RE: KDB: SMP locking problem

[Sanders, Richard]

> This should have the same effect and is much simpler.  It will only
> unlock and clear the state in the same logic flow that got the lock.
> 

Looks great to me.