Subject: Re: Feedback on Dec. 12 current kernel
To: Allen X Briggs <briggs@csugrad.cs.vt.edu>
From: Michael Finch <finchm@cray-ymp.acm.stuorg.vt.edu>
List: macbsd-general
Date: 12/13/1993 02:37:21
>Allen X Briggs
> 
> > 
> > The "a" in the first line has a "'" over it.  It looks like a
> > high-ascii char.  I am booting on the display in slot 11 (a Rasterops
> > 264SE/30 card); this text is displayed on a 640x480 screen.  What's
> > up?  Should all screens show up here?
Do you mean Virtual terminals?  There is a kick butt program that
allows you to display different Virtual Terminals on different
monitors, its called con.  Normally, all Virtual Terminals appear
on the one grf device grf0.

> 
> Argh!!!  Yes.  The nubus line should look something like:
> grf0 at nubus0: 640 x 480 8 color MacVideo (RasterOps 264SE/30) display
> and the grf0 should be grf1.  The problem seems to be that we're not
> reading the info from the RasterOps card correctly.  I'll defer this
> to Mike "NuBUS" Finch...
The output looks radically different from the printf's I originally coded...
I'd tend to agree with Allen "Kernel Panic" Briggs that the RasterOps card
is being read incorrectly.

> > Is there any data about the Rasterops card that you need to know?
I'd like to see the output from the Apple Developer program "Slots",
available on your better "Develop" CD's.

> 
> I may have to get one.  The bookstore here in town is selling 'em for
> $75...  Mike?  Have we seen all the possible bytelanes from nubus cards?
> Do PDS cards work differently somehow?
I'm fairly certain that I didn't handle all the possible byte lanes
cases, I coded stuff to read bytes stored longs(32 bits) apart
(tested), shorts(16 bits) apart(tested) and chars(8 bits) apart.
I only had short and long cards to test on, and I just figured that
the char case would work.  I'm not sure if the code works if the
longs are on odd bytes as well as even bytes, I think it does.  If
they used any other byte lanes then my code would fail to notice
the card and would be unable to return any info about it.  The
folks at Apple placed a test pattern in the roms of the card coded
in the byte lanes that the card uses, I look for that test pattern,
if its not found then you bus error.  This is commented in the
code, and I think i'm correct (its been awhile since I've looked
at that code) My guess is that the pds slot isn't exactly a nubus
slot, or Allen's code is broken, or there is a stray nutrino on
that card, because my code is 100% bug free.  I don't write bugs,
I find them.  ;-)
Basically, if there isn't a bus error then "No es mi problemo, es
Allen's code" ;-)



-Mike "Tact" Finch

------------------------------------------------------------------------------