Re: Pipelines with infinite recursion

"Toman_Vojtech@emc.com" <Toman_Vojtech@emc.com> writes:
> Detecting infinite recursion in any pipeline that is more complex than
> the (degenerated) example above would be difficult (recursion may be
> driven by dynamic logic), but perhaps we should at least say that the
> behavior of the processor is implementation-dependent in this case? Or
> do we want to have a dynamic error that could be used for reporting
> infinite recursion?

We have a warning in 2.12, Security Considerations, where we say a
"conformant implementation may raise dynamic errors, or take any other
corrective action, for any security problems that it detects."

I don't think we need to do any more in the spec.

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | Society is immoral and immortal; it can
http://nwalsh.com/            | afford to commit any kind of folly, and
                              | indulge in any kind of vice; it cannot
                              | be killed, and the fragments that
                              | survive can always laugh at the
                              | dead.--Henry Adams