On 05/29/2015 09:37 PM, Josh Hunt wrote:
On 05/29/2015 04:48 PM, Tom Herbert wrote:
Hi Josh,
Why did you need to move the resubmit label?
Grabbing nhoff out of the skb's cb didn't seem relevant anymore, unless
we're requiring the decapsulating code to update the control block
before it returns. Also, since we are returning nexthdr with the return
value it seems unnecessary to go through that work when we already have it.
Not that there has to be symmetry here, but for example, the v4 code
doesn't look up the protocol field again in the inner ip header. It just
uses the protocol value returned from the protocol handler.
Also, I'm skipping the pskb_pull(), which I assumed was left up to the
decapsulating code to setup the data pointer properly before returning.
Again, this is how the v4 code behaves.
The only thing left is the idev which I'm not sure about. Could that
change b/t calls?
Tom
Do you think the above is incorrect and we should just leave the
resubmit label where it currently is? If so, do you think we should just
bypass the nhoff/nexthdr checks and use the nexthdr value returned by
the protocol handler?
Thanks
Josh
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
