Subject: kern/9263: random kernel panics due to locking problems
To: None <gnats-bugs@gnats.netbsd.org>
From: Matthias Scheler <tron@lyssa.zhadum.de>
List: netbsd-bugs
Date: 01/20/2000 14:30:49
>Number:         9263
>Category:       kern
>Synopsis:       random kernel panics due to locking problems
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    kern-bug-people (Kernel Bug People)
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Jan 20 14:30:00 2000
>Last-Modified:
>Originator:     Matthias Scheler
>Organization:
Matthias Scheler                            http://www.sighardstrasse.de/~tron/
>Release:        2000/01/20
>Environment:
System: NetBSD lyssa.zhadum.de 1.4P NetBSD 1.4P (LYSSA) #0: Thu Jan 20 14:22:24 CET 2000 tron@lyssa.zhadum.de:/src/sys/compile/LYSSA i386


>Description:
For about a week my current i386 system is panicing on random occasions.
Because it always happened when X11 was running I couldn't see DDB output
and thus wasn't able to get a crash dump until I switched DDB off.
Afterwars I got a crash dump with this trace back:

(gdb) target kcore netbsd.0.core
panic: lockmgr: processor %lu, not exclusive lock holder %lu unlocking
#0  0xf02ec620 in deflate_window_in ()
(gdb) where
#0  0xf02ec620 in deflate_window_in ()
#1  0x5768000 in ?? ()
#2  0xf026ad53 in cpu_reboot ()
#3  0xf01b4f31 in panic ()
#4  0xf01a8cda in lockmgr ()
#5  0xf01aa30c in proclist_unlock_write ()
#6  0xf01a64ff in sys_wait4 ()
#7  0xf0270f79 in syscall ()
#8  0xf0100d65 in syscall1 ()
can not access 0xefbfd120, invalid translation (invalid PDE)
can not access 0xefbfd120, invalid translation (invalid PDE)
Cannot access memory at address 0xefbfd120.

Obviously a lock which wasn't locked is unlocked somewhere.

>How-To-Repeat:
I can't exactly repeat it but it always happend when X11 (XFree86 3.3.6
compiled from our X11 sources) was running.

>Fix:
Sorry, I don't know.

>Audit-Trail:
>Unformatted: