From owner-freebsd-hackers  Sat Jan  4 01:25:56 1997
Return-Path: <owner-hackers>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.4/8.8.4) id BAA08061
          for hackers-outgoing; Sat, 4 Jan 1997 01:25:56 -0800 (PST)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.8.4/8.8.4) with ESMTP id BAA08056;
          Sat, 4 Jan 1997 01:25:51 -0800 (PST)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.8.3/8.6.9) id UAA30161; Sat, 4 Jan 1997 20:21:29 +1100
Date: Sat, 4 Jan 1997 20:21:29 +1100
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199701040921.UAA30161@godzilla.zeta.org.au>
To: avalon@coombs.anu.edu.au, se@FreeBSD.ORG
Subject: Re: bootloader & memory test...
Cc: freebsd-hackers@FreeBSD.ORG, hans@brandinnovators.com
Sender: owner-hackers@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

>Besides the point, that it would be very complex to do the right thing
>if this code ever finds a problem, such a test would guarantee, that 
>the caches are "empty" every time you leave the idle loop. ;)

Actually, the test would want to work on non-cached pages to be sure
of testing main memory.  You could even run the code from a cached
pages to avoid losing I-cache, but I suppose a really good memory test
program should lock itself into the L1 cache to avoid side effects.
It would also disable interrupts ...  Not good for an idle loop.

Bruce