On Fri, Oct 26, 2012 at 11:32 AM, Jesse Gross <je...@nicira.com> wrote:
> On Sat, Oct 27, 2012 at 1:06 PM, Pravin B Shelar <pshe...@nicira.com> wrote:
>> diff --git a/include/linux/openvswitch.h b/include/linux/openvswitch.h
>> index 89feb61..d4a043c 100644
>> --- a/include/linux/openvswitch.h
>> +++ b/include/linux/openvswitch.h
>> @@ -182,6 +182,7 @@ enum ovs_vport_type {
>> OVS_VPORT_TYPE_UNSPEC,
>> OVS_VPORT_TYPE_NETDEV, /* network device */
>> OVS_VPORT_TYPE_INTERNAL, /* network device implemented by datapath */
>> + OVS_VPORT_TYPE_FT_GRE,
>
> I'm not sure what FT stands for. Future? In any case, can you add a
> comment next to it?
I meant Flow-based Tunneling, I will add a comment.
>
>> OVS_VPORT_TYPE_PATCH = 100, /* virtual tunnel connecting two vports
>> */
>> OVS_VPORT_TYPE_GRE, /* GRE tunnel */
>> OVS_VPORT_TYPE_CAPWAP, /* CAPWAP tunnel */
>
>> diff --git a/lib/netdev-vport.c b/lib/netdev-vport.c
>> index 621abd1..a864341 100644
>> --- a/lib/netdev-vport.c
>> +++ b/lib/netdev-vport.c
>> @@ -154,6 +154,9 @@ netdev_vport_get_netdev_type(const struct
>> dpif_linux_vport *vport)
>> case OVS_VPORT_TYPE_PATCH:
>> return "patch";
>>
>> + case OVS_VPORT_TYPE_FT_GRE:
>> + return "gre";
>
> I don't think that we want to introduce any uses of the symbolic name
> in userspace. In the next release, we'll drop the FT part of the name
> (but keep the same constant) so I think it's easier if we don't touch
> userspace.
>
ok.
> Also, did you look at the userspace code to see whether it can handle
> the transition by removing port types that it doesn't understand and
> recreating them?
Yes, I checked it and it works.
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
