Mark Brown <broonie@xxxxxxxxxx> 於 2020年8月25日 週二 下午6:08寫道：
>
> On Tue, Aug 25, 2020 at 06:01:46PM +0800, Gene Chen wrote:
> > Mark Brown <broonie@xxxxxxxxxx> 於 2020年8月25日 週二 下午5:34寫道：
>
> > > > LDO_VINx is HW design layout, so actually it can't be changed by device tree.
> > > > LDO_VIN1/LDO_VIN2 supply from VSYS, not regulator, so I think usually
> > > > not to show the supply from in device tree.
> > > > or I should declare a dummy reference to system power like "*-supply =
> > > > <&system_power>;"?
>
> > > When you say it's from the hardware design do you mean it's fixed by the
> > > silicon or is this something that's fixed in the board?
>
> > fixed in the board, not silicon.
>
> Ah, if it's fixed by the board that's exactly the sort of thing that
> should be in DT - the DT describes the board so if some other system has
> a different configuration then it will have a different DT. With supply
> properties you can just leave them missing in the DT, you're not
> supposed to but so many systems do it that the framework will handle it.
> Otherwise if you want to represent VSYS you can have a fixed voltage
> regulator with no control and hook them up to it.
ok, I think I will follow the framework to handle it.
I will add supply description in devicetree
LDO_VIN1-supply:
description: Input supply phandle(s) for LDO1/2/3 which is supply from VSYS
LDO_VIN2-supply:
description: Input supply phandle(s) for LDO5 which is supply from
VSYS
LDO_VIN3-supply:
description: Input supply phandle(s) for LDO6/7
