[ Sorry for the duplicate message for some of you.  I botched the headers
  on the original mail. ]

I've come across several instances where I need to fiddle with state that
is also touched by a timeout handler.  From a naming standpoint,
splsoftclock() sounds like the correct spl routine to use for protecting
these activities.  Unfortunately this only holds true if splsoftclock()
is used in a process context where the fact that it may actually lower
the ipl to softclock is not an issue.  As it currently stands, in most
cases you can rely on the fact that any spl level will also block callouts,
but this seems risky.  What I'd like to see happen is for either the
semantics of splsoftclock() to change, or to have a new spl introduced
(splcallout()??).  The hope is to provide a consistent interface across
all *BSDs which is why I've addressed this to all of the *BSD projects.
I look forward to you input on this proposal.

--
Justin