kardel@NetBSD.org said: > I can look into this tonight (in about 5h time) and see whether I get > the same symptoms in x86. It seems that the unsigned hardclock_ticks breaks the kernel timeout/callout code. That would not be i386 specific. best regards Matthias