On 20.10.2015 09:26, Alfonso Sanchez-Beato wrote:
> On Mon, Oct 19, 2015 at 3:55 PM, Denis Kenzior <denkenz@gmail.com> wrote:
>
>> Hi Alfonso,
>>
>> On 10/15/2015 11:47 AM, Alfonso Sanchez-Beato wrote:
>>
>>> ---
>>>    plugins/ril.c | 5 +++--
>>>    1 file changed, 3 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/plugins/ril.c b/plugins/ril.c
>>> index e31f24b..2edb0a9 100644
>>> --- a/plugins/ril.c
>>> +++ b/plugins/ril.c
>>> @@ -144,7 +144,9 @@ static void ril_radio_state_changed(struct ril_msg
>>> *message, gpointer user_data)
>>>                          break;
>>>                  default:
>>>                          /* Malformed parcel; no radio state == broken
>>> rild */
>>> -                       g_assert(FALSE);
>>> +                       ofono_error("%s: Bad radio state (%d), exiting",
>>> +                                                       __func__,
>>> radio_state);
>>> +                       exit(1);
>>>
>>
>> Can we just remove the ril modem in question instead of killing the entire
>> daemon?
>
>
> I think that if we start to do these things we need to add modem watches to
> udev/udevng/rildev as they keep lists with the modems and remove them in
> shutdown. Maybe we could try a modem reset here, within a limited number of
> retries instead?

We should avoid exiting in any case. The minimum we can do is to bring 
the whole modem down if we can't do anything more so we end up in the 
same state as we would when restarting ofono and we're about to setup 
the rilmodem again.

regards,
Simon