Subject: RE: PDP-11 thoughts
To: Johnny Billquist <bqt@Update.UU.SE>
From: None <allisonp@world.std.com>
List: port-vax
Date: 11/14/2000 09:20:13
On Tue, 14 Nov 2000, Johnny Billquist wrote:

An aside, I"d contend the the instruction set for PDP-11 is far better
than 80x86 (vastly better than 8086).  So the memory footprint would
be better.  that and PDP-11 And VAX instruction set are better matched
to C as a higher language.  

> Apart from the relocation, you can also set different protections on
> different pages, and also limit the actual size of a page to less than 8k,
> and even decide in which direction the page grows.

What about I&D space allowing concurrent data and code spaces?  PDP-11s
(some) permit this for greater space than would be available as 16bit
bare.  

> 
> Unless my brain fools me, segmented addressing works slightly different.
> I'm definitely no expert on the x86 architecture, so please correct me if

Your right. 

> In a segmented architecture you have one (or  perhaps two) registers which
> adds a constant to all your memory references. This register might also

For 8086 and kin you have CS(code), DS(data), SS(stack) and ES(extra) all
16 bit but left shifted and added to the 16 bit address.  This has a
granularity of 16bytes.  It allows a possible addressing (assuming no
overlap which is permitted)  of 64k per segment. 


> have some granularity to it. You cannot protect parts of the memory, and
> the code/data still have to be contigous, wherever it happens to be
> located. Also, I believe anyone can play with the segment registers, they
> are in no way protected agains manipulation.

The 80x86s (386 and above) add all sorts of other weirdness).  My $.02
is the X86 is not a pretty thing compared to PDP-11 or especially VAX
both of wich are more regular (orthoginal) in structure.  That really
makes the Linux (and other) code grow.  Comparisons are then senseless
to a degree.
 
Allison