AC> Am I right in thinking this is only supposed to lock ISA 
AC> machines? Because I have an ISA bus and it just hung the 
AC> VDM, and I killed it from another session.

No, it locks older machines (usually ISA based) that don't have an
internal watchdog timer.  IBM PS/2s, most Pentium machines, and EISA
(don't know about PCI) machines DO have a watchdog timer.  What it is,
is a separate timer chip that is programmed to start a timing loop. 
Every few seconds the machine is supposed to send an interrupt to reset
the timer.  Thus the timing loop ordinarily never runs to completion. 
If however the machine is set to disable interrupts AND then goes into
an infinite loop the machine without a watchdog timer will lock (OS/2
will not be able to switch sessions since it uses timer interrupts to
implement the preemptive multitasking)  On a machine WITH a watchdog
timer, the timing loop doesn't get reset and then the timer chip
independent of the CPU resets the interrupt controller and enables
interrupts thus allowing your multitasking OS to switch sessions.  BTW
this is a problem with any OS that is running true DOS code rather than
an emulation.  It is a problem under Warp, Windows, and Windows 95. 
You can disable the ability of a nonWarp program to disable the
interrupt controller by setting IOPL to NO and for more protection
PROTECTONLY to YES.  You won't be able to run DOS/Windows programs but
if you have a mission critical server you don't want to run them
anyway.  Windows NT gets around this by emulating DOS and not allowing
anyone direct access to the interrupt controller.  The drawback is the
inability to run some DOS programs particularly games.  At least with
Warp you have the choice.  And with a newer machine it is a moot point
anyhow.  Lets face it guys.  There ain't no such thing as a "crash
proof operating system"  but you can have some "crash protection" to
use IBM's term.

---
 X KWQ/2 1.2i X Say again all after crackle snarf hiss . . .

--- Maximus/2 2.02