[riot-devel] Border-router and multicast

Baptiste Clenet bapclenet at gmail.com
Fri May 13 20:45:36 CEST 2016


2016-05-12 21:03 GMT+02:00 Baptiste Clenet <bapclenet at gmail.com>:
> Thanks Martine, I will try tomorrow
>
> 2016-05-12 14:10 GMT+02:00 Martine Lenders <mail at martine-lenders.eu>:
>> Hi,
>>
>> Ah the problem seems to be that for some reason Linux elects a link-local
>> address for the ping, which is of course due to the fact, that there is no
>> global address assigned to the interface and that the hop limit seems to be
>> set to short (it is 1). So I did the following
>>
>> sudo ip route add ff1e::/16 dev tap0 table local  # add routing entry for
>> ff1e::1
>> sudo ip addr add affe::dc53:7dff:fe99:516e/64 dev tap0 # add global unicast
>> address for tap0
>> ping6 -t 2 ff1e::1 # ping with hop limit 2
>>
I did try but ping6 does not receive anything
Did you do any more set up?
What is this address dc53:7dff:fe99:516e?

#ip -6 route list table local
...
ff1e::/16 dev tap0  metric 1024
...

#ifconfig tap0
tap0      Link encap:Ethernet  HWaddr 26:C5:07:2F:43:E0
          inet6 addr: fe80::24c5:7ff:fe2f:43e0/64 Scope:Link
          inet6 addr: fe80::1/64 Scope:Link
          inet6 addr: affe::dc53:7dff:fe99:516e/64 Scope:Global
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:58 errors:0 dropped:0 overruns:0 frame:0
          TX packets:326 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:500
          RX bytes:4706 (4.5 KiB)  TX bytes:29442 (28.7 KiB)


>> Now I get one reply back, and if you add affe::/64 to the fib of the border
>> router the ethos interface (wasn't able to test that) it should work.
>>
>> Cheers,
>> Martine
>>
>>
>> 2016-05-12 13:34 GMT+02:00 Martine Lenders <mail at martine-lenders.eu>:
>>>
>>> Hi Baptiste,
>>> yeah in that case multicast seems like the best thing to go. But how about
>>> using an actual global address then: (ff:1e::1 e.g.; 10 flag =
>>> non-permanent, e scope = global). With the hint by Kaspar I was able to get
>>> the pings through the interface, but apparently the border router does not
>>> forward the address.
>>> Will investigate.
>>>
>>> Cheers,
>>> Martine
>>>
>>> 2016-05-12 11:39 GMT+02:00 Baptiste Clenet <bapclenet at gmail.com>:
>>>>
>>>> Hi,
>>>>
>>>> Martine, what I want:
>>>> My set up:
>>>> |Linux + RIOT BORDER ROUTER|
>>>>
>>>> |RIOT gnrc_networking A|
>>>> |RIOT gnrc_networking B|
>>>> |RIOT gnrc_networking C|
>>>> |RIOT gnrc_networking D|
>>>>
>>>> I run a program on Linux which sends some UDP data to A B C D nodes.
>>>> I want to use multicast in my program instead of one board by one bard
>>>> Is it clearer?
>>>>
>>>> Cheers,
>>>>
>>>> 2016-05-11 21:55 GMT+02:00 Kaspar Schleiser <kaspar at schleiser.de>:
>>>> > Hey,
>>>> >
>>>> > On 05/11/2016 09:42 PM, Martine Lenders wrote:
>>>> >> (2) I'm not sure you add multicast routing entries this way in Linux.
>>>> >
>>>> > You don't. If really desired, you have to use the "local" routing
>>>> > table.
>>>> >
>>>> > # ip -6 route list table local
>>>> >
>>>> > Kaspar
>>>> > _______________________________________________
>>>> > devel mailing list
>>>> > devel at riot-os.org
>>>> > https://lists.riot-os.org/mailman/listinfo/devel
>>>>
>>>>
>>>>
>>>> --
>>>> Baptiste
>>>> _______________________________________________
>>>> devel mailing list
>>>> devel at riot-os.org
>>>> https://lists.riot-os.org/mailman/listinfo/devel
>>>
>>>
>>
>>
>> _______________________________________________
>> devel mailing list
>> devel at riot-os.org
>> https://lists.riot-os.org/mailman/listinfo/devel
>>
>
>
>
> --
> Baptiste



-- 
Baptiste


More information about the devel mailing list