Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!zephyr.ens.tek.com!gvgpsa!gold.gvg.tek.com!shaunc
From: shaunc@gold.gvg.tek.com (Shaun Case)
Newsgroups: comp.os.msdos.programmer
Subject: Re: 8088 vs 8086
Message-ID: <2601@gold.gvg.tek.com>
Date: 22 Jun 91 23:45:57 GMT
References: <1991Jun21.030948.10951@cs.mcgill.ca> <1991Jun21.173218.1912@mccc.edu>
Organization: Grass Valley Group, Grass Valley, CA
Lines: 30

In article <1991Jun21.173218.1912@mccc.edu> pjh@mccc.edu (Pete Holsberg) writes:

(re: 8086 vs 8088)

>
>For all practical purposes, yes.  I think that there's slightly
>different instruction fetch queue, but neither a user nor a programmer
>would be aware of it.

Someone mentioned that they are different lengths -- what if you have some
self modifying code that modifies something that is already in the queue?
Do the 8086/8 catch this and re-fetch (or modify within the queue) the 
altered instructions?

It is my understanding that running self-modifying code is a definite 
no-no on the 386 and 486, although I think this is due to pipelining
rather than a fetch queue.  

Am I right or wrong?

(Please, no style flames on SMC, I don't write it (any more...)  :-> )

Shaun


-- 
shaunc@gold.gvg.tek.com  

-- 100,000, perhaps 200,000 or more Iraqis died in a "Turkey Shoot"
   inappropriately called a "war."  -- Michael Albert