Frank Kardel <kardel@netbsd.org> writes:
>>some of the stuff one wants to
>>do in the kernel needs to have a good idea of the time.
>
> It is certainly faster in the kernel (no news here).
> I am wondering in what kind of time you are actually interested. Are you
> interested in a time related to UTC? Are you interested in a time with
> a frequency controlled to be close to the norm (SI second).

The latter is more important than the former...

> Or are you interested in measuring just time differences.

That is probably the most important thing to me.  The things I care
about for very high precision timers including profiling kernel
activity, precisely pacing isochronous processes, etc.  For those
things, measuring times and being able to know what interval has
passed is reasonably important -- knowing the time of day is
not. However, other people may need precise time of day for various
purposes, such as high speed packet capture and logging, where time of
day is also important.

In any case, I think it is okay if the time gets somewhat inaccurate
for a moment when clock frequncy changes, so long as it remains
reasonable while the clock frequency is stable.

>>Well, as a fallback, that's at least a lot better than the ancient PC
>>timer chip. How hard would it be to make that work correctly?
>
> Thats work in progress - Matthias Drochner already sent me a
> preliminary implementation.

Good to hear...

Perry