[PATCH 01/35] mfd: ab8500-gpadc: Implemented suspend/resume

[Ulf Hansson]

On 25 February 2013 14:51, Mark Brown
<broonie at opensource.wolfsonmicro.com> wrote:
> On Mon, Feb 25, 2013 at 10:27:36AM +0100, Ulf Hansson wrote:
>> On 22 February 2013 11:38, Mark Brown
>
>> > Are you sure this actually does what you think it does, especially when
>> > run on modern kernels?
>
>> Not sure, what you are thinking of more precisely here. Runtime pm has
>> been in the kernel for quite some time now.
>
> Yes, thanks - I was aware of that.  The integration between runtime and
> system PM has been an area that's had some development though.
>
>> Anyway, to make it a bit clearer, we switch the regulator on/off at
>> the runtime suspend/resume callbacks. We want to take similar actions
>> in device suspend/resume.
>> To accomplish this a pm_runtime_get_sync is done in suspend and vice
>> verse in resume, otherwise you can not safely handle the regulator.
>
> Are you absolutely positive that with modern kernels your get actually
> resumes the device?

Yes, runtime resume is always ok, But you can not runtime suspend the
device, since the device suspend layer prevent this with a
pm_runtime_get_noresume before calling the device suspend callback.

Kind regards
Ulf Hansson