From owner-freebsd-hackers  Sun Jun 11 17:55:13 1995
Return-Path: hackers-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id RAA28238
          for hackers-outgoing; Sun, 11 Jun 1995 17:55:13 -0700
Received: from gndrsh.aac.dev.com (gndrsh.aac.dev.com [198.145.92.241])
          by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id RAA28230
          ; Sun, 11 Jun 1995 17:55:09 -0700
Received: (from rgrimes@localhost) by gndrsh.aac.dev.com (8.6.11/8.6.9) id RAA01675; Sun, 11 Jun 1995 17:55:18 -0700
From: "Rodney W. Grimes" <rgrimes@gndrsh.aac.dev.com>
Message-Id: <199506120055.RAA01675@gndrsh.aac.dev.com>
Subject: Re: Problem with 2940
To: gibbs@freefall.cdrom.com (Justin T. Gibbs)
Date: Sun, 11 Jun 1995 17:55:18 -0700 (PDT)
Cc: evanc@synapse.net, hackers@freebsd.org
In-Reply-To: <199506112347.QAA24633@freefall.cdrom.com> from "Justin T. Gibbs" at Jun 11, 95 04:47:05 pm
X-Mailer: ELM [version 2.4 PL24]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 1469      
Sender: hackers-owner@freebsd.org
Precedence: bulk

> 
> >System: P75 on SiS chipset, 32 MB RAM (both 60 ns), Seagate Hawk 2 GB
> >(SCSI), Toshiba 3401B external CD (SCSI), Adaptec 2940 and a Surecom
> >(NE2000-compatible) PCI ethernet.
> >
> 
> Can you reproduce these errors with the external and internel cache
> of your pentium disabled?  Parity is enabled and reported (when incorrect)
> by the driver.  You may also want to find out if your drive has parity
> enabled.  I've been using a 4GB Hawk to test the 2940 code for some time,
> so I don't think that there is some funny interaction between the two. Your
> problem sounds like a cache coherency bug in your hardware.

Or display hardware that is adding a bit to the character, or main
memory (do you have 36 bit simms and a chip set that does memory
parity creation and checking)?  I find that it is very hard to believe
you even got far enough to get the system booted if the data was really
this way in memory or on disk, all your binaries should be corrupt.

This also tends to rule out the cache, as your binaries should be blowing
chunks all over the place (or maybe they are and you failed to mention
that).

I am really really suspecting display hardware here, can you telnet in
and see if the files look fine over a telnet connection????

Do cksums on the files match those from a good system??

-- 
Rod Grimes                                      rgrimes@gndrsh.aac.dev.com
Accurate Automation Company                   Custom computers for FreeBSD