Thanks Pete!
Your answer addresses the main thing I tried to achieve. I guess it is not 
that easy to implement an easy "workaround" for that quing of interrupts, 
like a shortcut for some special interrupts. And still it would be highlly 
unpredictable how long time every queue would take to service.

But on the other hand if I keep the interrupts to a minimum then the NetBSD 
kernel is somehow relative predicatble in terms of guaranteed interrupt 
service, right?

So for some low sampling applications the kernel might work as it is, by 
blocking some other interrupts. How do I block interrupts, is that done 
through standard kernel function calls? This sound more like an device 
driver issue, right?

/magnus

-- 
Dr. Magnus Sethson
Email: magnus@sethson.net magse@ikp.liu.se
EOF