try_to_freeze() called with IRQs disabled on ARM

[Tejun Heo]

On Wed, Aug 24, 2011 at 12:17:03AM +0200, Tejun Heo wrote:
> if (freezing() && IRQ disabled) {
>       bust on IRQ;
>       try_to_freeze();
>       replug IRQ;
> }
> 
> But, that can't be right. The current code isn't triggering warning
> from scheduler code, right? If the above is the case, it should be
> triggering that. What am I missing?

I think the refrigerator() code was actually doing that through
spin_[un]lock_irq(), so it was accidentally masking the problem.  It
definitely seems to need fixing.

Anyways, for now, we can do two things,

1. if (freezing()) { irq_save; try_to_freeze(); irq_restore; } w/ BIG
   FAT UGLY comment.

2. Drop might_sleep() from try_to_freeze().  Moving it to
   refrigerator() wouldn't help much.  It would just trigger more
   sporadically during freeze, which is arguably worse than now.

I'd prefer #1 given that it documents the breakage while also
restoring the IRQ state afterwards FWIW.

Thanks.

-- 
tejun