On Tue, Jan 05, 2021 at 10:09:21AM -0500, Jim Quinlan wrote:

On Tue, Jan 5, 2021 at 9:01 AM Mark Brown [off-list ref] wrote:

quoted

quoted

For us, the supplies are for the EP chip's power.  We have the PCIe
controller turning them "on" for power-on/resume and "off" for
power-off/suspend.  We need the "xxx-supply" property in the
controller's DT node because of the chicken-and-egg situation: if the
property was in the EP's DT node, the RC  will never discover the EP
to see that there is a regulator to turn on.   We would be happy with

quoted

Why can't the controller look at the nodes describing devices for
standard properties?

It just feels wrong for the driver (RC) of one DT node to be acting on
a property of another driver's (EP) node, even though it is a subnode.

This is something we do for other buses, for example where there's
device specific tuning that is actually implemented in the controller
hardware.

There is also the possibility of the EP driver acting upon the
property simultaneously; we don't really have control of what EP
device and drivers are paired with our SOCs.

If the device is trying to do something with a supply that's a standard
part of the bus outside of the bus it seems like that's going to lead to
problems no matter what, due to the discovery issues the device must be
coordinating with the bus somehow.

In addition, this just pushes the binding name issue down a level --
what should these power supplies be called?  They are not slot power
supplies.  Can the  Broadcom STB PCIe RC driver's binding document
specify and define the properties of EP sub-nodes?

I assume the supplies have some name in the PCI specs, whatever names
are used there would probably be appropriate.