Re: [patch] lockless, scalable get_pid(), for_each_process()

[Ingo Molnar]

On Wed, 18 Sep 2002, William Lee Irwin III wrote:

> > i agree that this is okay, as an added mechanism. Nevertheless this does
> > not eliminate the 'box locks up for seconds' (or even minutes) situation.  
> 
> The lockups I see range from hours to "it spun over the weekend, time to
> pull the plug".

this can happen if there's a genuine PID space squeeze wrt. nr_threads -
that is solved by adding Linus' suggestion to the PID allocator. I believe
you saw that problem, not any inherent get_pid() algorithmic inefficiency.

nevertheless we do lock up for 32 seconds if there are 32K PIDs allocated
in a row and last_pid hits that range - regardless of pid_max. (Depending
on the cache architecture it could take significantly more.)

	Ingo