Re: a different approach to scheduling issues
Richard Gooch (rgooch@atnf.csiro.au)
Thu, 1 Oct 1998 11:00:15 +1000
Nathan Hand writes:
>
> > Not quite laughable... With all the recent discussion I was talking to a
> > friend of mine, and we were discussing the same idea. Although, to me,it
> > makes more sense to make it a compile time option, much like the network
> > packet scheduler... Then those who try something new will not necessarily
> > have thier system hosed if the kernel can't load the module.
>
> A solution here is to have a default round-robin scheduler that is
> active when no "modularised schedulers" are loaded. This could let
> you do something like
>
> % rmmod qnx-scheduler
> % insmod lb-scheduler
>
> Without due concern. The default round robin scheduler need not be
> smart - it wouldn't need to be - it would just need to be fair and
> free from starvation.
>
> However I don't think modularised schedulers are nearly as trivial
> as you and the poster before you are suggesting.
Agreed. The proposal I saw looks really fancy and powerful, but I
suspect it's awfully bloated. Just to have the basic framework in
place would probably make the core scheduler 10 times slower than a
simple scheduler (like Linux).
The main reason my separate RT run queue patch is mostly cost-neutral
(with a small bias to better performance) in the general case is
because I was able to take out code from the critical path.
Regards,
Richard....
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/