Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 13 Sep 2006 15:01:28 +0300
From:      Toomas Aas <toomas.aas@raad.tartu.ee>
To:        questions@freebsd.org
Subject:   Kernel panic with 5.5, possibly in propagate_priority
Message-ID:  <4507F318.80405@raad.tartu.ee>

next in thread | raw e-mail | index | archive | help
Hello!

Last October, after I upgraded the OS on my IBM x225 from RELENG_5_3 to
RELENG_5_4, I experienced a kernel panic:

http://makeashorterlink.com/?S167211CD

As noted in the article referenced above, I disabled debug.mpsafenet
and debug.mpsafevm. This was really just a guess. Seems like the
guess was lucky, because I never had another kernel panic after that.

Yesterday I upgraded this server from RELENG_5_4 to RELENG_5_5 and
decided to try to re-enable debug.mpsafenet and debug.mpsafevm. Today
in the middle of the day I got a kernel panic.

As also promised in the article above, I had built the debug kernel
and enabled kernel crash dumps. However, the machine just seems to freeze
when it panics and doesn't really generate a crash dump into the location
that I specified in /etc/rc.conf. So I still don't have much very useful
debug information.

The panic message is pretty similar to the one in the message above:

--------------------------------------------------------------------------
kernel trap 12 with interrupts disabled
Fatal trap 12: page fault while in kernel mode
cpuid=1; apic id = 06
fault virtual address: 0x24
fault code = supervisor read, page not present
instruction pointer = 0x8:0xc05276ae
stack pointer = 0x10:0xe83aab20
frame pointer = 0x10: 0xe83aab48
code segment = base 0x0, limit 0xfffff, type 0x1b
               = DPL 0, pres 1, def32 1, gran 1
processor eflags = resume, IOPL = 0
current process = 12064 (httpd)
trap number = 12
panic: page fault
cpuid = 1
spin lock sched lock held by 0xc279c480 for >5 seconds
--------------------------------------------------------------------------

Closest... um... thing to the instruction pointer 0xc05276ae seems to be
propagate_priority:

[heerold] ~> nm -n /boot/kernel/kernel | grep c05276
c052762c t propagate_priority

I found another thread discussing an issue which seems vaguely
similar, but I'm not enough of a FreeBSD kernel expert to be sure:

http://makeashorterlink.com/?G457131CD

After reading this thread - should I, perhaps, add NO_ADAPTIVE_MUTEXES to
my kernel config (and remove ADAPTIVE_GIANT)? What are other people's
experiences running FreeBSD 5.x on dual-processor IBM xSeries 225 box?
Am I the only one doing this?

For now, I just disabled debug.mpsafenet and debug.mpsafevm again and
I hope it works out as well as it did last time.

-- 
Toomas Aas



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4507F318.80405>