Open vitalivanov opened 10 years ago
From what I can see, OFTest relies heavily on loxigen. That project has recently added support for 1.3.1 so OFTest needs to be updated accordingly. When/if this will happen though I'm unsure?
Fwiw, we've been using OFTest internally to do lots of OF-1.3 testing, so all of the knobs are there. The problem is that OF1.3 is hard to test without making assumptions about the packet processing pipeline (that is, the order and capabilities of the match action tables). The set of tests that you can run without making assumptions about the packet processing pipeline is fairly trivial and not terribly useful.
OpenFlow in general hasn't really solved that problem - it's not really an OFTest issue.
On Thu, Nov 27, 2014 at 3:59 AM, Stephen Finucane notifications@github.com wrote:
From what I can see, OFTest relies heavily on loxigen https://github.com/floodlight/loxigen. That project has recently added support for 1.3.1 so OFTest needs to be updated accordingly. When/if this will happen though I'm unsure?
— Reply to this email directly or view it on GitHub https://github.com/floodlight/oftest/issues/126#issuecomment-64781873.
I believe parsing/packing of table_feature_prop types have not yet been resolved. Other than that, I think everything is fine. https://github.com/floodlight/loxigen/issues/198
True -- I had forgotten about that. Is this something that's significantly an issue? If so, I might be able to get someone on our side to look into it.
On Mon, Dec 1, 2014 at 9:21 AM, Jonathan Stout notifications@github.com wrote:
I believe parsing/packing of table_feature_prop types have not yet been resolved. Other than that, I think everything is fine.
— Reply to this email directly or view it on GitHub https://github.com/floodlight/oftest/issues/126#issuecomment-65100185.
There are two main cases were these message types are used in conformance style tests.
Can anyone describe the status of openflow 1.3 protocol support? What is done and what is to be done?
TIA