Huawei me906s-158

[Bjørn Mork]

Andreas Fett <andreas.fett at secunet.com> writes:
> On 12/04/16 14:13, Bjørn Mork wrote:
>> Andreas Fett <andreas.fett at secunet.com> writes:
>>> System (cold) boots with /sys/bus/usb/devices/1-2/bConfigurationValue of
>>> "2".
>>>
>>> echo 0 > /sys/bus/usb/devices/1-2/bConfigurationValue
>>> echo 3 > /sys/bus/usb/devices/1-2/bConfigurationValue
>> 
>> Hmm, rereading this while trying to make a case out of it, I realize
>> that I missed the fact that you need to switch to configuration 0
>> (i.e. unconfigured).  For some reason I thought you had to switch to
>> configuration 1.  But that doesn't work either, does it?
> No. It always starts with "2" and we have to go back to "0" and  then to
> "3" to get MBIM working. Switching to "1" and then to "3" (with or
> without delay) does not work.
>
>> Then I guess we have a problem, because AFAICS the kernel will have to
>> make some decision by default.  We could in theory make the
>> configuration switching always go via the unconfigured state, but adding
>> some arbitrary delay between the states is out of the question.  That's
>> better left for userspace then.
> That I understand, and I don't even like the arbitrary part in
> userspace. Is there no observable feedback whatsoever when the modem
> has completed the configuration switch?
>
>>> echo Y > /sys/class/net/wwan0/cdc_mcn/ndp_to_end
>> 
>> A device quirk for this one is simpler though.  Just one question to be
>> sure: It does not work without this?
> It doesn't. The probing works and even the connect and reporting of
> address info, but I can't use it for actual traffic.

Thanks for confirming.  I'll prepare a device quirk patch for the last
issue then.

Don't know what to do about the first problem.  Leaving it for userspace
for now.  The best I can think of is adding a hint to the
bConfigurationValue documentation in
https://www.kernel.org/doc/Documentation/ABI/stable/sysfs-bus-usb

But I don't think anyone ever reads that...



Bjørn