Ralf Baechle (ralf@julia.de) writes:
> There is a project to port Linux to the 8086 ... The project makes
> use of segment pointers which makes it a bit more difficult to port
> the code ...
Of course Unix started on PDP-11s, which didn't have virtual memory
hardware [and would support 15 users on a 48Kbyte machine, but that's
another story]. But they did have translation on a large-page basis.
8086 segments can provide something similar, so long as your
applications can live in a 64K I + 64K D space.
You could conceivably build a usable system on an untranslated CPU with
position-independent code. But the MIPS architecture is hostile to
position-independent code too...
ELKS is in danger of addressing a vanishing hardware base. Embedded
systems seem to be separating into ultra-low-cost 8-bit and 4-bit
systems, which are far too small to provide anything like POSIX; and
bigger systems, where 16-bit CPUs and machines without memory
translation are rapidly dying off - it just isn't worth the software
hassle for the miniscule savings involved. IDT's R4640 is an oddball
throwback, which bears the dread marks of a single customer's
requirement. NEC's Vr4300 appears to be more successful, partly
because it offers complete R4000 functionality at a similar price.
Regards,
Dominic Sweetman phone: +44 171 700 3301
Algorithmics Ltd home: +44 171 226 0032
3 Drayton Park fax: +44 171 700 3400
London N5 1NU email: dom@algor.co.uk
ENGLAND. www: http://www.algor.co.uk
ftp: ftp.algor.co.uk
|