> The existing code (performing SET_PORT for Ethernet port) is not a
 > bug, bu it is unnecessary.  It turns into a bug if patch 1/2 is
 > applied.

So if someone doing a bisection lands between 1/2 and 2/2 then mlx4 is
broken?  Can the order of patches be reversed so that we avoid having a
broken intermediate state?  Or should this just be one big patch?
_______________________________________________
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