Subject: port-amiga/4867: panic: MMU fault in tsleep()
To: None <gnats-bugs@gnats.netbsd.org>
From: None <frueauf@ira.uka.de>
List: netbsd-bugs
Date: 01/22/1998 11:22:06
>Number:         4867
>Category:       port-amiga
>Synopsis:       panic: MMU fault in tsleep()
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    gnats-admin (GNATS administrator)
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Jan 22 02:35:01 1998
>Last-Modified:
>Originator:     Thorsten Frueauf
>Organization:
private
	
>Release:        <NetBSD-current source date> NetBSD 1.3
>Environment:
	
System: NetBSD NetBSD 1.3 NetBSD 1.3 (MILANO-DEBUG) #0: Thu Jan 15 21:46:33 MET 1998 frueauf@NetBSD:/home/i41am1/NetBSD-1.3-src/sys/arch/amiga/compile/MILANO-DEBUG amiga

Amiga 3000 (m68030 CPU/MMU m68882 FPU)
real  mem = 16777216 (2048 pages)
avail mem = 14385152 (1756 pages)
using 115 buffers containing 942080 bytes of memory
memory segment 0 at 7000000 size 1000000
memory segment 1 at 0 size 200000
mainbus0 (root)
clock0 at mainbus0: CIA B system hz 100 hardware hz 709379
Calibrating delay loop... 
diff 1027 us, new divisor 334/1024 us
diff 1023 us, new divisor 334/1024 us
diff 1024 us, new divisor 334/1024 us
a34kbbc0 at mainbus0
ser0 at mainbus0: input fifo 512 output fifo 32
par0 at mainbus0
kbd0 at mainbus0: CIA A type Amiga
ms0 at mainbus0
ms1 at mainbus0
grfcc0 at mainbus0
grf0 at grfcc0: width 640 height 400 colors 4
ite0 at grf0: rows 50 cols 79 repeat at (30/100)s next at (10/100)s has keyboard
fdc0 at mainbus0: dmabuf pa 0x1e4bf8: dmabuf ka 0x2de0bf8
fd0 at fdc0 unit 0: 3.5dd 80 cyl, 2 head, 11 sec [9 sec], 512 bytes/sec
ahsc0 at mainbus0
scsibus0 at ahsc0: 8 targets
ahsc0: target 0 now synchronous, period=208ns, offset=12.
sd0 at scsibus0 targ 0 lun 0: <FUJITSU, M2622F-512, 0405> SCSI1 0/direct fixed
sd0: 314MB, 1429 cyl, 7 head, 64 sec, 512 bytes/sect x 644868 sectors
ahsc0: target 1 now synchronous, period=208ns, offset=12.
sd1 at scsibus0 targ 1 lun 0: <IBM, DORS-32160, S82C> SCSI2 0/direct fixed
sd1: 2063MB, 6703 cyl, 5 head, 126 sec, 512 bytes/sect x 4226725 sectors
aucc0 at mainbus0
audio0 at aucc0
zbus0 at mainbus0
le0 at zbus0 pa 0xe90000 man/pro 2167/201: address 00:60:30:00:06:e5
le0: 16 receive buffers, 4 transmit buffers
10 views configured


