Re: stress testing and loadavg

linux kernel account (linker@nightshade.ml.org)
Sat, 9 Aug 1997 16:25:48 -0400 (EDT)


On Tue, 5 Aug 1997, Alan Cox wrote:

> > its minimum run-queue. I hope people would realize that putting in system
> > calls just to get back to the minimum run-queue is a brain-damaged thing
> > to do. Of course, I'm sure compiler writers realize that optimizing for
>
> Doesn't hold water. If I do stupid tricks to get 99.95% of the CPU time
> by misusing scheduling tricks and kill the machine because I know its
> a brain damaged scheduler ..
>
> Alan
>

Every time there is a new and intresting feature submitted people start
making assumptions about what is broken before they try the patch out.. If
you had bothered to pay attention before commenting you would have found
out that a process cannot excede it's initial run que under any
circumstnace... So calling a bunch of syscalls to be labled as nice could
only get you where you started.. And if you do something like that then
start while looping you will quickly be demoted.. Important tasks that eat
cpu (number crunches, X, mpeg players) should be run with nice -1.. I've
created a nice C program that when suid0 users can use to make their
important tasks nice -1... Otherwise the tasks might be trampled.. No
biggie.. Maby you should just make users normally run at nice 2 so they
can nice 0 and nice 1 their own tasks, no need for funny suid wrappers..

The non-exec patch was trampled with this sort of invalid complaint.. If
you think something is broken: Test it. make a parch for it, or ignore it.
If it was crap in the first place then dont worry about it..