Re: [patch] "HT scheduler", sched-2.5.63-B3

Dimitrie O. Paun (dpaun@rogers.com)
Thu, 6 Mar 2003 19:09:38 -0500


On March 6, 2003 07:02 pm, Linus Torvalds wrote:
> Btw, "success" is often "being good enough". We shouldn't _suck_, but you
> should always remember the old "perfect is the enemy of good" thing.
> Trying to get perfect behaviour under some circumstances often means that
> you suck under others, and then the right answer is usually not to try to
> be so damn perfect, but "just being good".

Very much so! I can not agree more with you that the lack of a default
good policy is stupid -- this is why X was in such a bad state for years!
I could write pages on this subject, but I'm glad other people share my
ideas in this matter :)

That being said, I have nothing against an automatic algorithm. If we can
get that one to do what we want, there's obviously no need for knobs. All
I'm saying is that *if* there are cases where we need help, we should not
have to tweak the priority of a process, but create a different knob.
Priority is a different beast from interactivity. For example, it's clear
that good interactive behaviour _requires_ information only available in
the kernel. Which makes the interactivity know (if at all required) more
of a hint rather than a set in stone thing like the priority.

But I guess all this is academic. If the automatic stuff just works,
we're better off anyway.

-- 
Dimi.

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/