> I believe that we can avoid problems only if we will have a lock in
 > the ibv_context that handles the completion channels.

Yes, that is the idea I had too.

 > If you think that adding a lock to the ibv_context is good enough i
 > will send you a patch that implements this solution.

Please do.  I don't see anything objectionable to that solution.

 - 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