I keep getting hung up on the additional complexity introduced by the new
Physical_Endpoints table proposed here:
http://openvswitch.org/pipermail/dev/2016-March/068705.html
I wanted to see how much work it would be to implement a software L2 gateway
while trying to minimize the amount of change required to do so.
The real code changes here are really quite small. A new port type called
"gateway" is introduced that works *very* close to the existing localnet ports.
I'm interested in feedback on this alternative approach.
Thanks,
Russell Bryant (2):
ovn: Minor refactoring.
ovn: Add minimal software l2 gateway.
ovn/controller/binding.c | 12 ++-
ovn/controller/ovn-controller.8.xml | 15 ++--
ovn/controller/ovn-controller.c | 2 +-
ovn/controller/patch.c | 68 ++++++++-------
ovn/controller/patch.h | 2 +-
ovn/controller/physical.c | 10 ++-
ovn/ovn-nb.xml | 19 +++++
ovn/ovn-sb.xml | 38 +++++++++
tests/ovn.at | 164 ++++++++++++++++++++++++++++++++++++
9 files changed, 285 insertions(+), 45 deletions(-)
--
2.5.5
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
