Cinterion plugin - unsupported ^SIND="simstatus"

[Aleksander Morgado]

On Sat, Dec 30, 2017 at 4:30 PM, Colin Helliwell
<colin.helliwell at ln-systems.com> wrote:
>
>> On 30 December 2017 at 15:07 Colin Helliwell wrote:
>>
>>
>> The Cinterion plugin tries 'AT^SIND="simstatus",2' in after_sim_unlock(). I have two Cinterion modems, neither of which - according to their AT Command Set spec - support the simstatus indicator on this command, and so instead return "+CME ERROR: 21" (invalid index).
>> Nonetheless, the operation continues with all 15 of its retries. Should there perhaps be a 'if (!response)' trap in simstatus_check_ready() so that it bombs out?
>> I ask for comment, since I'm not sure (a) exactly what the retries are waiting for; and, in particular (b) what the response would be while its waiting. (I don't have a modem which does support the indicator).
>>
>
> I imagine not even getting a good response can be treated as unsupported, let alone needing to parse it for ^SIND content . Hence
>
>
>  plugins/cinterion/mm-broadband-modem-cinterion.c | 99 ++++++++++++++++++++----
>  1 file changed, 6 insertions(+), 1 deletions(-)
>
> diff --git a/plugins/cinterion/mm-broadband-modem-cinterion.c b/plugins/cinterion/mm-broadband-modem-cinterion.c
> index 64c5e08..1034d5c 100644
> --- a/plugins/cinterion/mm-broadband-modem-cinterion.c
> +++ b/plugins/cinterion/mm-broadband-modem-cinterion.c
> @@ -1590,7 +1629,12 @@ simstatus_check_ready (MMBaseModem  *self,
>      const gchar           *response;
>
>      response = mm_base_modem_at_command_finish (self, res, NULL);
> -    if (response) {
> +    if (!response) {
> +        /* "simstatus" not supported, go on anyway */
> +        g_task_return_boolean (task, TRUE);
> +        g_object_unref (task);
> +        return;
> +    } else {
>          gchar *descr = NULL;
>          guint val = 0;

Does the modem support AT^SIND at all? Or just the "simstatus"
indication isn't supported in the command?
As with other things, the best way to maintain this upstream is to try
to query first for the support of the specific feature and, if
unsupported, ignore it from then on. In this case the support checks
would be for either the whole command (e.g. AT^SIND=?) or for the
actual "simstatus" indication (e.g. looking for "simstatus" in
AT^SIND=? if that makes sense).


-- 
Aleksander
https://aleksander.es