Re: [PATCH v2 0/2] regulator_sync_state() support

[Mark Brown]

On Thu, May 28, 2020 at 12:06:08PM -0700, Saravana Kannan wrote:

> The simplified explanation of the problem is, for regulators left on by
> the bootloader, we want to keep them on until all the consumers are
> probed. This is because we need to protect consumer-A from turning off a
> shared regulator used by consumer-B. Once consumer-B (and all the other
> consumers come up), they can do it themselves and the regulator
> framework no longer needs to keep the regulator on.

> So, this is not just about module or device probe ordering between
> suppliers and consumers. Even if we get the probe order prefectly right,
> it still won't solve this problem.

This logic seems to be circular - can you be concrete please?

> We can eventually extend this to also cover voltage and other
> properties, but in this patch series I want to get this right for
> "enabled/disabled" first.

I'm quite worried about the extension to voltage changes.

Attachment: signature.asc
Description: PGP signature