Hi Marcel,

>> +
>> +/*
>> + * Callback from provisioning plugin.
>> + * settings: list of struct ofono_gprs_provisioning_data
>> + *
>> + * It is responsibility of callback function to free settings-list
>> + * settings-list elements must be freed with ofono_gprs_provisioning_data_free()
>> + */
>> +typedef void (*ofono_gprs_provision_cb_t)(GSList *settings, void *userdata);
>> +
>> +struct ofono_gprs_provision_driver {
>> +	const char *name;
>> +	int priority;
>> +	void (*get_settings) (struct ofono_modem *modem,
>> +				ofono_gprs_provision_cb_t cb,
>> +				void *userdata);
>> +};
> 
> So here is something we need to talk about. The nettime plugin
> infrastructure using a "pseudo" atom. So do we wanna do the same here or
> do we wanna change the nettime atom to something simple like this.
> 
> I like to have consistency here. Aki, Denis, thoughts?
> 

The nettime and history plugins can create their own D-Bus interfaces.
Hence the need for probe and remove.  Since the provisioning driver
might want to pre-parse its database, I believe it should follow the
same approach...

Regards,
-Denis