Re: [PATCH net-next v6 0/3] The huawei_cdc_ncm driver / E3276 problem

["Bjørn Mork" <bjorn@mork.no>]

Pasi Kärkkäinen <pasik@iki.fi> writes:
> On Fri, Mar 14, 2014 at 10:24:51AM +0100, Bjørn Mork wrote:
>> Pasi Kärkkäinen <pasik@iki.fi> writes:
>> 
>> > # ifconfig wwp0s26u1u5i1
>> > wwp0s26u1u5i1: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
>> >         inet 10.8.219.204  netmask 255.255.255.248  broadcast 10.8.219.207
>> >         inet6 fe80::e5b:8fff:fe27:9a64  prefixlen 64  scopeid 0x20<link>
>> >         ether 0c:5b:8f:27:9a:64  txqueuelen 1000  (Ethernet)
>> >         RX packets 0  bytes 0 (0.0 B)
>> >         RX errors 1493  dropped 0  overruns 0  frame 0
>> >
>> > Note the RX errors.. ? 
>> 
>> 
>> Yes, that is not expected and is quite possibly hiding the explanation.
>> 
>
> Ok, good to know.
>
>> Could you do a usbmon dump to try to capture the contents of the USB
>> buffers we drop here?  See
>> https://www.kernel.org/doc/Documentation/usb/usbmon.txt
>> 
>> Note that the text interface may truncate the buffers too early for the
>> result to be useful.  wireshark/thsark will for example allow you to
>> capture the whole buffer (when snooping on the "usbmonX" interface - not
>> the netdev!)
>> 
>
> I'll try that.. 
>
>> We also have a few debug messages in the driver for different types of
>> errors we drop.  Turning on those might explain something (assuming you
>> have DYNAMIC_DEBUG enabled):
>> 
>>   ethtool -s wwp0s26u1u5i1 msglvl rx_err on
>> 
>> Note that these will be logged at "debug" level, so you need to turn on
>> logging/showing such messages as well.
>> 
>
> # uname -a
> Linux localhost.localdomain 3.13.6-200.fc20.x86_64 #1 SMP Fri Mar 7 17:02:28 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux
>
> # grep -i DYNAMIC_DEBUG /boot/config-3.13.6-200.fc20.x86_64
> CONFIG_DYNAMIC_DEBUG=y
>
> # echo 8 > /proc/sys/kernel/printk
> # cat /proc/sys/kernel/printk
> 8       4       1       7
>
> # ethtool -s wwp0s26u1u5i1 msglvl rx_err on
>
> Is that enough?

I believe so, but I don't feel too steady on all the dynamic-ness here.

> I don't see any errors logged in "dmesg" output,
> but the RX errors counters do keep increasing in ifconfig output.. 

OK, let's see if we can figure it out based on the usbmon data.




Bjørn