> When i tried to recover the QP and modify it's state to RTS i had to > use the flag IBV_QP_CUR_STATE > because the internal QP structure assumed that the QP state is RTS > (although query QP noticed otherwise). > > I created and tested 2 patches for both mthca and mlx4 drivers to > update the internal QP state when a > successful query QP was executed. > > Will you accept those patches?
Sounds fine, although I think we need to take a little care about how we update the QP state (locking) now that there are two ways to do it. And obviiously it's hard to make a definitiev statement about patches I've never seen... _______________________________________________ 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
