On 13:41 Mon 27 Jul , Hal Rosenstock wrote: > > It could be done this way but it's unclear whether that's the best approach > going forward. In the future, to minimize the changes on an update due to > topology change, we're not sure how much state will need to be kept from > one iteration to the next.
When we will know we will do accordingly. But now mesh node structures are useless between sweeps and freed anyway, so let's just do it in a right place. Sasha _______________________________________________ general mailing list [email protected] http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
