Jonathan S. Shapiro
shap at eros-os.com
Tue May 22 13:54:56 EDT 2007
On Tue, 2007-05-22 at 13:23 -0400, Christopher Nelson wrote:
> > It also seems like it could induce false sharing between
> > threads and processes, where delays in fault handling by
> other threads
> > or processes could cause those to propagate and slow down
> groups who's
> > processing is otherwise unrelated. Of course, this assumes
> that you
> > are using these handlers for various types of notification
> > memory mapping / etc.)
> Now that it is clearer how the fault handler works, do you
> still have
> this concern?
> No. It seems that the only concern I would still have is that
> external fault handlers seem very heavy in comparison to activation
> handlers. Therefore, you are in a space/complexity trade-off. If
> these critical paths that activation handlers are blocking are so
> systemically significant that using external fault handlers will
> provide a boost sufficient to negate the memory overhead - then I say
> go for it.
Umm. Chris? The memory overhead of the activation handler is actually
*higher* than the memory overhead of the external handler...
> It's a shame, though. I was looking forward to using activations as
> lightweight asynchronous event handlers. :-)
How lightweight they are is entirely a function of the hardware
user/supervisor crossing overhead.
More information about the coyotos-dev