Re: [PATCH/RFC] ARM: mach-shmobile: sh7372 power domain prototype

[Magnus Damm <magnus.damm@gmail.com>]

Hi Rafael,

2011/4/12 Rafael J. Wysocki <rjw@sisk.pl>:
> Hi,
>
> To start with, I think we'll need to do that a bit differently, because
> apparently some clock management code may be shared with OMAP.
>
> Most imprtantly, the PM core is going to be changed so that power domain
> callbacks will be executed instead of and not in addition to the bus type
> callbacks, so platform_pm_runtime_suspend() and platform_pm_runtime_resume()
> will have to be replaced by the power domain callbacks entirely.
>
> I'm working on those changes right now, but I'd like to make the core
> changes first and get everybody involved to agree with them.

Thanks for your help!

> Some more comments below.

[snip]

> The above things don't look quite right.  I think the connections between
> device objects and power domain objects should be set up in the board file.

Yes, no doubt about that!!!

> This doesn't look nice to me.

I have to agree with you on that one. =)

This prototype code is rather filthy and can easily be seen as one big
layering violation.

However, just so you understand the reason behind this:

I simply posted the patch to publish _some_ code that exercises the
Runtime PM callbacks in combination with a bit of power domain code.
This so anyone with sh7372 can do some simple testing of drivers using
Runtime PM.

I'd like to see the drivers hardened in parallel with Runtime PM power
domain support. When we have something nicer we will of course switch
to using that. Feel free to post a better version that is properly
abstracted. =)

Thanks,

/ magnus