[PATCH] arm64: Kconfig: select PM for ARCH_VEXPRESS

[arnd@arndb.de (Arnd Bergmann)]

On Friday, June 17, 2016 4:49:34 PM CEST Sudeep Holla wrote:
> 
> On 17/06/16 16:07, Arnd Bergmann wrote:
> > On Friday, June 17, 2016 3:59:34 PM CEST Sudeep Holla wrote:
> 
> [...]
> 
> >> To be more precise genpd/runtime code return same error -ENODEV when
> >> PM_GENERIC_DOMAINS=n and if the no power domain is found with
> >> PM_GENERIC_DOMAINS=y . The AMBA bus code just checks for -EPROBE_DEFER
> >> to stop adding devices, while proceeds to access the device otherwise.
> >>
> >> However on Juno, the coresight needs to powered up before accessing it.
> >> Well all is fine if you specify power domains in DT and have
> >> PM_GENERIC_DOMAINS=y. However if PM_GENERIC_DOMAINS=n the AMBA ignores
> >> the error(-ENODEV) from runtime PM and access the device resulting in
> >> boot hang when adding amba devices.
> >>
> >> I thought we need is to have code to check for power domains in DT for
> >> the device even when PM_GENERIC_DOMAINS=n and return an error that can
> >> be used to skip registering the device and clearly identifies with other
> >> allowed errors. It may need a bit of surgery and closer look IMO. But
> >> when asked for opinion, I was suggested to do this. I admit that I
> >> didn't propose my thoughts then :)
> >
> > Ok, so we also need PM_GENERIC_DOMAINS, not just PM, right?
> >
> Yes, but latter won't select former. Generally I have seen domain
> controller code selecting PM_GENERIC_DOMAINS if PM=y

Then at least we need the vexpress code to select both as well,
otherwise you could end up with PM_GENERIC_DOMAINS disabled in
configurations that have none of the other code.

> > It also sounds like there is no easy way to just turn on the clocks
> > in the amba bus handler when PM_GENERIC_DOMAINS is turned off, other
> > than reimplementing PM_GENERIC_DOMAINS?
> >
> 
> Turn on power domain, not just clocks(just to be precise), yes. It needs
> some thoughts as there are some assumptions on which existing platform
> work. Adding the support to handle the scenario I mentioned earlier is
> not trivial. Also the driver to handle the device may not be compiled in
> when PM_GENERIC_DOMAINS=n or when PM=n. So I was thinking of not
> accessing the device and hence skip adding it.

I see. Another idea I had is flawed, too: we can't easily make ARM_AMBA
depend on PM_GENERIC_DOMAINS, because the main relevant use case I
can see for !PM builds is a kernel that is only used in virtual machines,
but we still want to use the PL01x serial port driver in that configuration.

I'd say let's select both PM_GENERIC_DOMAINS and PM on VEXPRESS for now,
(almost) as you have in your patch.

	Arnd