> Then in the interim, how do we address the interoperability issue between 
 > say
 > for example, Topspin and IBM HCAs using connected mode -switch to datagram 
 > mode? 

I think that we need to understand the root cause of the issue and try
to solve it without going to a non-zero retry count if possible.
Right the theory is that it is a bug in the CM, namely that the
respective HCA ack delays are not taken into account.  So we should
fix that bug.

 - R.
_______________________________________________
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

Reply via email to