Ken writes:
> The objective of this emulator would be to PRECISELY
> emulate the PA-RISC instruction set.

The instruction set is well-defined and presents few *functional* challenges
(it's difficult to get the best *performance* out of because of the much
larger machine state that's exposed to the programmer and OS in a RISC
architecture when compared to, say, a VAX).

The tricky bit is the emulation of the rest of the HPe3000 *system*, the
chipsets, I/O busses, I/O interface cards, and peripherals that are required
to turn a bare PA-RISC CPU chip in to an HPe3000 computer system.

So for example I would want to be able to emulate things like the bus
interface to a SCSI card, the SCSI card, the SCSI DISK attached to the SCSI
card (for doing virtual disk drives), etc.

I need to be able to emulate the hardware interfaces to the LAN card and map
the raw network frames onto a raw host network interface or decompose the IP
traffic and do dynamic NAT translation onto an available Layer 4 host
interface.

There is firmware that sets up the system environment that needs to be
emulated/replaced or perhaps licensed from HP (actually emulating its
functionality is pretty easy).

One of the issues that is yet to be addressed is that in the original
scenario where the emulator maker(s) entered into an agreement with HP, HP
would have been asked to agree to provide documentation, source-code access,
etc., to assist in the development of the emulator.  I presume that if HP
wants the concept of the emulator to proceed that they will still be willing
to provide this assistance to some reasonable number (at least two :-)
prospective emulator developers.

Gavin