Converting to dev_pm_ops (remove annoying warnings)

Converting to dev_pm_ops (remove annoying warnings)

[Ben Dooks]

I have a couple of questions about changing drivers to use the
dev_pm_ops method within drivers using the device driver model.

1) Should the suspend/resume code still be wrapped in #ifdef CONFIG_PM
   and supply a NULL 'struct dev_pm_ops' wen CONFIG_PM=n

2) Is it allowable to submit changes to move drivers to dev_pm_ops
   during the -rc phase, or do we have to wait for the merge window
   before sending these? The number of warnings on a system that
   relies heavily on platform devices is annoying (S3C24XX ARM).

-- 
Ben (ben@fluff.org, http://www.fluff.org/)

  'a smiley only costs 4 bytes'

Re: Converting to dev_pm_ops (remove annoying warnings)

[Magnus Damm]

Hi Ben!

On Tue, Jul 28, 2009 at 7:44 PM, Ben Dooks<ben-linux@fluff.org> wrote:
> I have a couple of questions about changing drivers to use the
> dev_pm_ops method within drivers using the device driver model.
>
> 1) Should the suspend/resume code still be wrapped in #ifdef CONFIG_PM
>   and supply a NULL 'struct dev_pm_ops' wen CONFIG_PM=n
>
> 2) Is it allowable to submit changes to move drivers to dev_pm_ops
>   during the -rc phase, or do we have to wait for the merge window
>   before sending these? The number of warnings on a system that
>   relies heavily on platform devices is annoying (S3C24XX ARM).

To begin with: I'm the one to blame for the annoying printouts. If
people dislike the printout then I'm all for disabling it later on in
the -rc phase. As you understand the goal is to motivate people to
update their drivers. =)

I'm not the right person to answer the merge policy question, but
wouldn't it be good to submit patches regardless of when they will be
picked up? OTOH, I guess knowing when helps if you have a bunch of
other changes queued up for your drivers though...

Any thoughts?

Cheers,

/ magnus