machine can not reboot itself while ed1 is present (sources of 19th feb 2004)

[Dave Cuthbert]

Matthew Dillon wrote:
    It sounds like an interrupt is not getting deasserted and the system
    is livelocking until the card gets pulled out and stops deasserting
    the interrupt.  My guess.
Hm.  It would be useful, from a debugging/development standpoint, if the 
kernel could heuristically detect this.

I'm thinking of something like:

interrupt_loop_counter = 0;

while (interrupt_asserted()) {
    if (++loop_counter > MAX_INTERRUPT_CYCLES) {
        disable_interrupt();
        print_a_message_about_this();
        enter_debugger();
        reenable_interrupt();
    }
    handle_interrupts();
}