Re: Problem with kernel-pll in 2.0.3x (at least)

Rogier Wolff (R.E.Wolff@BitWizard.nl)
Mon, 4 May 1998 11:49:52 +0200 (MET DST)

Ulrich Windl wrote:
> The point of you code is: You are doing basically the same as the NTP
> kernel code already does, i.e. remember fractional ticks and add them
> if the fraction is large enough.

Right.

> The correct solution would probably be to add a frequency offset that
> corrects the systematic error of the software to the offset that the
> user sees. (talking about MOD_FREQUENCY). Unfortunately the NTP code

That's what My patch does doesn't it?

> originally was designed with 32 bits in mind, and the overflow
> considerations limit the maximum correction. With a systematic offset
> of more than 400PPM there isn't much tolarence left for bad hardware
> (=quartz). Therefore I'd doe it outside of the NTP code, even if it
> wastes some cycles.

Ok.

> As long as the interrupt handling is architecture-specific, I don't
> see the advantage of the very universal code.

Huh? I've just looked at the timekeeping code, but its all
arch-independent. It would be nice to keep it that way.....

Roger. 

-- 
If it's there and you can see it, it's REAL      |___R.E.Wolff@BitWizard.nl  |
If it's there and you can't see it, it's TRANSPARENT |  Tel: +31-15-2137555  |
If it's not there and you can see it, it's VIRTUAL   |__FAX:_+31-15-2138217  |
If it's not there and you can't see it, it's GONE! -- Roy Wilks, 1983  |_____|


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu