Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!yale!cmcl2!acf5!bxj2161
From: bxj2161@acf5.NYU.EDU (bxj2161)
Newsgroups: comp.os.minix
Subject: Re: Compiling MINIX 1.3 with Turbo C 2.0
Message-ID: <1179@acf5.NYU.EDU>
Date: 8 May 90 17:35:26 GMT
Reply-To: bxj2161@acf5.UUCP ()
Organization: New York University
Lines: 21

In <1990May8.163109.26727@bach.amd.com> ching@pepsi.amd.com (Mike Ching) writes:
>........
>I may be misunderstanding what is being done here but I was under the
>impression that the BIOS routines couldn't be used since they are
>not reentrant. I've been looking at porting curses and it would be
>trivial if I could use interrupt 10 like the original does. Has
>anybody already ported curses? It all started when I decided to port
>pcomm which requires curses which requires ???.

Well, some BIOS routines are not reentrant, some are, but it doesn't matter.
MINIX is a special operating system, it uses message passing instead of
shared memory.  So as long as you obey the rule that no two tasks
use the same BIOS int call, there will never be reentry.  (The key point is
that tasks will not be reentered.)
In fact, I'm using int 21 for my MS-DOS file access, and it works happily.
The next step I am going to try my system on RLL, SCSI, and ESDI harddrives;
If these succeed, people out there won't need to worry about HD interfaces
and partitions if they don't want to.

Benchiao Jai
jai@cs.nyu.edu