From mboxrd@z Thu Jan  1 00:00:00 1970
Message-ID: <3BBB2F5A.B7F211CE@princeton.edu>
From: Martin Harriss <martin@Princeton.EDU>
MIME-Version: 1.0
To: 9fans@cse.psu.edu
Subject: Re: [9fans] Plan 9 installation
References: <20011003053303.Q28720@cackle.proxima.alt.za>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Date: Wed,  3 Oct 2001 11:31:38 -0400
Topicbox-Message-UUID: fc9c6084-eac9-11e9-9e20-41e7f4b1d025

A while back, just for giggles, I threw the install diskette into a
GXi.  Similar VGA and ethernet to what you have.  I got the VGA to come
up simply by inserting the appropriate string into vgalib; it was some
S3 variant.  Since this was only a basic sanity check to see if the
machine would run Plan9, I never did the full install - thus I didn't
get to check out the ethernet.  However, as that machine is no longer
serving its original purpose I will soon be trying to do the full
install.

As far as the ethernet is concerned, have you tried turning off the plug
and play?  (To do that you need the configuration program which you can
download from 3Com.)

Martin


Lucio De Re wrote:
> 
> The problem host is a DELL Optiplex Gs+ 5120L.
> 
> The problem hardware bits are the 3Com ethernet adapter and the
> Trio VGA adapter.
> 
> The S3 Trio 64V+ (N1C3CF - if that means anything, the 86C765 may
> well) doesn't have a matching entry in vgadb, bit I found something
> else that resembled it.  The results were disappointing, the screen
> mapping seemed totally wrong, to the extent that I cannot really
> describe it.  What I do remember, though, is the band of random-looking
> bits across the middle of the screen.  I only tried 800x600x8 with
> any success.
> 
> Even more intriguing is the Ethernet card.  The chip is marked
> Lucent, the Parallel Tasking version, and on it I found 9640S and,
> in case it again means anything, 40-01304.
> 
> The only noticeable effect of having ether0=elnk3 in plan9.ini has
> been to freeze the display at the point where 9load would scan the
> ethernet adapter (that's what it looked like to me).  Strangely,
> but maybe not so strangely, I have seven 3C509B-TPOs here and in
> trying to determine why the NetBSD netboot complained about not
> being able to reset the eeprom, I found that plan9 (or more likely
> 9load) behaved similarly on one of these adapters as it did on the
> Dell.
> 
> It sounds to me like there is a bete noir amongst the 3Com adapters
> and any suggestions on resolving this problem (from sources closer
> to the designers, I imagine) will be very welcome.  The 3C509B-TPOs
> have a Lucent chip very similar to the Dell's, except that there
> is no device-like number:
> 
>         3COM
>     40-0411-001
>  9804KGA25
> 
> more or less in the above fashion is all I can see besides the
> Parallel Tasking logo (I see no "LUCENT" in this here case, amazing
> how one makes invalid assumptions).
> 
> ++L