There was no concept of name/address space or device virtualization in the Lisp Machine operating systems. I don't know the instruction set, but I have never seen references to privileged instructions or protection rings. Not that that would be an impediment - you would really want to do virtualization at the software level in a Lisp OS anyway.
They had a single address space, tasks were more like threads in a modern system.
The MIT/LMI/TI ones didn't take interrupts, the microcode needed to poll the interrupt register at suitable points.
Re:The Scheme Machine (1994) [pdf]
by mrguyorama on Tuesday August 07, 2018 @04:14PM
(#17710286)
It could have been due to their lack of speed and performance. Can't run multiple users' software if it can barely run one user's software
Re:The Scheme Machine (1994) [pdf]
by convolvatron on Tuesday August 07, 2018 @04:58PM
(#17710737)
more importantly the whole point of the symbolics was the environment. one hi definition mono display and one space keyboard meant one 3600
How is this a limitation in the hardware? Operating system could switch between tasks using context switching right? Or are you suggesting these machines didn't have any mechanism to make sense of time/ticks or didn't have interrupts?