[support] NEPL Tunnel ID

[Arnaud Ebalard]

Hi Ben,

"Ben McCarthy" <b.mccarthy at lancaster.ac.uk> writes:

> I had a problem quite a while ago where I needed to identify which MRs
> tunnel a packet had been sent via. So if a HA has 10 connected and active
> MRs I need the HA to be able to determine which MR-HA tunnel a packet
> (specifically a BU) was sent down. Unfortunately due to the nature of what
> I'm doing (MANEMO) I can't simply identify which tunnel a packet was sent
> down based on address, so I came up with a intermediary solution where I
> appended information to a BU that allowed me to identify the tunnel in user
> space after it had been delivered. Unfortunately as the scenarios I am
> supporting have become more complex I have come across a flaw in this
> intermediary approach. Ideally it would be perfect if I could receive a BU
> and read some netlink variable from user space that identifies which tunnel
> it arrived via. but nothing is ever that simple and I think I determined
> that wasn't possible the last time I looked at this.
>
> So does anyone know if this would be possible from within the kernel (by
> hacking the v6 tunnel interface code a little?) and maybe have any hints how
> I might best go about achieving this?

I haven't check but will the iif in ha_recv_bu_worker() in ha.c not
provide the info about which tunnel iface the BU was received from when
it came by a tunnel? 

Cheers,

a+