If it is ip6 traffic why you don't use ip6-unicast feature arc?
ip6-input node can take care for ip6 hdr checks and pass packets to your 
feature node..

-- 
Damjan

> On 16 Apr 2018, at 09:44, Sara Gittlin <[email protected]> wrote:
> 
> Thank you Damjan,
> The my-forwarder plugin performs  l3 lookup  (3D array - on some bytes of the 
> dest-IP (v6) and  decides about the output interface.
> then sends to interface-output node - this works perfectly (with static 
> arp/ndp installed in the dest/source machines) when dest-ip and src-ip are in 
> the same subnet.
> i need a solution for:
>  - not in the same subnet 
> - ARP/NDP responder  - i will cancel the static ARP/NDP
> 
> please tell me if you need more details. 
> Regards
> -Sara
> 
> 
> 
> 
> 
> On Sun, Apr 15, 2018 at 10:05 PM, Damjan Marion <[email protected] 
> <mailto:[email protected]>> wrote:
> Dear Sara,
> 
> It is perfectly fine to send all ARP packets (identified by ethertype) to 
> arp-inut  node from my-forwarder-plugin.
> I'm just afraid that if you go that route that this is just a start of 
> problems....
> 
> Can you describe your use case in more details, we may come up with better 
> solution...
> 
> -- 
> Damjan
> 
>> On 15 Apr 2018, at 15:58, Sara Gittlin <[email protected] 
>> <mailto:[email protected]>> wrote:
>> 
>> Hi 
>> my VPP contains a customized forwarder plugin 
>> 
>> the graph is looking like :
>> 
>> device-input--> my-forwarder-plugin --> ethernet-output (or error drop)
>> this is good  as long as the traversed traffic is within the same subnet
>> but if not - i need to integrate/reuse VPP   ARP/NDP nodes  and the node 
>> which rewrite the L2 header
>> i do not want to pass any ip node .
>>
>> Appreciate your help  what nodes should be added, how the new graph should 
>> look like ?
>> -Sara
>> 
> 
> 

Reply via email to