Re: tasklet_kill will always hang for recursive tasklets on a UP

[Juergen Quade <quade@hsnr.de>]

On Tue, Aug 26, 2003 at 04:38:02AM -0300, Werner Almesberger wrote:
> Nagendra Singh Tomar wrote:
> > I fail to understand this. How can we say that its not a bug. If we 
> > support recursive tasklets, we should support killing them also. If we can 
> > do it why not do it. Is there any reason for that.
> 
> It's just a question of how you define "to kill" :-) But the
> naming is ambiguous, because people may indeed expect
> tasklet_kill to work like kill(2).
> ...
> Example: if a tasklet allocates some resources, and frees
> them when running the next time, you'd need a flag that
> tells the caller(s) of tasklet_kill whether there are
> still such resources that need freeing.

Is it really used in this way somewhere?

I always thought, a tasklet is a self-contained
(lets say stateless) function.

For more we have kernel-threads.

> The current mechanism makes sure that the tasklet will
> execute one last time, if scheduled before tasklet_kill.

If your tasklet has states, you can't know, in which
state it is, when you call "tasklet_kill".
Can this work reliable?

        Juergen.