Newsgroups: comp.os.minix
Path: utzoo!utgpu!cunews!bnrgate!bmerh408!bmerh451!dgraham
From: dgraham@bmerh451.bnr.ca (Douglas Graham)
Subject: Re: old minix articles
Message-ID: <1991Mar13.052041.5276@bmerh408.bnr.ca>
Sender: news@bmerh408.bnr.ca (Usenet News Admin)
Organization: Bell Northern Research, Ottawa, Canada
References: <1991Mar12.093943.21936@kithrup.COM> <1991Mar12.195103.284@ccs.carleton.ca>
Date: Wed, 13 Mar 91 05:20:41 GMT

In article <1991Mar12.195103.284@ccs.carleton.ca> jstewart@ccs.carleton.ca (John Stewart) writes:
>>	From: graham@sce.carleton.ca (Doug Graham)
...
>Weird!  The machine that this article originated on (sce) stopped running
>News last September or thereabouts and was entirely shutdown in January.
>These are truly "ghost" postings!

The machine maybe, but the rumours of *my* death are greatly exaggerated.

Since I'm here, I may was well talk about Minix a bit.  I recently bought
a 386 with piles of RAM, and Minix is much more fun on it than it was
on my old XT.  It sure is nice to be able to compile GNU software straight
out of the box.  However, the single threaded file system still manifests
itself in quite noticable ways.  For instance, even with a fast hard disk,
it can take quite a while to flush a 300K buffer cache to disk.  During
this time, not much else happens.  This is especially noticable when using
a shell (or editor) that puts the console into RAW mode, and does it's own
echoing.

Oh, and one question that I'm surprised I haven't seen asked here before.
I'm sure I must be missing something obvious, but what do I need to
do to speed up the typematic rate on the AT keyboard?  Not knowing anything
about the innards of the keyboard controller (yet), I tried making a BIOS
INT 16h call somewhere in the code before the switch to protected mode,
and before the BIOS vectors had been moved, but this only caused the
keyboard to freeze up.
---
Doug Graham dgraham@bnr.ca	Bell-Northern Research, Ottawa Ontario Canada