[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <ab7ef74f-294b-af2b-cbd4-d500fc0423ef@6wind.com>
Date: Mon, 5 Sep 2016 12:05:57 +0200
From: Nicolas Dichtel <nicolas.dichtel@...nd.com>
To: David Lebrun <david.lebrun@...ouvain.be>,
Roopa Prabhu <roopa@...ulusnetworks.com>
Cc: netdev@...r.kernel.org
Subject: Re: [RFC 2/9] ipv6: sr: add code base for control plane support of
SR-IPv6
Le 01/09/2016 à 15:21, David Lebrun a écrit :
> On 08/29/2016 05:31 PM, Roopa Prabhu wrote:
>> This looks fine. But, i am just trying to see if this can be rtnetlink.
>> Have you considered it already ?.
>> We would like to keep the API consistent or abstracted to accommodate SR-MPLS in the
>> future too. so, any abstraction there will help.
>>
>> what is your control-plane software using this ?
>> quagga or any-other routing daemon ?
>>
>> Thanks,
>> Roopa
>>
>
> rtnetlink is used for the routing part through the lwtunnels. In this
> specific patch, the API is intended to configure namespace-wide
> parameters such as a map of hmackeyid -> hashfn,secret, without
> connection to particular routes. I agree that the tunsrc parameter could
> be provided through rtnetlink, but the hmac data cannot (which by the
> way does not exist for SR-MPLS).
But it's closely related to routes. You can add a new family RTM_foo. It will be
easier to implement userspace part (only one netlink family is involved).
Regards,
Nicolas
Powered by blists - more mailing lists