Re: [PATCH 5/8] lte: activate default bearer on network registration

[Jonas Bonn <jonas@southpole.se>]

[-- Attachment #1: Type: text/plain, Size: 2054 bytes --]

Hi Denis,

On 04/13/2017 05:14 PM, Denis Kenzior wrote:
> I haven't tested this, but I suspect that calling "Start network" for
>>> the same profile again will fail...??? A profile number corresponds
>>> to a Context ID...???  Maybe...
>>
>
> I would imagine that a profile has the same meaning as the 27.007 
> equivalent of +CGDCONT <cid>

Yes, it does.  Others have tested modifying context parameters with AT 
commands and the changes are visible in the corresponding QMI profile.  
However, this could theoretically be firmware dependent as it's not 
actually specified anywhere that it must be so.

>
>> I tested this and "start network" fails when attempting to connect a
>> profile a second time.
>>
>
> Makes sense.  Do note that the same APN can be (in theory) double, 
> triple, etc activated.  But this requires a separate cid for each 
> instance.

Yes, exactly.  Multiple profiles with identical parameters can be 
activated, as far as I can tell.
>>> Given all of the above, for QMI this means:
>>>
>>> i)  When the netreg atom detects registration to an LTE network then
>>> we can call "Start network" immediately.
>
> gprs atom driver needs to tell ofono core that a context is active. 
> This means gprs driver should call ofono_gprs_cid_activated
>
>>> ii)  Start network will give us the packet-handle for the default
>>> bearer.  This can be held by the netreg atom.
>
> why netreg atom?  This should be managed by gprs-context.

Yes, you're right.  I also came to this insight after writing this email! :)

I originally implemented this as described above but have now posted a 
new series that works as you described.  It really hasn't been easy to 
understand how this all was supposed to work but what I've got now seems 
reasonable and functions well.

One question:  ofono_gprs_notify_bearer()... why do we need this when 
netreg already provides the technology to the GPRS atom?

Thanks for taking the time to help me work through all of this! Much 
appreciated.

/Jonas