complex power sequences

Ulf Hansson ulf.hansson at linaro.org
Tue Jul 1 09:56:55 PDT 2014


On 20 June 2014 17:42, Arnd Bergmann <arnd at arndb.de> wrote:
> On Thursday 19 June 2014 15:04:50 Ulf Hansson wrote:
>> +Power sequence DT bindings
>> +
>> +Each power sequence method has a corresponding "power-method" property string.
>> +This property shall be set in a subnode for a device. That subnode should also
>> +describe resourses which are specific to that power method.
>> +
>> +Do note, power sequences as such isn't encoded through DT. Instead those are
>> +implemented by each power method.
>> +
>> +Required subnode properties:
>> +- power-method: should contain the string for the power method to bind.
>> +
>> +       Supported power methods: None.
>> +
>> +Example:
>> +
>> +Note, the "clock" power method in this example isn't actually supported, but
>> +used to visualize how a childnode could be described.
>
> I'm not too thrilled about adding another top-level concept for these.

I agree, but when you collects the requirements from the discussions
we have had around this topic - I don't think we can find another
solution. But I might be wrong.

> This seems to duplicate some things that pm-domains do, but does them
> in a somewhata different way. Would it be possible to instead integrate
> it into the pm-domain code?

No, I don't think so.

That main argument would be that runtime PM is not fine grained
enough, but there are several other reasons.

Please refer to previous discussions.

>
> I also agree with Olof that having a standalone child device node is
> not the best representation. If you want to represent an SDIO device
> device that has some references to clocks, regulators, etc, then put
> that device into the tree and give it those properties.

Could you elaborate on why?

Where would a card (SDIO/SD/MMC) be better placed - unless as a child
node under a mmc host device?

> That would also let you worry about the sequencing in driver code rather
> than trying to come up with a completely generic model for it.

So in principle your are suggesting to "pre-probe" all discoverable
devices/buses, not just for SDIO ( aka mmc subsystem).

Will that even work for modules?

Kind regards
Uffe

>
>         Arnd



More information about the linux-arm-kernel mailing list