Sean Hefty wrote
How do we know that the user hasn't tried to destroy the id from another
callback? We need some sort of state check here.
fixed
+static int cma_netdev_align_id(struct net_device *ndev, struct rdma_id_private
*id_priv)
nit - function name isn't clear to me. Maybe something like
cma_netdev_change_handler()? Although I'm not sure that netdev change is what
the user is really interested in. What they really want to know is if IP
address mapping/resolution changed. netdev is hidden from the user.
I changed the function name to cma_netdev_change as it checks if there
was some netdev change between the time of this ID address resolution to
when the netdev event was delivered. The user doesn't get explicit
notification from the rdma-cm on netdev change but rather on address
change as you suggested next.
Maybe call this RDMA_CM_EVENT_ADDR_CHANGE?
done
Or.
_______________________________________________
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
