[PATCH 2/5] drivercore: Add driver probe deferral mechanism

[Ming Lei]

On Fri, Oct 14, 2011 at 12:04 AM, Alan Stern <stern at rowland.harvard.edu> wrote:
> On Thu, 13 Oct 2011, Ming Lei wrote:
>
>> Hi,
>>
>> On Thu, Oct 13, 2011 at 10:31 PM, Alan Stern <stern at rowland.harvard.edu>
>> >> Maybe we should understand the correct model of the ordering constraints
>> >> for the multiple device dependancies first, could you give a description or
>> >> some examples about it?
>> >
>> > The requirement is that devices must be capable of resuming in the
>> > order given by dpm_list, and they must be capable of suspending in
>> > the reverse order.
>> >
>> > Therefore if device A can't work unless device B is functional, then B
>> > must come before A in dpm_list.
>>
>> If all devices can support async suspend and resume correctly, looks like
>> the device order given by dpm_list is not needed any longer, doesn't it?
>
> It _is_ needed, because the user can disable async suspend/resume via
> /sys/power/pm_async.
>
> Also, not all devices do support async suspend/resume.

Basically, the devices which don't support async suspend/resume
should have out of tree PM dependency. If out of tree PM dependency
issue is solved, all devices can support async suspend/resume.

thanks,
-- 
Ming Lei