From owner-freebsd-questions  Fri Nov 21 04:31:43 1997
Return-Path: <owner-freebsd-questions>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.7/8.8.7) id EAA04355
          for questions-outgoing; Fri, 21 Nov 1997 04:31:43 -0800 (PST)
          (envelope-from owner-freebsd-questions)
Received: from gwdu60.gwdg.de (gwdu60.gwdg.de [134.76.10.60])
          by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id EAA04348
          for <freebsd-questions@FreeBSD.ORG>; Fri, 21 Nov 1997 04:31:39 -0800 (PST)
          (envelope-from kheuer@gwdu60.gwdg.de)
Received: (from kheuer@localhost)
	by gwdu60.gwdg.de (8.8.5/8.8.5) id NAA14339;
	Fri, 21 Nov 1997 13:31:32 +0100 (CET)
Date: Fri, 21 Nov 1997 13:31:32 +0100 (CET)
From: Konrad Heuer <kheuer@gwdu60.gwdg.de>
Message-Id: <199711211231.NAA14339@gwdu60.gwdg.de>
To: freebsd-questions@FreeBSD.ORG
Subject: Problems with 2.2.2-RELEASE and more than 64 MB of main memory?
Content-Type: text/plain
Mime-Version: 1.0
X-Mailer: NCSA Mosaic 2.7b5 on Unknown Platform
X-URL: http://www.freebsd.org/mailto.html
Sender: owner-freebsd-questions@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

I run 2.2.2-RELEASE on a 128 MB / AMD K6-233 machine. To make use of
more than 64 MB memory, I added

options        MAXMEM=131072           #Memory size

to the kernel configuration file before creating the custom kernel.

Now I've serious problems concerning the stability of the system.
If there's load on the system, it crashes from time to time.
I increased the memory access time from 60 ns to 70 ns in the BIOS
setup - no change.

So I started the system in single user mode many times and ran a
simple test program included here (after setting the datasize process
limit to unlimited):

#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>

#define MEGABYTE  1048576
#define MEMBLOCK  1 * MEGABYTE

int main(int argc, char *argv[])
{
  char *buffer;
  long i;
  long memsize = MEMBLOCK;

  while (buffer = malloc(memsize)) {
    memset(buffer, 0, memsize);
    for (i = 0; i < memsize; ++i)
      if (*(buffer + i))
        fprintf(stderr, "%s: read error at: %x (%d)\n", argv[0], i, i);
    printf("%s: %4d MB\n", argv[0], memsize / MEGABYTE);
    free(buffer);
    memsize += MEMBLOCK;
  }
}

In the range above about 75 MB memory allocated and used, the kernel
panics. Error messages are like this:

vm_page_free: pindex(16415), busy(0), PG_BUSY(0), hold(0)
panic: vm_page_free: freeing busy page

As soon as I omit the `options MAXMEM' line in the kernel configuration
file and thus only use 64 MB of main memory, the system works very fine.

So - I'm not sure: Is possibly one of my four SIMM modules (each 32 MB)
physically defect, or is there a problem with the kernel and more than
64 MB?

Thanks for any answer
Konrad Heuer (kheuer@gwdg.de)