Path: utzoo!attcan!uunet!mcvax!ukc!stl!stc!datlog!gis
From: gis@datlog.co.uk ( Ian Stewartson )
Newsgroups: comp.os.minix
Subject: Minix 1.2 - Some Problems ?
Message-ID: <1188@dlvax2.datlog.co.uk>
Date: 17 May 88 14:23:00 GMT
Reply-To: gis@datlog.co.uk ( Ian Stewartson )
Organization: Data Logic Ltd, Queens House, Greenhill Way, Harrow, London.
Lines: 44

I am in the process of setting up MINIX 1.2 on our PCs (IBM XTs and 
DELL System 200 - AT Compatible) and have come up against a number of
problems.

1)  When I execute a rebuilt MINED, it generates a trap at location 0xe6a
    which is a mov (bx), #0 instruction.  The trap number is 13 - General
    Protection Violation.  Any ideas ?

2)  I have tried to port EMACS to MINIX.  I realise this is probable a
    major problem given the current limitations of MINIX.  However, the
    EMACS main.c causes the Compiler to trap at location 0xf8dc in /lib/cem.
    I do not have the compiler sources.  Any suggestions?

3)  I had problems with the original tty driver (text following a space
    would be blanked until the space moved off the screen), so I turned to
    Jim Paridis's Console & tty driver which now works quite happily.
    However, Tom Frauenhofer (5281@cci632.UUCP/tvf@cci632.UUCP) mentions so
    mods to the driver.  Have I missed these or is he discussing Jim's
    driver itself?  As we have EGAs on our AT, I had to make some changes to
    get it to work, along with the changes to support a UK keyboard and
    supporting the keypad correctly.  Otherwise, Jim's drivers seem to work
    very well.  I can now connect to our VAX and I suspect the next set
    will be to generate either kermit or UUCP?? for MINX.

    As a side issue, how to you tell if there is an EGA board attached.  I
    know how to find out about MONO and CGAs.

4)  Has anyone improved the illegal interrupt interface.  Once you get one,
    you either never get out of it or have to press DEL to kill the process.
    The option available appears to depend on which interrupt vector is used. 

5)  When I was modifying kernel/clock.c to interface to the CMOS Real Time
    clock, I came across a problem with long division.  Sometimes I would
    get an interrupt 0 - divide error (by zero?).  This only happened when
    the time was between 30 and 50 minutes past the hour.  I have since
    resolved the problem by using repeated subtraction (which may be faster).
    However, has anyone else seen a similiar problem with long division.

Thanks in advance for any help.

Ian Stewartson
Data Logic Ltd, Queens House, Greenhill Way, Harrow, Middlesex, HA1 1YR, UK.
(Phone) +44 1 863 0383 (Telex) 888103 (Fax) +44 1 861 2010
(Network) gis@datlog.co.uk or ukc!datlog!gis