[OpenWrt-Devel] Huawei 3372 NCM support on CC not working

[chrono]

> Great docs of what you have tried!

I can't really expect help or a magic solution without giving
detailed info about what I've tried. Besides, since we're all
connected, it might help to fix these issues in general because
I have a device to test and other devs might see something I've
overlooked or simply am not aware of so we may get faster to
a solution that may work for everyone ;)

> There are two more things I can think of that is worth testing:
> 1) set the wwan0 interface up before connecting (AT^NDISDUP):
>      ifconfig wwan0 up
> 2) turn off ARP:
>      ifconfig wwan0 -arp
> 

IIRC I've already mentioned that I've tried the -arp approach
on the openwrt test and also did this on the gentoo box:

wwan0: flags=4291<UP,BROADCAST,RUNNING,NOARP,MULTICAST>  mtu 1500
         inet 10.115.201.248  netmask 255.255.255.240  broadcast 
10.115.201.255
         ether 00:1e:10:1f:00:00  txqueuelen 1000  (Ethernet)
         RX packets 3262  bytes 150052 (146.5 KiB)
         RX errors 0  dropped 0  overruns 0  frame 0
         TX packets 711  bytes 140192 (136.9 KiB)
         TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

> And note that pinging the "gateway" isn't a good test, contrary to
> normal networks.  The "subnet" and "gateway" are faked by the firmware.
> In reality, the gateway address is probably assigned to another user. 
> If
> the firmware does things properly, then it won't answer the ping but
> forward those icmp echo requests to the mobile network like any other
> packet.  And many mobile networks will never forward icmp echo to end
> users, so you will not get a reply even if the address happens to be in
> use by someone.

Thanks for the explanation, I usually use the google dns (8.8.8.8) as
a default test-ping IP, since it's easy to remember and is usually
never offline but since I never got a reply from it my thinking was
to try to reach the next hop (the GW) to see if at least this would
work (to exclude any other upstream routing issues)

> To test this you should find some address which is known to be 
> reachable
> and responding, and add a route to that (or a default route). I.e.
>  ip route add 8.8.8.8/32 via 10.115.201.241

8.8.8.8         10.115.201.241  255.255.255.255 UGH   0      0        0 
wwan0

# ping 8.8.8.8
--- 8.8.8.8 ping statistics ---
30 packets transmitted, 0 received, 100% packet loss, time 28999ms

> If you turn off arp then you can simply ignore the gateway address and
> refer to the netdev instead:
>   ip route add 8.8.8.8/32 dev wwan0

8.8.8.8         0.0.0.0         255.255.255.255 UH    0      0        0 
wwan0

# ping 8.8.8.8
--- 8.8.8.8 ping statistics ---
22 packets transmitted, 0 received, 100% packet loss, time 20999ms

The thing that puzzles me most is that either I'm totally doing 
something
wrong here or we're fighting some kind of regression, since people in
austria have obviously had success before, which influenced my choice
to get this device, since it's so simple to get online with it... NOT :)

http://www.lteforum.at/mobilfunk/huawei-e3372-openwrt.1739/#post-26074

There also seems to be an external driver package from huawei,
but I'd rather have this fixed natively and open-source, without
having to rely on binary driver packages from huawei.
_______________________________________________
openwrt-devel mailing list
openwrt-devel at lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel