On Tue, Jun 11, 2019 at 03:00:00PM -0700, Gregory Rose wrote: > I see the source of the confusion here. The vlan1 interface is added to the > OVS bridge and the port type > is internal. Here's a better way to look at this: > > Net Device > <Can originate packets> OVS switch > |------------------| |-----------------| > | | | | > | vlan1 |<----------->| ovsbr1 | > | | ^ | | > |------------------| | |-----------------| > | > | > <internal port> <----------------- Think of that as a > virtual Ethernet cable > > When you ssh to a destination via the vlan1 interface then the vlan1 > interface is generating the packets. If it has a tcp checksum offload > capability then it would use it but that will depend on the master device > it is controlled by. This port is in no way owned by OVS. OVS has simply > added it to the bridge using a virtual port which is by convention called > an 'internal' port. But think of it as the cable connecting your virtual > device 'vlan1' to the OVS bridge 'ovsbr1'. > > Does that help explain?
Only if you can tell me what the vlan1 device is. :-) I had assumed this was a first-class concept within ovs; after all, when you create a bridge you get one of these. It's created with Debian's ifupdown integration (an “OVSIntPort”-type interface), which seems to do: ovs_vsctl -- --may-exist add-port "${IF_OVS_BRIDGE}"\ "${IFACE}" ${IF_OVS_OPTIONS} -- set Interface "${IFACE}"\ type=internal ${OVS_EXTRA+-- $OVS_EXTRA} > No, the internal ports are the virtual interfaces between physical/virtual > devices (vlan1) and the OVS bridge (ovsbr1). So what are the vlan1 devices? Who should I nag to get them to pad their packets correctly? :-) /* Steinar */ -- Homepage: https://www.sesse.net/ _______________________________________________ discuss mailing list disc...@openvswitch.org https://mail.openvswitch.org/mailman/listinfo/ovs-discuss