Hello Wenyou,

[snip]

quoted

quoted

If then device is not being loaded from the DT (and it shouldn't be,
the device looks like it should be instantiated directly by the MFD as
it can't exist separately to that MFD) an OF table will do nothing.

Then he should remove the .of_compatible from the MFD cell definition.

I tried it,

But if removed this .of_compatible, and reserved the OF table.

What we said is that you should remove both the mfd cell
.of_compatible field and the OF table. Of course removing one but
leaving the other is not correct.

the &pdev->dev->of_node is NULL, the driver fails to get the configuration value from DT,

I see, you mean that you are not able to lookup the
"active-semi,vsel-high" property.

It seems the OF table still doesn't work. Where is wrong?

I think the problem is with your DT binding. You have a PMIC dev node
with compatible "active-semi,act8945a" that has a child node with
compatible "active-semi,act8945a-regulator" which in turn has the
"regulators" node.

I believe what Mark says is that there shouldn't be a compatible and
node for the regulators IC since is part of the PMIC. IOW, you can't
have it as a standalone node in the DTS.

Could you help give some suggestion?

I would just remove the "active-semi,act8945a-regulator" node and make
"active-semi,vsel-high" a property of the "active-semi,act8945a" node.
That way you can remove the mfd cell .of_compatible and OF table in
the regulator driver and read the "active-semi,vsel-high" using the
platform device's parent of_node.

But it's better if you wait for Mark's opinions before re-spining your patches.

Best regards,
Javier