Re: [PATCH] machine_power_off: not only local_irq_disable but also do disable preemption

[Russell King - ARM Linux <linux@arm.linux.org.uk>]

On Sun, Jul 06, 2014 at 12:20:03AM +0530, pawandeep oza wrote:
> preempt_check_resched would check TIF_NEED_RESCHED
> #define preempt_check_resched() \
> do { \
> if (unlikely(test_thread_flag(TIF_NEED_RESCHED))) \
> preempt_schedule(); \
> } while (0)
> 
> there is a chance that just beofre we disabled irqs, somebody would have
> marked the flag to current, and
> later on, it might happen that, current gets replaced by the process which
> tries to hold a spin_lock which has already been previosuly held by CPU1
> when
> was being plugged out by smp_send_stop.

And preempt_schedule() contains:

        /*
         * If there is a non-zero preempt_count or interrupts are disabled,
         * we do not want to preempt the current task. Just return..
         */
        if (likely(!preemptible()))
                return;

where preemptible() is defined as:

	(preempt_count() == 0 && !irqs_disabled())

which means... if interrupts are disabled (as they are) we don't
return from preempt_schedule() without doing anything.

Scheduling with interrupts disabled is a bug.  If you need to add
preempt_disable() before local_irq_disable() to prevent it, then
there's a bug somewhere else, and we don't "fix" that by adding
preempt_disable().  The real bug needs to be found and fixed.

-- 
FTTC broadband for 0.8mile line: now at 9.7Mbps down 460kbps up... slowly
improving, and getting towards what was expected from it.