From owner-freebsd-stable  Fri Feb 28 13:23:18 1997
Return-Path: <owner-stable>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.5/8.8.5) id NAA18633
          for stable-outgoing; Fri, 28 Feb 1997 13:23:18 -0800 (PST)
Received: from mh1.cts.com (root@mh1.cts.com [205.163.24.66])
          by freefall.freebsd.org (8.8.5/8.8.5) with ESMTP id NAA18627
          for <stable@freebsd.org>; Fri, 28 Feb 1997 13:23:16 -0800 (PST)
Received: from io.cts.com (io.cts.com [198.68.174.34]) by mh1.cts.com (8.8.5/8.8.5) with ESMTP id NAA29974; Fri, 28 Feb 1997 13:23:10 -0800 (PST)
Received: (from mdavis@localhost) by io.cts.com (8.8.5/8.6.9) id NAA24625; Fri, 28 Feb 1997 13:15:23 -0800 (PST)
From: Morgan Davis <mdavis@io.cts.com>
Message-Id: <199702282115.NAA24625@io.cts.com>
Subject: Re: Random dives out of make world
To: jkh@time.cdrom.com (Jordan K. Hubbard)
Date: Fri, 28 Feb 1997 13:15:23 -0800 (PST)
Cc: stable@freebsd.org
In-Reply-To: <12289.856654034@time.cdrom.com> from "Jordan K. Hubbard" at "Feb 22, 97 03:27:14 pm"
X-Mailer: ELM [version 2.4ME+ PL19 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-stable@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

Jordan K. Hubbard writes:
> > I've had the same problems with a fresh 2.1.7 sup....no one knows
> >why it's ha ppening....now I'm sure it's not my computer if others are
> >having the same prob lems.  Could the 2.1.7 source be corrupted?
> 
> If it were, I think I'd have had more failure reports than just 2. :-)
> I've been compiling and running from the 2.1.7 sources for weeks now.
> This is quite mysterious.
> 
> 					Jordan

There is nothing wrong with the integrity of the source files, because
I've been building 2.1.7-stable on a 486/66 with no problems.  But the
same sources on a P120 at work have never been able to build without
some bizarre, random error situation in cc1 ("fatal signal 11",
"address not computable", "internal compiler bug", and similar obscure
errors).

Per the advice given from participants in this list, I've done the
following over the span of several days:

	* Set BIOS to its defaults where possible
	* Disabled external CPU cache
	* Disabled all BIOS (and video) caching
	* Increased RAM wait states and I/O refresh delay timing
	* Disabled all APM settings
	* Verified SCSI chain termination

Basically, I selectively adjusted these one at a time, rebooted, and
tried a "make world".  If it failed, I put the setting back to a
default (or conservative) state, and then tried the next.  But none of
these worked.

Before totally giving up and RMA'ing the motherboard, there was one
last BIOS item I decided to change: Internal CPU cache.  I disabled it
and started the make world process.  This has effectively turned the
computer into a 286, as "make world" has been running for almost 12
hours now and it's only in "libc".  But, it hasn't crashed!  So, maybe
it's just a bad CPU.  I've got a replacement to try on Monday.

But, I was wondering if by disabling the internal CPU cache that it
could be masking some other real culprit.  Any comments?

--Morgan