On 1/22/21 12:42 AM, Ben Pfaff wrote: > On Wed, Jan 13, 2021 at 05:32:52PM +0100, Dumitru Ceara wrote: >> On 1/13/21 2:56 AM, Ilya Maximets wrote: >>> On the side note: don't we need to increase 'cond_seqno' here since >>> we're practically dropping one of the condition requests? User might >>> wait for this sequence number, but 'req_cond' will be cleared and >>> never be acked. >> >> I think you're right, we should be increasing 'cond_seqno' to match the >> logic in ovsdb_cs_db_set_condition() (or ovsdb_idl_db_set_condition() in >> the old code). >> >> I can take care of this as a follow up patch as it seems to be >> pre-existing issue of the IDL code. > > I found that any minor change to the IDL behavior in this area tended to > break ovn-controller, which seems to have really brittle dependencies on > exactly how the IDL behaves.
Yep. I think, narrowing down direct idl usage from the ovn-controller side is one of the main ToDo items for the refactoring we talked about on some of the previous OVN meetings. > > Thanks for the reviews. I got the impression from these and from the > discussion on IRC this morning that it was OK to push this to master, so > I did. > Sure, thanks! I think we also agreed to apply the series to branch-2.15 (it was already created last week). Will you do that? Best regards, Ilya Maximets. _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
