| Message ID | 20241008091009.1567305-1-sean@geanix.com (mailing list archive) |
|---|---|
| State | Changes Requested |
| Headers | show |
| Series | [RFC] atmodem: let +COPS: report access technology | expand |
Hi Sean, On 10/8/24 4:10 AM, Sean Nyekjaer wrote: > When using the SIMCom A7672E-FASE in gsm mode, gsm is never reported as > access technology. Fix this by letting +COPS: report it. > --- > For EUTRAN networks at_creg_cb() is setting the correct technology. > But as I read it, CREG doesn't differentiate between GSM and UMTS. > Can we fix this by using the technology from +COPS: ? > > drivers/atmodem/network-registration.c | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/drivers/atmodem/network-registration.c b/drivers/atmodem/network-registration.c > index 46b8512a..ccd01234 100644 > --- a/drivers/atmodem/network-registration.c > +++ b/drivers/atmodem/network-registration.c > @@ -353,6 +353,9 @@ static void cops_cb(gboolean ok, GAtResult *result, gpointer user_data) > op.status = 2; > op.tech = tech; > > + /* Set current access technology */ > + nd->tech = tech; > + Maybe behind a vendor check with an explanation? COPS isn't really meant to be used for this as it can contain the cell capabilities rather than the actual registration technology. > DBG("cops_cb: %s, %s %s %d", name, nd->mcc, nd->mnc, tech); > > cb(&error, &op, cbd->data); Regards, -Denis
Hi Denis, On Wed, Oct 09, 2024 at 10:47:33AM +0100, Denis Kenzior wrote: > Hi Sean, > > On 10/8/24 4:10 AM, Sean Nyekjaer wrote: > > When using the SIMCom A7672E-FASE in gsm mode, gsm is never reported as > > access technology. Fix this by letting +COPS: report it. > > --- > > For EUTRAN networks at_creg_cb() is setting the correct technology. > > But as I read it, CREG doesn't differentiate between GSM and UMTS. > > Can we fix this by using the technology from +COPS: ? > > > > drivers/atmodem/network-registration.c | 3 +++ > > 1 file changed, 3 insertions(+) > > > > diff --git a/drivers/atmodem/network-registration.c b/drivers/atmodem/network-registration.c > > index 46b8512a..ccd01234 100644 > > --- a/drivers/atmodem/network-registration.c > > +++ b/drivers/atmodem/network-registration.c > > @@ -353,6 +353,9 @@ static void cops_cb(gboolean ok, GAtResult *result, gpointer user_data) > > op.status = 2; > > op.tech = tech; > > + /* Set current access technology */a So something like: if (nd->vendor == OFONO_VENDOR_SIMCOM_A76XX) ? > > + nd->tech = tech; > > + > > Maybe behind a vendor check with an explanation? COPS isn't really meant to > be used for this as it can contain the cell capabilities rather than the > actual registration technology. And an explanation from the SIMCom datasheet? > > > DBG("cops_cb: %s, %s %s %d", name, nd->mcc, nd->mnc, tech); > > cb(&error, &op, cbd->data); > > Regards, > -Denis /Sean
Hi Sean, > > So something like: > > if (nd->vendor == OFONO_VENDOR_SIMCOM_A76XX) ? >>> + nd->tech = tech; >>> + Yep. >> >> Maybe behind a vendor check with an explanation? COPS isn't really meant to >> be used for this as it can contain the cell capabilities rather than the >> actual registration technology. > > And an explanation from the SIMCom datasheet? > Yeah, that would be great. Regards, -Denis
diff --git a/drivers/atmodem/network-registration.c b/drivers/atmodem/network-registration.c index 46b8512a..ccd01234 100644 --- a/drivers/atmodem/network-registration.c +++ b/drivers/atmodem/network-registration.c @@ -353,6 +353,9 @@ static void cops_cb(gboolean ok, GAtResult *result, gpointer user_data) op.status = 2; op.tech = tech; + /* Set current access technology */ + nd->tech = tech; + DBG("cops_cb: %s, %s %s %d", name, nd->mcc, nd->mnc, tech); cb(&error, &op, cbd->data);