🚀 go-pugleaf

In article <id.9FYU.UBJ@ferranti.com> peter@ferranti.com (peter da silva) writes:
>In article <1992Nov11.134556.20150@super.org> rminnich@super.org (Ronald G Minnich) writes:
>> Only if you don't care about protection, right?
>From an interrupt routine???

What can I tell you. The people who write the OS for these machines, 
at least a few years ago, told me they switch stacks on interrupt. One reason 
given is protection.  Ask them. Any of you MCP hackers reading this?

If memory serves so did the HP3000. Any HP people care to comment?

>In any case, the Burroughs A-series did security entirely in software. The
>hardware memory protection was to enforce type correctness, not security
>policy.

Not really. First off, don't extrapolate what you may have read in Bell&Newell
about the 6700/7800/7900 to the A-series. They are not the same. 
Secondly, past some very early machines there was always bounds checking to 
make sure you stayed in your stack. So not "entirely". 
Thirdly, with the advent of E-mode in 
ca. 1985 or so you find a set of memory mapping registers which look much like
what you find on PDP-11s that ran out of address bits. And for the same reason:
ca. 1982 or so Burroughs began to realize that 20 bits of (48-bit word) 
address was not going to be enough for much longer. 
My E-mode spec. is ca. 1983 and an internal document anyway, 
so is hopelessly out of date, 
so possibly someone more current can comment here on what E-mode provides
in the way of mapping/protection. 

Finally, the correct tense (for now :-) ) is "the A-series DOES". :-)
ron
-- 
you want windows? I got windows. 4608 by 3600,    | rminnich@super.org
16 channels of sound, 4' by 6'. Fun.              | (301)-805-7451 or 7312
Most abused words: open, scaleable, [more later]  |