edward shu wrote:
>
>>>>
>>>
>>> srn is for userland consumption. The primary difference between
>>> srn and SIGTHAW, is that it provides syncronous access, so that it
>>> can also be used for suspend notification. If all you care about is
>>> a notification that resume is complete, SIGTHAW is sufficient. If
>>> you also want suspend notification and/or require a mechanism for
>>> notifying the framework of suspend or resume errors, then srn may be
>>> a better choice.
>>
>> We only need to perform work on resume, so it sounds like SIGTHAW is
>> reasonable.
> I am worried that there are some cards that require special device
> context saving in the suspend process.
> And they are required to reset the original video mode in the resume
> process.
All that will happen before getting to the point where the SIGTHAW is
delivered. The card is supposed to do the POST (and restoring) of the
card all by itself and we won't interfere at all with it.
When the SIGTHAW is delivered we are in the position of checking if
X.org is up (in which case we basically do nothing) or not (in which
case we push down to vbiosd a request to reset the saved VESA mode).
- Enrico
>
>>
>> -jan
>
