Can you run Sys V/AT on a 386 box?

[Steven C. Neighorn]

In article <225 at opel.UUCP> johnk at opel.UUCP (John Kennedy) writes:
>I have recently confirmed that the 286-COFF executables compiled on
>System V/AT seem to run reliably on a 386 machine running Sys V 3.2
>(AT&T 6386 WGS).  
>It would make sense, as the 386 is supposed to be a true superset of the
>286.

And the fact that there are special programs (/bin/[ix]286emul) that
help out with emulation. That's why 286 binaries used to need read
permission (ie r-xr-xr-x instead of just --x--x--x) or they wouldn't
execute. I believe this is fixed now, at least in 3.2 it is.

Also, not all 286 COFF binaries will run, even with the emulation. Two
notables were a hacked 5.51 sendmail (ioctl differences) and the development
system for Unify, a relational database package.

>Is there any reason that a Sys V/AT wouldn't run on a 386 CPU board?
>Granted, there are unique initializations for the 386 processor, 
>but shouldn't that be handled by the BIOS before boot?

There may be an equation for answering this question definitively, but in
my experience you can only go on a case by case basis. Microport's Sys V/AT
ran fine (ie normal 286 operation speed and reliability) on Intel OEM
platforms, but would not run on an Everex System 3000 '386 box. It might
have been BIOS related, and the machine in question has had a BIOS upgrade,
but the interested parties are no longer interested in V/AT.

I think the only way to find out for sure is to actually try it. Great advice,
I know, but as it has been said in the movies, "It's the only way to be sure."
-- 
Steven C. Neighorn           !tektronix!{psu-cs,nosun,ogccse}!qiclab!neighorn
Sun Microsystems, Inc.      "Where we DESIGN the Star Fighters that defend the
9900 SW Greenburg Road #240     frontier against Xur and the Ko-dan Armada"
Portland, Oregon 97223          work: (503) 684-9001 / home: (503) 641-3469