From mboxrd@z Thu Jan  1 00:00:00 1970
Message-ID: <001a01bfd8ee$bf9df740$8105d2cc@san.rr.com>
From: "Eric Dorman" <edorman@san.rr.com>
To: <9fans@cse.psu.edu>
References: <3948a500.11168068@news.arnhem.chello.nl> <8idb8s$1oq8$2@pandora.alkar.net>
Subject: Re: [9fans] Re: S3 775/325
Date: Sat, 17 Jun 2000 23:30:50 -0700
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Topicbox-Message-UUID: c2789324-eac8-11e9-9e20-41e7f4b1d025

----- Original Message -----
From: "Wladimir Mutel" <mwg@alkar.net>
To: <9fans@cse.psu.edu>
Sent: Friday, June 16, 2000 6:57 AM
Subject: [9fans] Re: S3 775/325

> Oscar Gala <proto@phatspace.org> wrote:
> > I have a S3 Trio64 videocard with the 86C775 chipset and a S3 Virge
> > with the 86C325 Chipset,
> > but I can't get them to work with the Plan9 installation.
> 86c775/86c785 (Trio64V2) works. I just added such line (marked with
> +) in /lib/vgadb (and before, in vgadb on boot diskette):
[xx]hi,

Gotta be careful here, because an 86C775 isin't quite a Trio64
but has some more regs that might have to be filled in.  An example
is the C775 on the HP Vectra VL5/166.  On that box regs
0x90 and 0x91 didn't get filled in correctly, resulting in the
left half of the display repeated twice.  I did the following:
init:
 vga->crt[0x90] = 0x0;
 vga->crt[0x91] = 0x7F;
load:
 vgaxo( Crtx, 0x90, vga->crt[0x90] );
 vgaxo( Crtx, 0x91, vga->crt[0x91] );
looking over the docs today leads me to think that 0x90
is sufficient.

The BIOS string is also different (ah, love HP sometimes).
vgadb:
    0xC004E="86C775"

I have a true-blue Trio64 card to compare against to make
sure I get the correct identification so real regs get filled in,
then I'll generate a patch.

--eld