> > isn't it better to do the opposite?
> > ie. callers of kauth_authorize_action should not assume it never sleep.
> 
> given the replacement of suser() with kauth_authorize_action() (or, the
> authorization wrappers) we might have some places where the code
> actually expects it don't sleep, no?

maybe.  but it's better to fix them.

YAMAMOTO Takashi