From rdroot@purple.rahul.net Fri Mar  5 01:43:52 1999
Return-Path: <rdroot@purple.rahul.net>
Received: from purple.rahul.net (purple.rahul.net [192.160.13.88])
	by hub.freebsd.org (Postfix) with ESMTP id CB67815159
	for <FreeBSD-gnats-submit@freebsd.org>; Fri,  5 Mar 1999 01:43:50 -0800 (PST)
	(envelope-from rdroot@purple.rahul.net)
Received: (from root@localhost)
	by purple.rahul.net (8.9.2/8.9.2) id BAA08721;
	Fri, 5 Mar 1999 01:43:33 -0800 (PST)
	(envelope-from rdroot)
Message-Id: <199903050943.BAA08721@purple.rahul.net>
Date: Fri, 5 Mar 1999 01:43:33 -0800 (PST)
From: dhesi@rahul.net
Sender: rdroot@purple.rahul.net
Reply-To: dhesi@rahul.net
To: FreeBSD-gnats-submit@freebsd.org
Subject: 3.1R page fault while in kernel mode, from de driver
X-Send-Pr-Version: 3.2

>Number:         10397
>Category:       kern
>Synopsis:       3.1R page fault while in kernel mode, from de driver
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    freebsd-bugs
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Fri Mar  5 01:50:00 PST 1999
>Closed-Date:    Tue Jun 6 02:55:53 PDT 2000
>Last-Modified:  Tue Jun 06 02:57:11 PDT 2000
>Originator:     Rahul Dhesi <dhesi@rahul.net>
>Release:        FreeBSD 3.1-RELEASE i386
>Organization:
a2i network
>Environment:

	FreeBSD 3.1-RELEASE, i386, ASUS P2B-s motherboard, 
	Tulip-based 100/10 Ethernet card, SCSI disk drives

>Description:

	About once every 2-3 days there is a kernel panic due to
	'page fault while in kernel mode'.  I have generated a
	debugging kernel and collected some crash dumps, so detailed
	debugging information should be available.  Selected portions
	of a gdb session are given below.  The complete kernel
	configuration, and a more detailed gdb session log, have
	been made available at this URL:

	   http://www.rahul.net/dhesi/freebsd/crash/

	Excerpts from gdb session:


Fatal trap 12: page fault while in kernel mode
fault virtual address   = 0x48
fault code              = supervisor write, page not present
instruction pointer     = 0x8:0xf01eb2ed
stack pointer           = 0x10:0xf02c22d0
frame pointer           = 0x10:0xf02c2310
code segment            = base 0x0, limit 0xfffff, type 0x1b
                        = DPL 0, pres 1, def32 1, gran 1
processor eflags        = interrupt enabled, resume, IOPL = 0
current process         = Idle
interrupt mask          = net tty
trap number             = 12
panic: page fault

syncing disks...

...

(kgdb) where
#0  boot (howto=260) at ../../kern/kern_shutdown.c:285
#1  0xf015db30 in at_shutdown (
    function=0xf02b686f <__set_sysinit_set_sym_memdev_sys_init+1115>, arg=0x0,
    queue=12) at ../../kern/kern_shutdown.c:446
#2  0xf024c345 in trap_fatal (frame=0xf02c1fd8, eva=48)
    at ../../i386/i386/trap.c:942

...

