Subject: Re: BIOS flash fixes all uvm_pagefaults! (Was: Re: uvm_pagefaults in
To: Don Phillips <don@resun.com>
From: Peter. Bex <Peter.Bex@student.kun.nl>
List: tech-kern
Date: 10/17/2002 20:13:17
Quoting Don Phillips <don@resun.com>:
> >>>>> "Peter" == Peter Bex <Peter.Bex@student.kun.nl> writes:
> 
> Which motherboard do you have?  I'm running an MSI MS-6378 with a
> BIOS from June 2002.  Unfortunately for me, this looks to be the
> most current revision of the BIOS for this board (v3.1).

I have an AOpen AK73Pro(A) motherboard. The BIOS is an Award thingy.

> In any event, I plugged in a SCSI card and disabled the onboard IDE
> channels.  I ran into the same problem.  Wouldn't this have avoided the VIA
chip set issues?

It should, I think...
Maybe the kernel still detects the IDE. Perhaps you could try to remove the
drivers from the kernel?
Perhaps someone with more knowledge in this area could tell me if it matters or not.

> I changed the memory timing to 3 T, and set the speed to 100MZ
> (using PC133 SDRAM).  Same problem.
> Peter> I think this is enough proof that flashing the BIOS does fix
> Peter> the VIA chipset bug Frank van der Linden mentioned. Thanks
> Peter> Frank!
> 
> At least for your motherboard.  For now, it looks like I'm SOL.
> 
> I'm going to rebuild a system based upon the IN530 and a K6-2
> processor and see if I still reproduce the problem.  It was because
> of problems with 1.5.2 on that configuration that I ended up in the
> current configuration.  IIRC, I didn't try 1.6 on the IN530.  Since
> the IN530 is SIS based, this should eliminate the VIA chip problem.

Well, good luck!

> Do we need to start a list of motherboards and their
> working/notworking status?  1/2 :-)

That's actually not such a bad idea!
Perhaps a database with known bugs in devices and solutions or work-arounds.
Perhaps even a list of special patches which shouldn't be incorporated in the
main source tree?
Then people wouldn't crowd these mailing lists if a problem is known.
I did try a search in the NetBSD search engine, but that didn't turn up anything
useful. Especially since I had no idea what the problem was.
If such a database is built up in a nice way, users could search by either
problem or device.
At least people would know wether to complain about things or not.
 
> Regards,
> -- 
>   Don Phillips         don@resun.com
>   Escondido, Calif.    My opinions are just that, and no more.

Best of luck with your system!

Peter Bex

"The process of preparing programs for a digital computer
is especially attractive, not only because it can be economically
and scientifically rewarding, but also because it can be an aesthetic
experience much like composing poetry or music."
        --Donald Knuth