Issue

• The system encountered a null pointer in the queue_process() routine. This caused a page fault and the ensuing code triggered a call to netconsole to report the error. As the queue_lock was already held when the page fault occurred and it attempted to acquire the lock again in netpoll_queue(). This lead to a recursive attempt to take the lock. As it's a spin lock the process entered a loop that it could never return from resulting in the NMI watchdog issuing an NMI that caused an outage.