problems with huawei E3372 alias Telekom Speedstick V

[Bjørn Mork]

Thomas Schäfer <tschaefer at t-online.de> writes:

> Hi, 
>
> ** (file-roller:8837): WARNING **: Couldn't register with accessibility bus: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.
>
> I made some tests again.
>
> First of all I noticed, windows uses the "normal" at-channel, without ncm-
> Header. So I tried it again with /dev/ttyUSB0 instead of cdc-wdm.
>
> I found also the explanation why my first test with kernel 3.11/modemmanager 
> has failed. mmcli did claim to be connected, but mm has tried to dial 
> *99#...for ppp-mode  instead of ndisdup. The LED of the device did not lie.
>
> After using minicom (with ndisdup) dhcpd got an IP-address via wwan0. Here the 
> story ends - no real connection. But the behavior was repeatable, so I made an 
> usb-trace.
> Please have a look on it. Maybe it is a mac-address-mess again or something 
> similar.

So this works better with 3.11 than with 3.17?  That's not good... 

But at least it shows us that this modem does not have any problems with
the 16bit headers.  There DHCP offer and ack are nicely wrapped as 16bit
versions.

And it also seems to handle our failure to align the packets as
requested. I'll look further into that - I suspect we never really have
done this correctly.  Or I might have messed it up at some point.  I've
never had access to any test device requesting any remainder != 0 so I
could have broken it without noticing.

But in any case: This modem deals with it, so that's not why we have
problems.  The problem seems to be a lack of replies to the ARP requests
we send.  You could try turning off ARP on the wwan0 interface and see
if that helps.  Something like "ifconfig wwan0 -arp" should do.

I do wonder if that maybe should be the default for the huawei_cdc_ncm
driver?

But this still doesn't answer why we don't get any DHCP replies in
3.17.  Could explain why you couldn't make the connection work with
static addressing though, so it might be worth testing.


Bjørn