Jonathan Stone writes:
> The issue I really wanted to address was the ntp_adjtime() syscall
> which was *not* in the NetBSD kernel last time I looked, even as an
> option.  I guess I'm a time vulture. I find high-quality time
> synchronization invaluable for performance measurement and tuning of
> networking code.

If you are refering to the "PLL in the kernel" stuff, if you can port
it cleanly, why don't you do so and submit it for inclusion in the
source tree? I agree it would be a win, though I don't remember how
clean it is.

Perry