Re: excessive kworker activity when idle. (was Re: vma corruption intoday's -git)

[Linus Torvalds]

And with your fixed trinity thing, I can now do this:

On Wed, Mar 30, 2011 at 9:08 PM, Dave Jones <davej@xxxxxxxxxx> wrote:
>
> git clone git://git.codemonkey.org.uk/trinity
> cd trinity ; make
> ./trinity --mode=random -i
>
> let it run for a while, then kill it.   kworker should then go nuts.

I really thought "hey, that reproduced it", because "top" did indeed
show kworker going nuts after a while. And perf did show it doing all
that

                   |          queue_work
                   |          schedule_work
                   |          flush_to_ldisc

that looks like the re-scheduling.

But when I killed trinity, kworker immediately calmed down too.

Which I'd actually expect, since any tty it is using (and making cause
problem) should calm down when it's closed - even if it was stuck
before.

But maybe I haven't triggered just the right condition to actually get
it into some really infinite case.

Regardless, I'll put my money where my mouth is, and try to remove the
crazy re-flush thing.

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