#21 0xf015db30 in at_shutdown (
    function=0xf02b686f <__set_sysinit_set_sym_memdev_sys_init+1115>, arg=0x0,
    queue=12) at ../../kern/kern_shutdown.c:446
#22 0xf024c345 in trap_fatal (frame=0xf02c2294, eva=72)
    at ../../i386/i386/trap.c:942
#23 0xf024c023 in trap_pfault (frame=0xf02c2294, usermode=0, eva=72)
    at ../../i386/i386/trap.c:835
#24 0xf024bc9a in trap (frame={tf_es = 16, tf_ds = 16, tf_edi = -257310720,
---Type <return> to continue, or q <return> to quit---
      tf_esi = -257223552, tf_ebp = -265542896, tf_isp = -265542980,
      tf_ebx = -253346544, tf_edx = -257223552, tf_ecx = 0,
      tf_eax = -1649606648, tf_trapno = 12, tf_err = 2, tf_eip = -266423571,
      tf_cs = 8, tf_eflags = 66051, tf_esp = -257310720, tf_ss = -64551872})
    at ../../i386/i386/trap.c:437
#25 0xf01eb2ed in tulip_rx_intr (sc=0xf0a9c000) at ../../pci/if_de.c:3602
#26 0xf01eba33 in tulip_intr_handler (sc=0xf0a9c000, progress_p=0xf02c2338)
    at ../../pci/if_de.c:3976
#27 0xf01ebb91 in tulip_intr_normal (arg=0xf0a9c000) at ../../pci/if_de.c:4165

>How-To-Repeat:
>Fix:
	
	None known to me.

>Release-Note:
>Audit-Trail:

From: Alexander Langer <alex@cichlids.com>
To: freebsd-gnats-submit@freebsd.org, dhesi@rahul.net
Cc:  
Subject: Re: kern/10397: 3.1R page fault while in kernel mo
Date: Sun, 21 Nov 1999 15:39:47 +0100 (CET)

 What about this PR?
 Does the problem still exist on recent releases (both, current and
 stable) for you, Rahul?
 
 Thank you
 
 Alex
 
 

From: Rahul Dhesi <dhesi@rahul.net>
To: Alexander Langer <alex@cichlids.com>
Cc: freebsd-gnats-submit@freebsd.org
Subject: Re: kern/10397: 3.1R page fault while in kernel mo 
Date: Sun, 21 Nov 99 17:52:06 -0800

 Hi, I have not encountered this recently.  Of the two machine
 on which I observed the problem:
 
 - One has been offline.  I was planning to have the motherboard replaced,
   to see if that was what caused the problem.  Haven't done it yet.
 - The other is running 3.1-RELEASE but I am not really stressing it,
   and the problem does not reappear.  So perhaps the problem requires a
   high degree of filesystem activity.
 
 On a third similar machine, running 3.1-STABLE from a few months ago,
 I do not observe the problem even under load.
 
 Rahul
 
 > Date:  Sun, 21 Nov 99 15:39:47 +0100
 > From:  Alexander Langer <alex@cichlids.com>
 > To:    freebsd-gnats-submit@freebsd.org, dhesi@rahul.net
 > Message-Id: <199911211439.PAA15268@cichlids.cichlids.com>
 > Subject: Re: kern/10397: 3.1R page fault while in kernel mo
 
 > What about this PR?
 > Does the problem still exist on recent releases (both, current and
 > stable) for you, Rahul?
 > 
 > Thank you
 > 
 > Alex
 
State-Changed-From-To: open->closed 
State-Changed-By: alfred 
State-Changed-When: Tue Jun 6 02:55:53 PDT 2000 
State-Changed-Why:  
Submitter announced destruction of reproduceable problem, if the problem 
still exists in 4.0 we can re-open the PR. 

http://www.freebsd.org/cgi/query-pr.cgi?pr=10397 
>Unformatted:
 >Hoe-To-Repeat: 
 
 	Occurs spontaneously.  The only activity on the machine is
 	that once an hour 'rsync' is used to make a backup of a
 	filesystem from another machine onto a local disk on this
 	machine.  Other than this, and the standard nightly cron
 	jobs that come with FreebSD, the machine is idle.
 
 	I need to deploy this paticular machine in production mode
 	soon.   To aovid the kernel panics, I will likely revert
 	it to FreeBSD 2.2.8 (and add an external SCSI card supported
 	by 2.2.8, instead of relying on the on-board SCSI which
 	2.2.8 does not support).  If any of the FreeBSD developers
 	would like me to keep this machine in it current configuration
 	to collect more debugging information, please let me know.
 	I will be happy to provide you with any information you need.
 
 	Otherwise I will wait 2-3 weeks to hear from you and then
 	downgrade it to 2.2.8.
 
