Unfortunately, this commit breaks many unit tests. Now we won't know whether the unit tests are failing because of this commit or because of some other regression.
On Thu, Jun 25, 2015 at 9:50 AM, Eitan Eliahu <[email protected]> wrote: > If it gets too involved have your user mode hack instead. > Thanks Alin. > Eitan > > -----Original Message----- > From: Alin Serdean [mailto:[email protected]] > Sent: Thursday, June 25, 2015 8:38 AM > To: Eitan Eliahu; Nithin Raju > Cc: Ben Pfaff; [email protected] > Subject: RE: [ovs-dev] [PATCH v2] dpif_probe_feature not suported in windows > datapath > > Eitan, > > Ok I will add it in the datapath. > > Why I am considering it as a hack, because if we see it in the datapath we > just reply with an error, there will be no logic behind it at the moment. But > maybe it makes some sense to have it for further user. > > Alin. > > -----Mesaj original----- > De la: Eitan Eliahu [mailto:[email protected]] > Trimis: Thursday, June 25, 2015 6:01 PM > Către: Alin Serdean; Nithin Raju > Cc: Ben Pfaff; [email protected] > Subiect: RE: [ovs-dev] [PATCH v2] dpif_probe_feature not suported in windows > datapath > > > > Alin, > We just need to add: > [OVS_FLOW_ATTR_PROBE] = { .type = NLA_FLAG }, to the flow policy. > This is not a hack. > If it is time consuming please go ahead with the uder mode change. > Thanks, > Eitan > > > -----Original Message----- > From: Alin Serdean [mailto:[email protected]] > Sent: Thursday, June 25, 2015 7:30 AM > To: Nithin Raju; Eitan Eliahu > Cc: Ben Pfaff; [email protected] > Subject: RE: [ovs-dev] [PATCH v2] dpif_probe_feature not suported in windows > datapath > > Both are hacks IMO. > > We could respond to DPIF_FP_PROBE but I prefer to do it in userspace until we > address the bigger problem in the datapath. > > Alin. > > -----Mesaj original----- > De la: Nithin Raju [mailto:[email protected]] > Trimis: Thursday, June 25, 2015 4:51 PM > Către: Eitan Eliahu > Cc: Ben Pfaff; Alin Serdean; [email protected] > Subiect: Re: [ovs-dev] [PATCH v2] dpif_probe_feature not suported in windows > datapath > >> On Jun 24, 2015, at 12:57 PM, Eitan Eliahu <[email protected]> wrote: >> >> >> I'm wondering if we could fail this call in the datapath level by examining >> the DPIF_FP_PROBE bit in the flags. >> If it is too hard we can still live with user mode code change. >> Thanks, >> Eitan > > That’s a good suggestion. We can go with Alin’s solution for 2.4 since we > exactly know what code we will be disabling, and work on the flow validation > logic for master. > > thanks, > -- Nithin > _______________________________________________ > dev mailing list > [email protected] > http://openvswitch.org/mailman/listinfo/dev _______________________________________________ dev mailing list [email protected] http://openvswitch.org/mailman/listinfo/dev
