[ovs-dev] [learning v2 14/19] ofproto: Reinterpret meaning of OpenFlow hard timeouts with OFPFC_MODIFY.

Ben Pfaff blp at nicira.com
Thu Sep 8 21:39:31 UTC 2011


On Thu, Sep 08, 2011 at 11:28:27AM -0700, Ethan Jackson wrote:
> Seems fine to me.  Out of curiosity: Is this compatible with the 1.0
> spec?  I wonder if anyone is relying on the old behavior?  Seems
> unlikely I suppose.

    As far as I can tell, OpenFlow 1.0 doesn't specify this.  It does say:

    For MODIFY requests, if a flow entry with identical header fields
    does not current reside in any table, the MODIFY acts like an ADD,
    and the new flow entry must be inserted with zeroed counters.
    Otherwise, the actions field is changed on the existing entry and
    its counters and idle time fields are left unchanged.

That is, it says explicitly that the "idle time" is not updated, but
it doesn't say anything about hard timeouts, so I'm taking the liberty
of interpreting it in the way most convenient for us.



More information about the dev mailing list