nvmf/rdma host crash during heavy load and keep alive recovery

[Steve Wise]

> >
> > the DNR bit should not be set normally, only when we either don't want
> > to requeue or we can't.
> >

So when a request is requeued, when is it restarted?  It is getting restarted on
a controller that is in recovery mode and hasn't setup the new qp.  So the
nvme_rdma_queue struct associated with the request is pointing to a freed
ib_pq...