On Mon, Apr 20, 2015 at 3:00 AM, frankzfz <[email protected]> wrote: > > Dear： > Test scenario: two KVM guests sitting in different hosts communicate to each > other with a GRE tunnel. > > > All interface MTU is default 1500 Bytes, from guest point of view, its skb > gso_size could be as bigger as 1448 Bytes, however after guest skb goes > through GRE encapuslation, individual segments length of a gso packet could > exceed physical NIC MTU 1500, which will be possible modified at recevier > side.So Sender have to fragment the IP datagrams that are larger than the > interface MTU . > > The packets is modified 42 bytes in ovs module at receiver side.42 bytes is > equal to GRE encapsulation header（Ethernet header（14） + IP Header （20）+ GRE > Header（8））. > > If interface MTU is modified 1458 Bytes, after guest skb goes through GRE > encapsulation, IP fragmentation will not occur at sender side.SKb is good at > receiver side. This mistake didn't happen at receiver side. > > Why SKB is modified 42 bytes at receiver side through OVS modules for > fragmentation packets.
What does it mean that the packet is modified? What specifically happens? However, as a general rule, Linux does not support fragmentation of tunnel packets and OVS inherits this behavior. _______________________________________________ discuss mailing list [email protected] http://openvswitch.org/mailman/listinfo/discuss
