On Sun, 3 Jan 1999, Andrea Arcangeli wrote:
>
> I think it will make more sense to not inherit the avg_slice from the parent.
No, I actually think it makes tons of sense. If only because we really
don't have any better guesses (just zeroing it is worse than guessing it
will behave similarly).
Ok, we don't have a precise better guess, but the situation is real.
It's wildly inaccurate for very common activities, even a kernel
build. Think about what happens to make, gcc, and their children as
they each begin to do their work.
One thing that _may_ make sense is to decide that when we "execve()" our
behaviour changes, and as such avg_slice might be different. However, I
again do not see that zeroing it is the right answer. I don't know what
the right answer is, but I bet that leaving it alone is likely to be "more
correct" than most other answers.
Here is a quick idea. How about halfing it? Then in about 2
timeslices we'll be approximate, even less if the child's average
slice is significantly smaller than the parents.
Just an idea, it's probably way off... so someone should test it
before considering it seriously.
Later,
David S. Miller
davem@dm.cobaltmicro.com
-
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/