After commit 
e9d8b2c2968499c1f96563e6522c56958d5a1d0d (xen-netback:
disable rogue vif in kthread context), a fatal (protocol) error would
leave the guest Rx thread spinning, wasting CPU time.  Commit
ecf08d2dbb96d5a4b4bcc53a39e8d29cc8fef02e (xen-netback: reintroduce
guest Rx stall detection) made this even worse by removing a
cond_resched() from this path.
Since a fatal error is non-recoverable, just allow the guest Rx thread
to exit.  This requires taking additional refs to the task so the
thread exiting early is handled safely.
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Reported-by: Julien Grall <julien.grall@linaro.org>
Tested-by: Julien Grall <julien.grall@linaro.org>
Acked-by: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
 
                goto err_rx_unbind;
        }
        queue->task = task;
+       get_task_struct(task);
 
        task = kthread_create(xenvif_dealloc_kthread,
                              (void *)queue, "%s-dealloc", queue->name);
 
                if (queue->task) {
                        kthread_stop(queue->task);
+                       put_task_struct(queue->task);
                        queue->task = NULL;
                }
 
 
                 */
                if (unlikely(vif->disabled && queue->id == 0)) {
                        xenvif_carrier_off(vif);
-                       xenvif_rx_queue_purge(queue);
-                       continue;
+                       break;
                }
 
                if (!skb_queue_empty(&queue->rx_queue))