Kernel config file: (remarks removed)
--- snip ---
include "arch/amiga/conf/std.amiga"
maxusers	16
options 	RTC_OFFSET=0
options 	M68030		# support for 030
options 	INET		# IP networking support (Required)
file-system 	FFS		# Berkeley fast file system
file-system 	MFS		# Memory based filesystem
file-system 	NFS		# Network File System client side code
file-system 	ADOSFS		# AmigaDOS file system
file-system 	CD9660		# ISO 9660 + Rock Ridge filesystem
file-system 	MSDOSFS		# MS-DOS filesystem
file-system 	FDESC		# user file descriptor filesystem
file-system 	KERNFS		# kernel data-structure filesystem
file-system 	NULLFS		# loopback filesystem
file-system 	PROCFS		# /proc filesystem
file-system 	UMAPFS		# NULLFS + uid and gid remapping
file-system 	UNION		# union filesystem
options 	NFSSERVER	# Network File System server side code
options 	FIFO		# FIFOs; RECOMMENDED
options 	QUOTA		# User and group quotas in FFS
options 	TCP_COMPAT_42	# compatibility with 4.2BSD TCP/IP
options 	COMPAT_43	# compatibility with 4.3BSD interfaces
options 	COMPAT_09	# compatibility with NetBSD 0.9
options 	COMPAT_10	# compatibility with NetBSD 1.0
options 	COMPAT_12	# compatibility with NetBSD 1.2
options 	COMPAT_SUNOS	# Support to run Sun (m68k) executables
options 	COMPAT_NOMID	# allow nonvalid machine id executables
options 	SYSVSHM		# System V-like shared memory
options 	SYSVMSG		# System V-like messages
options 	SYSVSEM		# System V-like semaphores
options 	KTRACE		# system call tracing support
options 	NKMEMCLUSTERS=256	# Size of kernel malloc area
options 	UCONSOLE	# anyone can redirect a virtual console
options 	SCSIVERBOSE	# Verbose SCSI errors
options		NTP		# NTP phase/frequency locked loop
options 	DIAGNOSTIC	# Extra kernel sanity checks
options 	DEBUG		# Enable misc. kernel debugging code
makeoptions DEBUG="-g"
options 	GRF_ECS		# Enhanced Chip Set
options 	GRF_NTSC	# NTSC
options 	GRF_PAL		# PAL
grfcc0		at mainbus0		# custom chips
grf0		at grfcc0
ite0		at grf0			# terminal emulators for grf's
a34kbbc0	at mainbus0		# A3000/A4000 battery backed clock
aucc*		at mainbus0		# Amiga CC audio
audio*		at aucc?
le*		at zbus0		# A2065, Ameristar, Ariadne
ahsc0		at mainbus0		# A3000 scsi
scsibus*	at ahsc0
sd*	at scsibus? target ? lun ?	# scsi disks
st*	at scsibus? target ? lun ?	# scsi tapes
cd*	at scsibus? target ? lun ?	# scsi cd's
ss*	at scsibus? target ? lun ?	# scsi scanner
ch*	at scsibus? target ? lun ?	# scsi autochangers
uk*	at scsibus? target ? lun ?	# scsi unknown
pseudo-device	loop			# loopback network interface
pseudo-device	bpfilter	16	# Berkeley packet filter
pseudo-device	view		10	# views
pseudo-device	pty		16	# pseudo ptys
config	netbsd root on ? type ?
--- snip ---

>Description:
	
This machine is running as WWW Server (www.de.netbsd.org - apache 1.2.5),
sup server (supfilesrv from 1.3 bin release) and offers common user
services.

I saw the following panic, which created a crashdump (I can make it available
if needed):

-rw-r--r--   1 root  wheel  18874900 Jan 19 13:51 netbsd.1.core

(gdb) target kcore netbsd.1.core
panic: MMU fault
#0  tsleep (ident=0x0, priority=-2147483648, wmesg=0x0, timo=13)
    at ../../../../kern/kern_synch.c:365
365             curpriority = p->p_usrpri;
(gdb) bt
#0  tsleep (ident=0x0, priority=-2147483648, wmesg=0x0, timo=13)
    at ../../../../kern/kern_synch.c:365
#1  0x16cae in tsleep (ident=0xdcc28, priority=4, wmesg=0x8f78e "scheduler", 
    timo=0) at ../../../../kern/kern_synch.c:359
#2  0x8f836 in scheduler () at ../../../../vm/vm_glue.c:361
#3  0xdaf4 in main (framep=0x111fb4) at ../../../../kern/init_main.c:401

Unfortunatly, this happened not just once. It is reproducable :-(

I _think_ that the supfilesrv process is causing that, but its totaly
unclear to me why. The category is choosen port-amiga, as such problems
obviously don't appear at e.g. sup.netbsd.org, although I am not too
sure if it is amiga or m68k only.

>How-To-Repeat:
	
It happens while a certain user trys to sup from this machine, the
last messages that makes it into /var/log/messages is:

Jan 19 13:41:51 NetBSD supfile[8466]: Error sending file list to client
Jan 19 13:41:52 NetBSD supfile[8466]: GOAWAY: Error sending file list to client

The strange thing is - suping from localhost or from a machine at the same
ethernet does _not_ cause this. I also tryed this same supfile from that
certain user - I was not able to reproduce it. But every time he starts
sup on his machine I get the crash.

>Fix:
	
I have no idea. But I am willing to try out every suggestion to nail it down.
If you need more information, feel free to ask.
>Audit-Trail:
>Unformatted:
panic: MMU fault in tsleep()