On Sun, 11 May 2008, Ingo Molnar wrote:
quoted text
> 
> but the problem is that by serializing the wakeup chains naively you 
> introduced a more than 50% AIM7 performance regression.

No, the problem is that the BKL shouldn't be a semaphore at all. 
Performance fixed.

After that, it's *purely* a correctness issue, and then we might as well 
be fair and not allow the stealing of semaphores from under waiters at 
ALL. Which is what Matthews original code did.

In other words, current -git is fine, and was already confirmed by Sven 
to fix the bug (before *any* of your patches were), and was earlier 
confirmed to fix the AIM7 performance regression (better than *any* of 
your patches were).  

So I fixed the problems last night already. Stop wasting everybody's time.

		Linus
--
unsubscribe notice
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/