Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watnot!watmath!clyde!rutgers!husc6!seismo!mcvax!botter!ast
From: ast@botter.UUCP
Newsgroups: comp.os.minix
Subject: Re: MINIX and the Tandy 1000
Message-ID: <1135@botter.cs.vu.nl>
Date: Wed, 8-Apr-87 05:22:16 EST
Article-I.D.: botter.1135
Posted: Wed Apr  8 05:22:16 1987
Date-Received: Sat, 11-Apr-87 07:14:02 EST
References: <1127@botter.cs.vu.nl> <472@moncol.UUCP>
Reply-To: ast@cs.vu.nl (Andy Tanenbaum)
Distribution: world
Organization: VU Informatica, Amsterdam
Lines: 26
Keywords: sob, whimper

In article <472@moncol.UUCP> john@moncol.UUCP (John Ruschmeyer) writes:
>Summary: MINIX won't run on the Tandy 1000.
>	Memory management on the 1000 is a tad different than on a real PC-
>	actually it's closer to that of a PCjr. In short, there is no dedicated
>	video memory; rather the top 16K is stolen for this purpose and somehow
>	mapped into the 0xB800 segment. Thus, ny 640K Tandy 1000 is really a
>	624K machine.
A number of people have sent me mail about problems with the Tandy.  Maybe
the problem is as follows.  MINIX expects a video RAM at address 0xB0000 if
the motherboard switches say there is a monochrome adapter installed and
at 0xB8000 if there is a color adapter.  If it tries to write to those
addresses and nothing is there, the computer will just hang.  Conceivably
something is going wrong with the mapping.  I would suggest writing a little
test program under MS-DOS or some other system that just writes characters
to 0xB8000 or 0xB0000 to see if that works.

Another problem is that MM may think memory is 640K instead of 624K.  It
may be necessary to patch something around line 5517.  Note that get_tot_mem
is an assembly code routine in the directory lib.

Finally there is the issue of whether the disk controller chip is really
compatible.  Probably the best thing to do is modify the MINIX disk driver
to make it a free-standing test program, compile it under MS-DOS or XENIX
or something else, and experiment.  

Andy Tanenbaum