On 04/20/2017 07:03 PM, Denis Kenzior wrote:
On 04/18/2017 02:50 AM, Jonas Bonn wrote:
> + if (strlen(info->apn) > 0)
> + ldd->apn = g_strdup(info->apn);
> + else
> + ldd->apn = NULL;
Why do you want to store the APN inside the driver data?
Just to do _something_ with the APN until the FIXME below gets fixed.
There's no reason to store it in the long run.
> + /*
> + * FIXME: The above is not sufficient. The APN needs to be set
> + * to the 'default' profile so that the modem uses it when
> + * connecting the default bearer.
> + */
> + ofono_warn("LTE APN needs to be written to default profile");
> + qmi_service_create(device, QMI_SERVICE_WDS,
> + create_wds_cb, lte, NULL);
So I talked to Marcel about this and it seems a QMI service cannot be
created multiple times for the same device. So I think this should
fail and you should be using qmi_service_create_shared instead.
OK, thanks for noticing this. I was, in fact, wondering about this, too.
The QMI changes that I have made in the last few weeks have created the
same service in multiple atoms (gprs, network-registration,
gprs-context). That _does_ actually work on my EC21 which is a pretty
new Qualcomm chip. But if older chips have problems with this then I
suppose this will need to be fixed in the other atoms, too. I will do so.