On Tue, Jul 25, 2006 at 09:43:15AM -0700, Jason Thorpe wrote:

> We can't really remove the "no sleep" restriction from kauth, because  
> listeners could be in userspace.

That certianly makes things a bit more interesting. This is starting to bite
people though. Can we agree to back the change out for the time being, at
least the spinlock assertion, until we have a locking protocol that fits?

Elad?

Thanks,
Andrew