Re: [PATCH 1/3] regulator: core: fix boot-on regulators use_count usage

From: Mark Brown
Date: Tue Oct 08 2019 - 08:51:47 EST

On Tue, Oct 08, 2019 at 08:03:11AM +0200, Marco Felsch wrote:
> On 19-10-07 19:29, Mark Brown wrote:
> > On Mon, Oct 07, 2019 at 11:34:29AM +0200, Marco Felsch wrote:

> > > Sorry that won't fix my problem. If I drop the regulator-boot-on state
> > > the fixed-regulator will disable this regulator but disable/enable this
> > > regulator is only valid during suspend/resume. I don't say that my fix
> > > is correct but we should fix this.

> > I'm having a bit of trouble parsing this but it sounds like you want the
> > regulator to be always on in which case you should use the property
> > specifically for that.

> Sorry my english wasn't the best.. Imagine this case: The bootloader
> turned the display on to show an early bootlogo. Now if I miss the
> regulator-boot-on property the display is turned off and on. The turn
> off comes from the regulator probe, the turn on comes from the cosumer.
> Is that assumption correct?

No, we shouldn't do anything when the regulator probes - we'll only
disable unused regulators when we get to the end of boot (currently we
delay this by 30s to give userspace a chance to run, that's a hack but
we're fresh out of better ideas). During boot the regulator state will
only be changed if some consumer appears and changes the state.

Attachment: signature.asc
Description: PGP signature