On Thu, Nov 27, 2014 at 09:34:06AM +0530, Shashwat Srivastava wrote:
> Thanks for the reply just wanted to confirm did you get any reply
> regarding the EXT-502 from ONF extensibility working group.
Yes, Jean Tourrilhes wrote:
If you assume "importance only", I would vote no. Trying to squeeze
a low importance flow entry by displacing more important flow entry
does not make sense.
IMHO, the text is ambiguous and could be interpreted both ways. The
prototype for EXT-192 refuses inserting the flow entry in that case
and returns OFPFMFC_TABLE_FULL. By the way, you may want to point
the developer to the prototype, that could help him...
We need to clarify that in the spec. At this point, it's unclear
which specification would carry that clarification, which is why I
put 1.6 and 1.4.1 as the target.
and I replied:
Thanks. I passed that along.
OFPFMFC_TABLE_FULL seems like a simple solution to me.
I pointed the developer to
https://github.com/jean2/ofsoftswitch13/commits/linda/ext-192-eviction,
I hope that's the right one.
> Also can you please let me know what will happen if the new flow enters
> the switch with same importance as the lowest importance in the flow
> table. Considering the earlier given example (new flow has importance=100
> and lowest flow in the existing flows has importance = 100), in this case
> will the existing flow will be evicted or the new flow will be dropped ?
I think I'd probably use OFPFMFC_TABLE_FULL for that case too.
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
