Prev: [PATCH 0/5] use idr kernel library instead of handles
Next: [PATCH v2] cpuidle: extend cpuidle and menu governor to handle dynamic states
From: Pavel Machek on 9 Aug 2010 03:30
Hi!

> > I've CC'ed Andrea because we were having an offline conversation about
> > whether ksmd (and his khugepaged) need to set_freezable(); and I wonder
> > if this swap bug underlies his interest, though he was mainly worrying
> > about I/O in progress.
>
> My opinion is that with current freezer model it is needed for suspend
> to disk. The kthread_should_stop seems useless for kswapd/ksmd, but
> I'm afraid it might get useful in the future so just to stay on the
> safe side I added it to khugepaged as well, but it's contributing to
> the pollution.
>
> I've no idea why the freezing isn't preemptive (through the scheduler,
> all these kernel threads are obviously lowlatency beasts) by default

Hibernation woulld like all tasks frozen with no locks held - so that
write to disk does not have to care about locking.

--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
First  |  Prev  | 
Pages: 1 2 3 4 5 6 7
Prev: [PATCH 0/5] use idr kernel library instead of handles
Next: [PATCH v2] cpuidle: extend cpuidle and menu governor to handle dynamic states