[U-Boot] [PATCH] rockchip: dts: rk3399: rockpro64: Provide init voltage

[Mark Kettenis]

> From: Jagan Teki <jagan at amarulasolutions.com>
> Date: Sat, 22 Jun 2019 20:17:13 +0530
> 
> On Sat, Jun 22, 2019 at 2:46 AM Mark Kettenis <kettenis at openbsd.org> wrote:
> >
> > Add missing regulator-init-microvolt property to vdd_log regulator.
> >
> > Signed-off-by: Mark Kettenis <kettenis at openbsd.org>
> > ---
> > Without this property the pwm regulator driver prints a warning and
> > the OpenBSD kernel crashes shortly after spinning up the big cores.
> >
> >  arch/arm/dts/rk3399-rockpro64-u-boot.dtsi | 4 ++++
> >  1 file changed, 4 insertions(+)
> >
> > diff --git a/arch/arm/dts/rk3399-rockpro64-u-boot.dtsi b/arch/arm/dts/rk3399-rockpro64-u-boot.dtsi
> > index dbfa4ba9f8..f7f26d584f 100644
> > --- a/arch/arm/dts/rk3399-rockpro64-u-boot.dtsi
> > +++ b/arch/arm/dts/rk3399-rockpro64-u-boot.dtsi
> > @@ -5,3 +5,7 @@
> >
> >  #include "rk3399-u-boot.dtsi"
> >  #include "rk3399-sdram-lpddr4-100.dtsi"
> > +
> > +&vdd_log {
> > +       regulator-init-microvolt = <950000>;
> > +};
> 
> This seems to be issue on whole? how does Linux work? if it required
> in Linux better make a change there and sync here.

It needs the fix to actually turn on the boot-on regulators as well.
But yes, with this property in the device tree, U-Boot enables the
regulator with the appropriate voltage and the crash I'm seeing is gone.

I deliberately made this change in the rk3399-rockpro64-u-boot.dtsi
file since the "regulator-init-microvolt" property is *not* defined by
the Linux devicetree bindings.  So I expect that a patch to change
this in the Linux device tree would be rejected.

Note that this regulator has a "regulator-boot-on" property.  So Linux
doesn't need to know the initial setting of this regulator as the
bootloader/firmware is supposed to set it up.

I don't know if folks running Linux are experiencing problems as wll;
I don't run Linux on my boards ;).  I suspect very few people are
actually using mainline U-Boot on their boards.  But it is also
possible that the Linux kernel does enable this regulator when it
boots, since it sees that it hasn't been enabled.  It does have some
fallback code to fix that up in case the bootloader/firmware didn't do
things properly.  But what voltage would it use?