Help on an invalid modem case

[Aleksander Morgado]

Hey,


> my team and I are using for testing purposes an 3G USB modem (Huawei E352)
> that has an unexpected behaviour with ModemManager 1.12.10 and later stable
> versions (1.12.12 and 1.14.0).
> The 3G modem is used as a broadband one and every time it disconnects, it
> encounters AT serial timeouts and is eventually declared as invalid by
> ModemManager then removed from modem list.
>
> Jul 02 08:42:27.226342 tap-00006711 ModemManager[1225]: <debug>
> [1593679347.226334] (ttyUSB1): --> 'ATE0<CR>'
> ...
> Jul 02 08:42:30.453896 tap-00006711 ModemManager[1225]: <debug>
> [1593679350.453872] (ttyUSB1): --> 'ATV1<CR>'
> Jul 02 08:42:33.679831 tap-00006711 ModemManager[1225]: <warn>
>  [1593679353.679807] (tty/ttyUSB1) at port timed out 2 consecutive times
> ...
> Jul 02 08:42:59.421551 tap-00006711 ModemManager[1225]: <error>
> [1593679379.421503] (tty/ttyUSB1) at port timed out 10 consecutive times,
> marking modem '/org/freedesktop/ModemManager1/Modem/0' as invalid
> ...
> Jul 02 08:43:02.667893 tap-00006711 ModemManager[1225]: <debug>
> [1593679382.667880] Modem (Huawei)
> '/sys/devices/pci0000:00/0000:00:1c.0/0000:01:00.0/usb4/4-1' completely
> disposed
>
> Only an external action (unplug/unplug, electrical action) makes it
> reappear, manual scan via mmcli does not.
>
> So I have two questions:
> 1) Is there something in the debug logs that may explain these systematic
> serial timeouts?
> I'm not sure ModemManager is at fault here (like some actions leading to
> unrecoverable errors) or the modem device is simply defective...
>
>
When the ttyUSB being used for data doesn't cleanly disconnect, it may be
because we're using the wrong TTY for data. What's the output of "mmcli -m
0" and the "lsusb -v" for the device?


> 2) When the modem is declared invalid, is a there a way using libmm-glib
> to detect or be notified of such event occurrences?
> The device disappearance at ModemManager and NetworkManager level is an
> obvious one but since it is quite a common event, we would like to know if
> something more specific can be observed.
>
>
The device disappearance from MM would be the best way I'd say.

-- 
Aleksander
https://aleksander.es
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/modemmanager-devel/attachments/20200703/9efbabba/attachment.htm>