Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!ucsd!sdcc6!sdcc19!sdcc15!pa1343
From: pa1343@sdcc15.ucsd.edu (John J. Marco)
Newsgroups: comp.os.minix
Subject: Re: Minix 1.3 and printers
Summary: Hercules card
Keywords: Printer, Hercules monograf card
Message-ID: <424@sdcc15.ucsd.edu>
Date: 9 Jun 89 20:18:56 GMT
References: <3096@ogccse.ogc.edu>
Reply-To: pa1343@sdcc15.UUCP (John J. Marco)
Organization: University of California, San Diego
Lines: 43

In article <3096@ogccse.ogc.edu> lyle@ogccse.UUCP (Lyle Cool) writes:
>I have Minix 1.3 (640K PC) running (finally) except for the printer
>driver. The printer works fine under DOS, but Minix doesn't seem to
>know that the printer is connected. Every attempt to use it results
>in :
>printer out of paper
>printer not on line
>printer error
>I have a Panasonic 1091 connected to the parallel port of a typical
>XT clone. If anyone can offer help, it would be most appreciated.
I'll bet you have a hercules card.  Why?  I had this same problem also
and only recently found out how to fix it.  The problem is that the 
IO address of the p[rinter port is calculated using the type of video card
that is in use.  two marcos are defined for this: PR_COLOR_BASE and 
PR_MONO_BASE.  Normally, get_chrome determines whether the current video card
is color or mono (or EGA).  for the color cards it reports a color adapter
to minix.  printer.c then uses this to set the printer IO port to 
PR_COLOR_BASE.  For mono cards, monocrhome is reported by get_chrome() and
then printer.c uses that to set the printer port to PR_MONO_BASE.
Here's the catch:
If get_chrome is called and the video is Hercules or compatible
it will return the monochrome value.  printer.c will then set the 
port address to PR_MONO_BASE.
BUT, after testing to find out what the printer port is with a hercules
    card I found out that the hercules card  makes the printer port
    address be PR_COLOR_BASE.  Thus the WRONG port address is being used
    for the printer port and the printer doesn't work.

My kludge to this problem was to define PR_MONO_BASE to be the same value
as PR_COLOR_BASE.  This gives the printer driver no choice but to use the
color printer IO address like the Hercules card requires.  I am considering
putting a small routine in klib88.asm to use the BIOS to find the printer
port, thus eliminating this problem altogether.

I hope this helps.

>Thanks,
>Lyle Cool

+-----------------------------------------+
| John J. Marco -- pa1343@sdcc15.ucsd.edu | 
| ...!uunet!sdcc15.ucsd.edu!pa1343	  |  
+-----------------------------------------+