[fso based] Resource Allocation

[Boris Stender]

Hi everyone!

I have some questions concerning the allocation of FSO resources:


1) If the "Display" resource is allocated (e.g. utilizing fsoraw + application) the device will not suspend. Other resources (of course except "CPU") do not show this blocking effect. So why this special handling for "Display"?


Annotation: So the following excerpt of "fsoraw -h" seems to make no sense?!

...
examples:
     fsoraw --resources CPU,Display tangogps
...


2) If solely the "CPU" resource is requested I would except that the Display will blank. This is not the case. Trying to have a pure GPS logger utilizing "fsoraw --resources CPU tangogps" doesn't bring the expected result. For the logger scenario a workaround for me was to enable screen-locking instead, but of course with much higher energy drain than wanted, especially due to unavoidable touchscreen presses inside my pocket.


Annotation: It seems that Neil Brown did some coding around this problem but I haven't tested it on current SHR-U yet. Furthermore I think this non blanking is a general problem (blame me if not).


http://neil.brown.name/blog/20090128025654

http://neil.brown.name/git?p=freerunner;a=blob;f=lock/lock.py;h=9adc658b7081572829f1fc3c866b58b257ccb280;hb=ceb0fdcd0dd60d8c5400bec58df7f0f901665bdc
 

3) Fsoraw from Nicola Ranaldo does a great job. Unfortunately, it does not allow to make some resoure re-allocation for already running applications. Sometimes it may be desired to switch allocation on the fly, without the need to stop/start the program or without the need to override allocations from settings menue. Would it make sense to have some kind of resource manager instead of current fsoraw or should this be done within FSO?


4) One more point concerning "Display" allocation. For me there are three use cases: a) display blanking (logger), b) display with auto-dimming (tangogps while riding my bike) c) display always on (navit). Would it be feasible to pimp fsoraw to have something like this:

ad a) fsoraw -r CPU tangogps
ad b) fsoraw -r CPU,DispDim tangogps
ad c) fsoraw -r CPU,DispNoDim navit



And as I am already writing to the community I've one more question concerning #2^10

5) Why is ti_calypso_deep_sleep = "never" used by default on SHR instead of "auto"? Is there a problem with the recamping detection algorithm/procedure?   


Cheers,
Boris
-- 
Neu: GMX DSL bis 50.000 kBit/s und 200,- Euro Startguthaben!
http://portal.gmx.net/de/go/dsl02