Re: [PATCH] Fix locking bug in vcpu_migrate

[Keir Fraser <keir.xen@gmail.com>]

On 22/04/2011 19:23, "John Weekes" <lists.xen@nuclearfallout.net> wrote:

> 
>> Nice that empirical evidence supports the patch, however, I'm being dense
>> and don't understand why order of lock release matters.
> 
> I was taught long ago to release them in order, but as I think about it,
> I agree with you that there doesn't seem to be a scenario when the
> release would matter.
> 
> It's odd that it seemed to lead to such a big difference for me, then.
> I'll do some further tests -- maybe I changed something else to cause
> the behavior, or the problem is more random than I thought and just
> hasn't occurred for me yet in all the new tests.

Thanks!

 -- Keir

>> perhaps you've merely perturbed a fragile pos code that's
>> broken in some other way.
> 
> That's entirely possible.
> 
>> Also the last hunk of your patch is broken -- in the final else clause you
>> call spin_unlock_irqrestore on the wrong lock. This is very definitely a
>> bug, as irqs should never be enabled while any schedule_lock is held.
> 
> Definitely. I had fixed that here but sent an old version of the patch
> -- a boneheaded mistake.
> 
> -John