[ovs-dev] [PATCH 1/2] [RFC] datapath: tunneling: Allow vlan_tci to be passed to ovs_tnl_rcv()

[Jesse Gross]

On Thu, Apr 12, 2012 at 5:53 PM, Simon Horman <horms at verge.net.au> wrote:
> On Thu, Apr 12, 2012 at 05:45:33PM -0700, Jesse Gross wrote:
>> On Thu, Apr 12, 2012 at 12:36 AM, Simon Horman <horms at verge.net.au> wrote:
>> > The STT protocol allows a VLAN TCI to be passed as part of the
>> > STT header.  It seems appropriate to pass this TCI to ovs_tnl_rcv()
>> > and for other tunneling protocols to pass 0 to retain their existing
>> > behaviour.
>> >
>> > Signed-off-by: Simon Horman <horms at verge.net.au>
>>
>> Did you see my previous comment about pushing down the call to
>> vlan_set_tci() into the protocol handlers?  I just noticed that you
>> didn't respond to it before.
>
> Sorry, I appologise for that. I had read it but failed to respond.
>
> Is the idea that each implementation of update_header will
> call vlan_set_tci() ? If so, it might be nice to call vlan_set_tci(skb, 0)
> before calling update_header so that tunnelling protocols that
> are VLAN agnostic can just continue being ignorant.

I'm not sure that I understand.  update_header() is called on transmit
but the functions in this patch are for receive.  On the transmit side
we can't wipe out the vlan before calling update_header because STT
needs that information to store in its header.