On Fri, Aug 1, 2014 at 5:22 PM, Ben Pfaff <[email protected]> wrote: > A dpif reports EEXIST if a flow put operation that should create a new flow > instead attempts to modify an existing flow, or ENOENT if a flow put would > create a flow that overlaps some existing flow. The latter does not always > indicate a bug in OVS userspace, because it can also mean that two > userspace OVS packet handler threads received packets that generated > the same megaflow for different microflows. Until now, userspace has > logged this, which confuses users by making them suspect a bug. We could > simply not log ENOENT in userspace, but that would suppress logging for > genuine bugs too. Instead, this commit drops DPIF_FP_MODIFY from flow > put operations in ofproto-dpif, which transforms this particular > problem into EEXIST, which userspace already logs at "debug" level (see > flow_message_log_level()), effectively suppressing the logging in normal > circumstances. > > It appears that in practice ofproto-dpif doesn't actually ever need to > modify flows in the datapath, only create and delete them, so this > shouldn't cause problems. > > Suggested-by: Jesse Gross <[email protected]> > Signed-off-by: Ben Pfaff <[email protected]>
Acked-by: Jesse Gross <[email protected]> _______________________________________________ dev mailing list [email protected] http://openvswitch.org/mailman/listinfo/dev
