Re: [patch] lockless, scalable get_pid(), for_each_process() elimination, 2.5.35-BK

From: Rik van Riel (riel@conectiva.com.br)
Date: Wed Sep 18 2002 - 12:00:19 EST

On Wed, 18 Sep 2002, Linus Torvalds wrote:
> On Wed, 18 Sep 2002, Rik van Riel wrote:
> >
> > On second thought ... yes there's a reason. Suppose you have
> > 100000 threads on your box already, how long is it going to
> > take to walk them all to figure out the pid distribution ?

> The pid space is not a uniform distribution, which your made-up-example
> depends on. So you usually walk the 100000 threads _once_, and then you
> don't have to walk them again for quite a long time.

Agreed, you're right there. On the other hand, walking the threads
_once_ will take 1.5 minutes on a 500 MHz PII (according to Ingo's
measurements).

That's about 18 times the timeout for the NMI oopser and will cause
people real trouble.

cheers,

Rik 

-- 
Spamtrap of the month: september@surriel.com

http://www.surriel.com/		http://distro.conectiva.com/

-
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/

This archive was generated by hypermail 2b29 : Mon Sep 23 2002 - 22:00:23 EST