20 Feb
2025
20 Feb
'25
5:34 p.m.
On Feb 20, 2025, at 5:04 PM, Mike Katz via cctalk
<cctalk(a)classiccmp.org> wrote:
The advantage of multiple stacks is that the system stack runs ring 0 protected tasks
(main operating system tasks). The user stack runs the user(s) tasks.
On systems with some kind of memory protection this keeps any non system task from
accessing any other non-system tasks memory and possibly keeps any individual system task
from accessing any other system (or non-system) tasks memory. On these systems, usually
an MMU is involved and an interrupt or trap instruction will cause the MMU to switch its
memory map to the system stack. Usually all interrupts go through the OS even if the
eventually interrupts handler code is in user memory.
That reminds me that the PDP-11 has the ability to send interrupts to modes other than
kernel mode, because the new processor status (which includes the mode) is part of the
interrupt vector. I doubt that has ever been done, though. One complication is that an
interrupt into user mode that interrupts kernel mode code would be a dead end, because the
RTI instruction can't restore the state then: in user mode it can't restore a PSW
value that says kernel mode. So a user mode vector would work only if the kernel is not
interruptible at all.
paul