From nobody@FreeBSD.ORG  Thu Jun  1 09:38:28 2000
Return-Path: <nobody@FreeBSD.ORG>
Received: by hub.freebsd.org (Postfix, from userid 32767)
	id F219B37BF16; Thu,  1 Jun 2000 09:38:26 -0700 (PDT)
Message-Id: <20000601163826.F219B37BF16@hub.freebsd.org>
Date: Thu,  1 Jun 2000 09:38:26 -0700 (PDT)
From: sussman@collab.net
Sender: nobody@FreeBSD.ORG
To: freebsd-gnats-submit@FreeBSD.org
Subject: infinite "microuptime() went backwards" messages -- caused by intense disk access
X-Send-Pr-Version: www-1.0

>Number:         18949
>Category:       kern
>Synopsis:       infinite "microuptime() went backwards" messages -- caused by intense disk access
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    freebsd-bugs
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Thu Jun 01 09:40:01 PDT 2000
>Closed-Date:    Mon Jun 12 01:59:33 PDT 2000
>Last-Modified:  Mon Jun 12 02:00:21 PDT 2000
>Originator:     Ben Collins-Sussman
>Release:        4.0
>Organization:
Collab.Net
>Environment:
FreeBSD newton.collab.net 4.0-RELEASE FreeBSD 4.0-RELEASE #0: Thu May  4 14:49:56 CDT 2000     root@newton.collab.net:/usr/src/sys/compile/NEWTON-KERNEL  i386
>Description:
This problem is identical, I think, to kern/18858, and possibly kern/18798.

The "microuptime() went backward" messages are so numerous, that my
system becomes destabilized;  the CPU load stays quite high,
everything slows down, I have to reboot.  There's no way to stop the
messages once they get started.
>How-To-Repeat:
Intense UDMA disk access on this particular motherboard.

I can *absolutely* reproduce it by doing
any sort of prolonged, intense disk access -- for example, building
the "ispell" port.  (ispell builds some huge indices.)

>Fix:
?

>Release-Note:
>Audit-Trail:
State-Changed-From-To: open->feedback 
State-Changed-By: phk 
State-Changed-When: Sat Jun 10 12:16:09 PDT 2000 
State-Changed-Why:  
Please try with a kernel without APM support 
Please include dmesg and sysctl kern.timecounter output. 

http://www.freebsd.org/cgi/query-pr.cgi?pr=18949 
State-Changed-From-To: feedback->closed 
State-Changed-By: alex 
State-Changed-When: Mon Jun 12 01:59:33 PDT 2000 
State-Changed-Why:  
The originator said, that the problem disappeared with a non-APM kernel. 

http://www.freebsd.org/cgi/query-pr.cgi?pr=18949 
>Unformatted:
