Subject: Re: SE/30 grf mmap() fix
To: Denny Gentry <denny1@home.com>
From: Scott Reynolds <scottr@og.org>
List: port-mac68k
Date: 03/17/1997 14:09:36
On Fri, 14 Mar 1997, Denny Gentry wrote:

> Masami and Ken Nakata wrote:
> > > in grf.c:
> > > 374c374
> > > <     *addrp += (unsigned long) gp->sc_grfmode->fboff & 0xfff;
> > > ---
> > > >     *addrp += (unsigned long) gp->sc_grfmode->fboff;
> > 
> > But I'm not sure of this one.  I vaguely remember this is necessary on
> > some (old, I think) models (probably those with 32-bit dirty ROM).
> 
>   Thats what I thought at first too: the SE/30 fboff was 0x8000,
> which looks a lot like a flag bit, and the SE/30 has dirty ROMs.
> However I've concluded that the offset really is 32767. It has two
> banks of video RAM, seperated by 32K, and the primary bank is the
> upper one.

Is that a typo and you meant 32768, or did you really mean 32767?  Did you
come up with an alternate solution?  I'm really not to keen on applying
the patch above until I understand exactly what it's accomplishing.

Ken, yeah, you sent me some information on that but I haven't been able to
unwind all of the data surrounding this yet.  Other things have taken a
higher priority, unfortunately.

--scott