Re: thread ids, task ids and subsystems

[Niels Möller]

Marcus Brinkmann <address@hidden> writes:

> On Mon, Apr 07, 2003 at 06:41:05PM +0200, Niels Möller wrote:
> > For a start, memory and cpu usage should be included in the counts for
> > the process that the task "belongs to".
> 
> That's probably harder to do (and more cpu intensive) than just listing each
> task individually.

I think it's the natural way to do limits and accounting of
unregistered tasks. The accounting id (associated with each task) is
all information task needs to do it. But I guess we'll see later if
it's feasible or not.

For instance, consider a process that spawns lots of short-lived cpu
intensive tasks. Their cpu usage must be added up and associated with
some living object (the process), and that can easily be done by task.
I don't think we want unixish concepts like waitpid and zombies on the
task level.

Regards,
/Niels