Re: netbsd-5 locking panic

[Hauke Fath <hauke%Espresso.Rhein-Neckar.DE@localhost>]

At 21:25 Uhr +0200 26.4.2009, Hauke Fath wrote:
>>(gdb) info line *0x12345678
>
>Thanks for the hint. I'll do that tomorrow, unless I get a coredump earlier
>- the console server is located behind the router, so breaking into ddb
>does not do me any good while I am at home...

So... getting up at half past five this morning, I found the machine had
been busy going up and down during the night

Apr 26 22:01:58 Vertatscha savecore: reboot after panic: panic: LOCKDEBUG
Apr 26 22:01:58 Vertatscha savecore: (null): Bad address
Apr 26 22:05:57 Vertatscha savecore: reboot after panic: panic: LOCKDEBUG
Apr 26 22:05:57 Vertatscha savecore: (null): Bad address
Apr 26 22:11:28 Vertatscha savecore: reboot after panic: panic: LOCKDEBUG
Apr 26 22:11:28 Vertatscha savecore: (null): Bad address
Apr 26 22:16:40 Vertatscha savecore: reboot after panic: panic: LOCKDEBUG
Apr 26 22:16:40 Vertatscha savecore: (null): Bad address
Apr 27 01:04:14 Vertatscha savecore: reboot after panic: panic: LOCKDEBUG
Apr 27 01:04:14 Vertatscha savecore: (null): Bad address
Apr 27 01:44:07 Vertatscha savecore: reboot after panic: panic: LOCKDEBUG
Apr 27 01:44:07 Vertatscha savecore: (null): Bad address
Apr 27 01:47:02 Vertatscha savecore: reboot after panic: panic: LOCKDEBUG
Apr 27 01:47:02 Vertatscha savecore: (null): Bad address
Apr 27 01:50:19 Vertatscha savecore: reboot after panic: panic: LOCKDEBUG
Apr 27 01:50:19 Vertatscha savecore: (null): Bad address
Apr 27 03:28:57 Vertatscha savecore: reboot after panic: panic: LOCKDEBUG
Apr 27 03:28:57 Vertatscha savecore: (null): Bad address
Apr 27 03:31:53 Vertatscha savecore: reboot after panic: panic: LOCKDEBUG
Apr 27 03:31:53 Vertatscha savecore: (null): Bad address

(never a usable core, though), and I figured that I'd have to roll back the
installation. I set ddb.onpanic=1, left for work, and fully expected to
find the machine hanging at the debugger prompt. Except - it didn't. I
haven't seen a panic since.

During the upgrade, I had activated hyperthreading, and I rebooted the
machine during lunch time to switch it off, figuring that besides being
pointless with pf, even simulated SMP probably wasn't helpful in my
situation.

I set ddb.onpanic=0 for the night, to avoid early phone calls, and sort of
restore the original configuration (minus hyperthreading). Right now I am
pretty unhappy, annoyed and confused. If there's one thing worse than bugs,
it's bugs that disappear when you look at them.

Thanks for listening,

        hauke



--
"It's never straight up and down"     (DEVO)