On Fri, 8 Jul 2011 20:27:34 +0100
Russell King - ARM Linux [off-list ref] wrote:

On Tue, Jul 05, 2011 at 09:49:04AM +0100, Marc Zyngier wrote:

quoted

@@ -43,6 +45,10 @@ static void twd_set_mode(enum clock_event_mode mode,
 		ctrl = TWD_TIMER_CONTROL_IT_ENABLE | TWD_TIMER_CONTROL_ONESHOT;
 		break;
 	case CLOCK_EVT_MODE_UNUSED:
+#ifdef CONFIG_ARM_GIC_PPI_MAP
+		disable_irq(clk->irq);
+		/* fall through */
+#endif

Eww, no.  This is wrong.  You can get an UNUSED event if the TWD clock
event device gets swapped out for some other clock event device (eg,
one with a higher preference.)

This is at best a hack around the problem.  The existing code deals with
hot unplug via this in arch/arm/kernel/smp.c:

/*
 * The generic clock events code purposely does not stop the local timer
 * on CPU_DEAD/CPU_DEAD_FROZEN hotplug events, so we have to do it
 * manually here.
 */
static void percpu_timer_stop(void)
{
        unsigned int cpu = smp_processor_id();
        struct clock_event_device *evt = &per_cpu(percpu_clockevent, cpu);

        evt->set_mode(CLOCK_EVT_MODE_UNUSED, evt);
}

Maybe that needs a local_timer_stop() ->twd_timer_stop() call rather
than trying to be clever by doing this via the set_mode() method.  That
would seem to be overall a cleaner solution.

I think your suggestion to use request_irq() should solve it nicely.
No need to disable the interrupt, it will be re-enabled on the next
setup_irq() when the CPU comes back up.

	M.
-- 
I'm the slime oozin' out from your TV set...