Hi Bernhard,

>> In theory you can use $GPSNMUN to enable GPS reporting on Ports 1 and 4
>> which seem to be fully functional.  However, it does look like on my
>> UC864-G, port #3 is used for the GPS stream.  Port #1 doesn't seem to
>> respond to anything.
>>
> 
> I just tried the following at my modem:
> 
> AT$GPSNMUN=1,1,1,1,1,1,1
> AT$GPSP=1
> 
> Which enables the GPS reporting on Ports 1 and 4 and adds $GPSNMUN: at
> the beginning of every GPS report e.g.
> 
> $GPSNMUN: $GPVTG,,T,,M,,N,,K*4E
> 
> Which would be easy to catch from ofono.
> 
> In my opinion the problem about this is that the GPS atom needs to be
> loaded for every telit modem, then try AT$GPSNMUN=1,1,1,1,1,1,1 and wait
> for the response. If the response is an error the atom knows that the
> modem does not support GPS. Could this check/enable also be added to a
> part of the telit plugin?
> 
> The udev approach has the benifet to be able to do
> 
> if (HasGPS)
>     ... = ofono_location_reporting_create(...);
> 
> or something like that at the telit plugin.
> 

I agree that your proposed approach is the right one.

>>
>> Yes and no.  ofono_modem_create has two modes of operation, one where a
>> known unique id is given.  This can come for example from the serial
>> number of the USB device.  Some manufacturers provide this information
>> properly, others do not.  If you cannot trust the USB serial ID, then
>> giving a NULL parameter to ofono_modem_create will tell oFono to
>> generate a unique id itself.  Your path will then look like /telit2 or
>> something like that.
> 
> Ah, ok then the patch is fine I think and I like to get it reviewed :)
> 
> But I can alter the patch to only search for the modem device and add
> the GPS check later to the telit plugin or the GPS atom, if you like.
> 

Patch has been reviewed, however we might have to examine whether we
need to tweak the udev rules for this serial number / modem path stuff
to work properly.  Marcel will have to take a look as I'm lost in that
part of the udev plugin.

Regards,
-Denis