Re: GEM and memory protection

[Julian Reschke-FJR010 <Julian_Reschke-FJR010@email.mot.com>]

schwab@LS5.informatik.uni-dortmund.de%INTERNET wrote:
> 
> Konrad Kokoszkiewicz <draco@mi.com.pl> writes:
> 
> |>
> |> > > That would be a reasonable compromise (IMHO) between an unreal dream (full
> |> > > protection) and hard reality (no protection).
> |> >
> |> > I really have a problem here: it means that correctly written programs
> |> > are
> |> > punished (real MP is taking away), just because some other programmers
> |> > don't code properly.
> |> >
> |> > If you really think that this needs to be done, PLEASE make it
> |> > configurable.
> |>
> |> The better solution would be to protect the shell against being killed
> |> because of a faulty AV-client.
> 
> Accessing it in supervisor mode should do it.  In supervisor mode you can
> access all memory attached to any process.

That's news to me.