Maybe a little bug inside "serial.c" driver file:
Line 666 : "sizeof addr" gives 4 on my side ?
I had to replace it with something that give 8 ie: IEEE802154_ALEN
// memcpy(addr, zbdev->data, sizeof addr);
memcpy(addr, zbdev->data, IEEE802154_ALEN);
Le 04/11/2010 14:20, [email protected] a écrit :
> I was on 64b linux
>
> On Thu, Nov 4, 2010 at 9:18 AM, [email protected]<[email protected]> wrote:
>> I'm at Linux Plumber's Conference for the rest of the week.
>> http://www.linuxplumbersconf.org/
>> I don't have access to the hardware.
>>
>> I had applied my patches on top of the linux-zigbee kernel tree. Is
>> that the base kernel you are using?
>>
>> I was running the wireshark directly on the 6lowpan interface. The
>> crash was on receives. I was unable to ping.
>>
>> On Thu, Nov 4, 2010 at 9:06 AM, Pierre-emmanuel Goudet
>> <[email protected]> wrote:
>>> Le 04/11/2010 12:38, [email protected] a écrit :
>>>> On Thu, Nov 4, 2010 at 7:31 AM, Pierre-emmanuel Goudet
>>>> <[email protected]> wrote:
>>>>> Hi Jon,
>>>>>
>>>>> I've made a try using two interfaces up on my serial device.
>>>>> One with your 6LowPan filter activated the other without.
>>>>> The sender is a device talking IPV6/RPL/6LowPan with a first DIO
>>>>> brodcasted.
>>>>>
>>>>> What i can see that could be some clues for your implementation is that
>>>>> the
>>>>> full 802.15.4 header as been modified (deleted) by 6LowPan activated
>>>>> interface, then my Wireshark cans no more decode the frame ?
>>>> I didn't fully implement the code sorting out long vs short addresse
>>> Do you think it's the reason why, quite all the 802.15.4 header is cleared
>>> (set to 0) ?
>>>>> Nevertheless, with this DIO broadcasted frame my Kernel does not crash it
>>>>> receives it on the 6LowPan activated interface ?
>>>> It will only crash if you run wireshark and watch the packets.
>>> I was running 2 wiresharks (1.5.0) instances watching packets, on my linux
>>> at this moment.
>>> One on each interface, the 6lowpan and the raw.
>>> But i forgot to join the two PCAP files. Here they are now.
>>>
>>> Notice that I've tarted from 2.6.35 kernel before patching with your code,
>>> and maybe the type of frame broadcast or not etc, long/ short ... is the
>>> reason of the crash on your linux ?
>>> What kind of frame (802.15.4/6Lowpan/IPV6) do you send when you see your
>>> kernel crashing ?
>>>>> Any idea ?
>>>>>
>>>>> Best regards
>>>>>
>>>>> Pierre-emmanuel
>>>>>
>>>>>> <[email protected]> wrote:
>>>>>>> Hi Jon, Dimitri, all,
>>>>>>>
>>>>>>> And many thanks.
>>>>>>>
>>>>>>> I tried all this yesterday.
>>>>>>> Great ! it works. I can see 6LowPan frames sended.
>>>>>>>
>>>>>>> Basic steps to get this working for whom who begin with all this (a
>>>>>>> bit
>>>>>>> like me) :
>>>>>>> =====================================================================
>>>>>>> * Create a serial Node with an 802.15.4 interface conform to Dimitri
>>>>>>> specification (Cf Linux-zigbee Wiki),
>>>>>>> and use an other node with Contiki IPV6 capabilities. Not necessary
>>>>>>> during
>>>>>>> these first tests except for RS/RA behaviour.
>>>>>>> * Get the git devel branch of linux-zigbee, (Starting from a debian
>>>>>>> dist
>>>>>>> at
>>>>>>> least 2.6.32)
>>>>>>> * Apply your (Jon smirl) patches (Kernel and lowpan), (use git-apply
>>>>>>> with
>>>>>>> your patch)
>>>>>>> * Compile the linux-zigbee kernel and lowpan tools
>>>>>>>
>>>>>>> * Connect your serial node ad use these commands to test :
>>>>>>> # izattach /dev/ttyUSBX
>>>>>>> Find the new wpan-phy with # iz list-phy
>>>>>>> # iz 6lowpan wpan-phyY
>>>>>>> # ifconfig wpan0 up
>>>>>>>
>>>>>>> (Use a sniffer Cf 15dot4 project from Colin O'Flyn to see frames on the
>>>>>>> real
>>>>>>> media with wireshark more recent
>>>>>>> than 1.3.3 the best would be 1.5.0 not my case today ...)
>>>>>> Run wireshark on your Linux box and capture from the network
>>>>>> interface. First packet received will crash the kernel. That's the bug
>>>>>> I never got sorted out.
>>>>>>
>>>>>>
>>>>>>> My current questions:
>>>>>>> ===================
>>>>>>> - My interface Hardware address is not setted correctly ... it seems
>>>>>>> that
>>>>>>> just 4 or 5 first bytes are setted ?
>>>>>>> (even so i see the correct address in the return "zb0x8d..." message
>>>>>>> with
>>>>>>> #dmesg )
>>>>>> I think Dmitry and I implement incompatible ways to set the address.
>>>>>> This needs to get sorted out to something that works.
>>>>>>
>>>>>>
>>>>>> - As discussed in previous posts in linux-zigbee mailing list,
>>>>>>> I had to have my channel setted (done as a hack of
>>>>>>> mac802154_wpan_xmit()
>>>>>>> from net/mac802.15.4/wpan.c) to see my packets sended.
>>>>>> there's no good UI for changing the channel at the moment. I just
>>>>>> change it in the source code.
>>>>>>
>>>>>>> - Without defining a global address this ping line
>>>>>>> # ping6 fe80::ff:ff00:1 -I wpan0 -c1
>>>>>>> gives a MAC header containing my own MAC address as source AND dest ?
>>>>>>> (seen
>>>>>>> on the sniffer)
>>>>>>>
>>>>>>> - With a Global IPV6 address defined on the interface
>>>>>>> # ip -6 addr add aaaa::1/64 dev wpan0
>>>>>>> # ping6 aaaa::ff:ff00:1 -I wpan0 -c1
>>>>>>> as got very strange LL addressing and lowpan compression with 16 bit
>>>>>>> source
>>>>>>> address ?
>>>>>> There is some stuff that needs to be sorted out in Contiki. Contiki is
>>>>>> always using 16b addresses in places where long and short address
>>>>>> should both be legal.
>>>>>>
>>>>>>
>>>>>>> For whom who are interested i join the PCAP files for initial RS/RA
>>>>>>> exchanges, and an example of Ping send with Local then global with a
>>>>>>> Contiki
>>>>>>> IPV6 device.
>>>>>>>
>>>>>>> Pierre-emmanuel
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>
>>>>
>>>
>>>
>>
>>
>> --
>> Jon Smirl
>> [email protected]
>>
>
>
------------------------------------------------------------------------------
The Next 800 Companies to Lead America's Growth: New Video Whitepaper
David G. Thomson, author of the best-selling book "Blueprint to a
Billion" shares his insights and actions to help propel your
business during the next growth cycle. Listen Now!
http://p.sf.net/sfu/SAP-dev2dev
_______________________________________________
Linux-zigbee-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-zigbee-devel
