On Mon, Mar 12, 2018 at 09:05:22PM +0200, Leon Romanovsky wrote: > On Mon, Mar 12, 2018 at 06:39:00PM +0000, Hefty, Sean wrote: > > > > rdma_join_multicast() includes state checks which are intended to > > > validate that the cm_id is in a valid state. It may make more sense > > > to expand those checks with a device check, since the state checks > > > alone are insufficient. > > > > > > Do you know why these checks not enough? Why did the change state > > > without cm_id->device? > > > > I believe the cm_id can be bound to a wildcard address (state RDMA_CM_ADDR_BOUND), which would leave device unset. RDMA_CM_ADDR_BOUND is used in both the wildcard and non-wildcard case. > > ok and one more question, the syzkaller reported failure in this line line: > 4170 if (rdma_protocol_roce(id->device, id->port_num)) > > Will the check of existence of id->device enough? Or should I need to check port_num too? Sure would be confusing if device was valid but port_num was not :( Jason -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
