Re: ZTE MF192 3G Modems

[Brian Gitonga Marete <marete toshnix com>]

On Fri, Jul 20, 2012 at 3:35 PM, Aleksander Morgado
<aleksander lanedo com> wrote:
>
>>>>>>> On Mon, Jul 9, 2012 at 9:25 PM, Brian Gitonga Marete <marete toshnix com> wrote:
>>>>>>>> Hello all,
>>>>>>>>
>>>>>>>> I have the same (or very similar) ZTE MF192 modem referred to by
>>>>>>>> Dmitry Morozhnikov here:
>>>>>>>> https://mail.gnome.org/archives/networkmanager-list/2012-February/msg00014.html
>>>>>>>> My modem is supplied by Orange Kenya, for use in their Orange Internet
>>>>>>>> 3G service.
>>>>>>>>
>>>>>>>> In summary, it also does NOT work for me with MM/NM out of the box.
>>>>>>>> What is more, the patch suggested by Dmitry in the mail thread
>>>>>>>> referenced above suffices at least establish an Internet connection
>>>>>>>> that is fully usable.
>>>>>>>>
>>>>>>>> Wvdial works well with this MF192 device, and attached is my wvdial
>>>>>>>> config as wvdial.conf. I have also attached wvdial.log, so that you
>>>>>>>> can see what wvdial uses to successfully drive it.
>>>>>>>>
>>>>>>>> I was also able to capture a USB trace (using Wireshark) of windows
>>>>>>>> driving the modem from within a Virtualbox VM. I think this will allow
>>>>>>>> you to develop full support for this modem. The capture trace reflects
>>>>>>>> all traffic to/from the modem from hotplug to a short time after I
>>>>>>>> successfully "connect" from withing the Windows program. Please see
>>>>>>>> the attachment zte-mf192.cap.xz for this captured trace. View it from
>>>>>>>> Wireshark.
>>>>>>>>
>>>>>>>> Also attached are the files modem-manager.log and network-manager.log,
>>>>>>>> showing the log of a failed connection attempt (before applying the
>>>>>>>> patch referenced above). Please also see lsusb.txt for USB ID details
>>>>>>>> about my modem.
>>>>>>>>
>>>>>>>> One last thing: I can confirm that, as in Dmitry case, the AT commands
>>>>>>>> ZPAS? and ZSNT=1,0,0 do not seem to work on this modem. (And, in fact,
>>>>>>>> I don't see them in the Windows trace (zte-mf192.cap.xz) which
>>>>>>>> contains many AT commands to the modem.
>>>>>>>>
>>>>>>>> If I had a better idea about the structure/architecture of MM, I might
>>>>>>>> be able to help better, with actual code :)
>>>>>>>>
>>>>>>>> Please let me know if you need any further information.
>>>>>>>>
>>>>>>>> Thanks!
>>>>>>
>>>>>> Hello, Brian!
>>>>>>
>>>>>> Recently, I wrote to this list about ZTE MF192. I got it to work with
>>>>>> dhclient, see my previous messages:
>>>>>> https://mail.gnome.org/archives/networkmanager-list/2012-June/msg00003.html and https://mail.gnome.org/archives/networkmanager-list/2012-April/msg00043.html
>>>>>>
>>>>>> Please, can you try my patches, and confirm them to work, or not.
>>>>>
>>>>> Hello Alexander,
>>>>>
>>>>> Thanks for your email!
>>>>>
>>>>> Unfortunately, the MF192 I have does not present with a
>>>>> pseudo-ethernet interface. Just the ttyACM* interfaces.  Dmitry
>>>>> Morozhnikov patch works for me. Do you have a suggestion that can help
>>>>> me expose a pseudo-ethernet interface if the device has one? Also,
>>>>> note that my device has the ID: 19d2:1519, and the latest
>>>>> usb_modeswitch data packages switches it to a device with ID:
>>>>> 19d2:1519.
>>>>
>>>> Hello Alexander,
>>>>
>>>> Sorry, I meant to say that my MF192 has ID 19d2:1517 initially and is
>>>> switched to a device with ID 19d2:1519 by usb_modeswitch.
>>>>
>>>> Sorry about the confusion!
>>>
>>>
>>> Hello all,
>>>
>>> With regard to the above, is Dmitry's patch acceptable, or does it
>>> need some work (I am happy to do such work).
>>>
>>
>> Dmitry's patch forces this specific modem to get connected with a
>> generic ATD call and PPP. But the modem supports Icera commands, and it
>> should better get connected using the built-in wwan interface. The
>> problem seems to be that the kernel doesn't export any wwan interface
>> here, so there's some work to be done likely there. Once the kernel
>> exports the wwan interface, I guess we should be able to connect with
>> the Icera subset of commands nicely.
>>
>> Dan, in the mean time, why don't we just check in _is_icera() if we have
>> a net port grabbed and if we don't, just warn and go on with the
>> Icera-less implementation?
>>
>
> Gah, I should have checked the repo before. This should already be
> handled in the following commit:
>
> http://cgit.freedesktop.org/ModemManager/ModemManager/commit/?h=MM_05&id=7dd42f1982720dd666bc4d7a4920f033bcf64662
>
> Brian: can you try to run the latest code from the MM_05 branch to see
> if it works for you?
>
 Ok. Will try this one. I should add that I have a full Wireshark USB
Capture of Windows driving the device. Windows uses AT commands over
the serial port(s) and starts PPP eventually. I can't attach this
trace here because it's too large. Can I send it directly to you?

-- 
Brian Gitonga Marete
Toshnix Systems - http://toshnix.com