Re: [PATCH 4/6] src/modem: notify lte driver before going online

[Giacinto Cifelli <gciofono@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 2673 bytes --]

On Fri, Oct 12, 2018 at 8:17 AM Jonas Bonn <jonas@southpole.se> wrote:
>
>
>
> On 12/10/18 05:04, Giacinto Cifelli wrote:
> > Hi,
> >
> > On Thu, Oct 11, 2018 at 10:19 PM Jonas Bonn <jonas@southpole.se> wrote:
> >>
> >>
> >> On 10/10/18 08:54, Giacinto Cifelli wrote:
> >>> ---
> >>>    src/modem.c | 2 ++
> >>>    1 file changed, 2 insertions(+)
> >>>
> >>> diff --git a/src/modem.c b/src/modem.c
> >>> index 9e254482..74dbe7ad 100644
> >>> --- a/src/modem.c
> >>> +++ b/src/modem.c
> >>> @@ -729,6 +729,8 @@ static void sim_state_watch(enum ofono_sim_state new_state, void *user)
> >>>        case OFONO_SIM_STATE_READY:
> >>>                modem_change_state(modem, MODEM_STATE_OFFLINE);
> >>>
> >>> +             ofono_lte_set_reg_info(modem);
> >>> +
> >> modem_change_state(...OFFLINE) results in the post_sim() implementation
> >> being called.  The drivers that implement the lte atom all call
> >> ofono_lte_create() there.
> >>
> >> Your .._set_reg_info function checks for the existence of the atom and
> >> does some work.  As such, why not just merge the contents of
> >> ofono_lte_set_reg_info() into the at_lte_probe function instead... from
> >> there, it's more obvious what's going on.
> > The point is that the function must be called before set_online
> > (following), and not after post_sim, even if they happen to be
> > together.
>
> set_online only happens in the following line if the modem is "always
> online", which is generally not the case.  For other modems, set_online
> happens elsewhere.
>
> > This is how it works for gprs-context: before activating the context,
> > apn and auth are set.
> >
> > It isn't so obvious to me that the probe function does the work. It is
> > in general the first one to be called, then there is a chance to set
> > the properties one by one, and eventually they are transferred to the
> > modem.
>
> Between "enable" and "online", you have a window to change the settings.
>
> For modems that are "always online", you pretty much have to go
> "enabled/online", perhaps failing to do so(?), change the settings,
> "disable" and "reenable" the modem.  Ugly, yes, but that's the way
> things look right now.
>
> Twiddling the LTE settings just results in context reconfiguration
> (CGDCONT)... this, presumably, does not have any effect on any already
> established context... correct me if I'm wrong.  As such, the changes
> don't really take effect until you've gone offline/online again, anyway.
>
> /Jonas

I have decided to go for the direct application. I am changing the code now.
This part is therefore no longer applicable.

Giacinto