Re: Task destruction

[Neal H. Walfield]

> The easiest approach (for the kernel/task server) is if there is
> just some magic token (send right in Mach) that is inherited from
> task to task and which can only be changed (or fetched in the case
> of a Mach send right) by something privileged like a proc server.
> Then any unregistered task is accounted as belonging to whatever
> registered task ancestor it had, such as an entire subhurd belonging
> to boot.

I had started to consider this problem, however, had not gotten very
far.  This approach does make a lot of sense.  The only open issue
then is how to determine who the privileged server.  Do you have any
ideas?  Should this just be some parameter?  The first to register as
such?  Or perhaps, you had something else in mind.