On Fri, Aug 1, 2014 at 10:56 AM, Jesse Gross <[email protected]> wrote:
> On Tue, Jul 29, 2014 at 5:11 PM, Ansis Atteka <[email protected]> wrote:
>> diff --git a/datapath/flow_netlink.c b/datapath/flow_netlink.c
>> index e1eadbb..41199ff 100644
>> --- a/datapath/flow_netlink.c
>> +++ b/datapath/flow_netlink.c
>> @@ -1051,10 +1052,7 @@ int ovs_nla_put_flow(struct datapath *dp, const
>> struct sw_flow_key *swkey,
>> const struct geneve_opt *opts = NULL;
>>
>> if (!is_mask) {
>> - struct vport *in_port;
>> -
>> - in_port = ovs_vport_ovsl_rcu(dp, swkey->phy.in_port);
>> - if (in_port->ops->type == OVS_VPORT_TYPE_GENEVE)
>> + if (swkey->tun_key.tun_flags &
>> TUNNEL_OPTIONS_PRESENT)
>> opts = GENEVE_OPTS(output,
>> swkey->tun_opts_len);
>> } else {
>> if (output->tun_opts_len)
>
> I think we can probably use this logic for both the mask and non-mask
> cases, right? TUNNEL_OPTIONS_PRESENT is set in the same cases that the
> mask tun_opts_len is set and is probably clearer.
I sent v4 for that.
_______________________________________________
dev mailing list
[email protected]
http://openvswitch.org/mailman/listinfo/dev
