Was PGP Source, Now : Copy Protection

[Ken Brown]

David_Biggins@usermgmt.com wrote:
> 
> The chances are indeed excellent that it does not intervene on an Add or
> any other non-privileged instruction - on current architectures, short
> of interpreting my entire code, could it sensibly do so?  Without
> crippling performance?

It depends, of course.

There have been operating systems which could intervene on such. Some
versions of IBM's old VM mainframe system, on some hardware, effectively
interpreted ALL user-issued instructions opcodes, and IIRC all versions
could be made to do so with a bit of tweaking. 

Some ADDs will be intercepted, in a multiprogramming environment. And of
course any instruction that addresses memory should be interpreted by
the virtual memory management.

I should think that any real OS could be made to intercept an ADD or any
other command if you could get at the source code. Of course without the
close handshaking between chip & code that would mess up performance
except that not many applications are CPU-bound these days and not many
of them are limited by the load-from-cache parts of the cycle. VM
performance was better than you'd expect (byte for byte it beat MVS on
the same hardware, not that that is difficult) but of course the OS,
microcode, and CPU were being designed by the same people to work
together & there were various tweaks to allow user code to be dispatched
directly in some circumstances.

As an aside, in a world of worms and viruses and Evil Spam,  that sort
of completely sandboxed virtual machine architecture is the only sort
that one really ought to trust for security and reliability. 

Ken