On Sat, Sep 29, 2007 at 10:00:36PM +0200, Nicolas Joly wrote:
> On Sat, Sep 29, 2007 at 03:09:49PM -0400, Blair Sadewitz wrote:
> > When are you getting the hard lockups?  I recently noticed some kind
> > of [what I believe to be] resource starvation (which I haven't seen
> > before while using FFS) during filesystem access where the HDD LED
> > stays on indefintely.
> 
> During kernel boot ... Due to instabilities problems, i removed all
> local sources changes and custom kernel config. Yesterday, i started
> to re-add all my devices. But with spdmem(4), my machine rocks solid
> when starting sshd (???) I then need to unplug the power cable to get
> it back, even reset or power button are useless.
> 
> For now, i reproduce it the following config:
> 
> include	 "arch/amd64/conf/GENERIC"
> options	 MULTIPROCESSOR
> options	 DIAGNOSTIC
> options	 LOCKDEBUG
> iic*	 at amdpm?
> spdmem*	 at iic? addr 0x50
> [...]
> spdmem*  at iic? addr 0x57

I just updated my sources again, and got hit by another one ... This
time, without spdmem(4) lines in kernel config.

[...]
Updating motd.
Starting ntpd.
Starting xdm.
Starting sshd.
fatal double fault in supervisor mode
trap type 13 code 0 rip ffffffff80102ba8 cs 8 rflags 13082 cr2  ffff800048f76fc8 cpl 5 r
sp ffff800048f76fd0
panic: trap
syncing disks... Kernel lock error: _kernel_lock: spinout

lock address : 0xffffffff80af9270 type     :               spin
shared holds :                  0 exclusive:                  0
shares wanted:                  0 exclusive:                  0
current cpu  :                  0 last held:                  1
current lwp  : 0xffff800049001b00 last held: 000000000000000000
last locked  : 0xffffffff804a3392 unlocked : 0xffffffff803eb0e3
curcpu holds :                  0 wanted by: 0xffff800049001b00

panic: LOCKDEBUG

Unfortunately, at that time i didn't have ddb enabled :-(

-- 
Nicolas Joly

Biological Software and Databanks.
Institut Pasteur, Paris.