On Wed, Feb 17, 2016 at 10:26 AM, Joe Stringer <j...@ovn.org> wrote:
> On 16 February 2016 at 12:35, Joe Stringer <j...@ovn.org> wrote:
>> On 16 February 2016 at 11:43, Jesse Gross <je...@kernel.org> wrote:
>>> On Tue, Feb 16, 2016 at 10:43 AM, Joe Stringer <j...@ovn.org> wrote:
>>>> diff --git a/datapath/linux/compat/lisp.c b/datapath/linux/compat/lisp.c
>>>> index e5a6a7fe00a4..0c81dd52ecab 100644
>>>> --- a/datapath/linux/compat/lisp.c
>>>> +++ b/datapath/linux/compat/lisp.c
>>>> @@ -114,6 +114,7 @@ struct lisphdr {
>>>> };
>>>>
>>>> #define LISP_HLEN (sizeof(struct udphdr) + sizeof(struct lisphdr))
>>>> +#define LISP_MAX_MTU (LISP_HLEN + sizeof(struct iphdr) + sizeof(struct
>>>> ethhdr))
>>>
>>> This doesn't look right to me - isn't this a very small value? Also,
>>> LISP doesn't have an inner Ethernet header, unlike the other tunnels.
>>
>> You're right, I messed this one up. Looks like it should be
>> IP_MAX_MTU - LISP_HLEN - sizeof(struct ethhdr)
>
> Or rather, this.
> (IP_MAX_MTU - LISP_HLEN - sizeof(struct iphdr))
>
> Though, the max differs for IPv4/IPv6 - Could drop it to using the
> ipv6hdr instead.
There's currently no support for LISP over IPv6 so I guess we don't
need to worry about it for the time being at least.
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
