On Wed, 2006-04-05 at 09:48 +0100, Stefan Seufert wrote:
> But we still should think about some mutexes protecting the
> manipulations of the processor status. The chance that a race conditions
> will happen seems to be small and the consequences seem to be benign but
> still operations such as
>
> if (processor->status == CHILD_STATUS_STANDBY)
> {
> _DBG("Activating child #%d", processor->id);
> processor->status = CHILD_STATUS_STARTING;
> }
>
> should be protected. I think the APR has some useful stuff in that area.
A lot of that stuff can probably be removed - there's a lot of cleanup I
haven't gotten around to yet. I wrote a lot of hacky code when I was
first getting this to work; I was under the (mistaken) impression that
each worker process needed its own socket, and they were created late
on, during processor startup. Now that each ServerEnvironment has a
socket that's created early on, there shouldn't be as much need for all
this keeping-tabs-on-everything code. Hopefully...
--
Sean Gabriel Heacock
Telana Internet Services
http://www.telana.com/
_______________________________________________
Peruser mailing list
[email protected]
http://www.telana.com/mailman/listinfo/peruser
