Re: [PATCH net-next 0/3] net: vrf: Add support for local traffic to local addresses

On 6/6/16 4:56 PM, David Miller wrote:

From: David Miller <davem@davemloft.net>
Date: Mon, 06 Jun 2016 15:12:28 -0700 (PDT)

quoted

From: David Ahern <redacted>
Date: Thu,  2 Jun 2016 13:15:09 -0700

quoted

Add support for locally originated traffic to VRF-local addresses,
be it addresses on enslaved devices or addresses on the VRF device:

 ...

Series applied, but I've been wondering what happens to hw offloads
when these VRF devices sit in the middle.

Does TSO et al. still occur properly?

It should, but I will double check and adjust if needed.

Actually I have to revert this series, ip6_input() is not an exported
module symbol.

If you're only build testing things like this with everything "=y",
please reconsider.

As I noted in the commit message "ip6_input is exported so the VRF 
driver can use it for the dst input function." It was dropped somewhere 
along the way of refactoring for upstream. Will send a v2.