Re: pci id database bug?

[Chris Turner <c.turner@xxxxxxxxxxxxxxxxxxx>]

On 10/23/12 02:40, Sascha Wildner wrote:

> Also, what's in your loader.conf. We had a similar problem recently somehow triggered
 > by the ale(4) module being in loader.conf, but it was kinda fixed. Not sure yet what
 > the real overall issue is.

For record I also had this 1x - after the same exact message -
after reboot and manual bootloader 'boot'  was OK, then I rebuilt kernel to include
the 'ale' change and it and didn't happen - but I haven't rebooted again either -
however, the previous 'problem' kernel was booted several times - so it seems sporadic.
My theory is that there might be a race/deadlock in cam or something,
since it seems to setup next, and kernel is MP by then,
but that's purely speculative based on dmesg sequence.

--8<--

$ cat /boot/loader.conf.local
#
# $Id$

# bootloader options

boot_verbose="yes"                      # enable verbose booting
loader_color="yes"                      # pretty unicorns
fred_disable="YES"                      # or not

# kernel modules

pf_load="yes"                           # pf for various nat / bridge stuff
uplcom_load="yes"                       # needed for /dev/ucom0
accf_data_load="yes"                    # network accept filtering
accf_http_load="yes"                    # HTTP network accept filtering

screensave_load="YES"                   # screensaver
screensave_name="warp_saver"            # warpspeed capn.

linux_load="yes"                        # leenukz
linprocfs_load="yes"                    # leenukz

# SysV IPC Tunables
# (pgsql, xine, jack, etc)

kern.ipc.semmni=1024
kern.ipc.semmns=4096
kern.ipc.semmnu=1024
kern.ipc.semmsl=1024
kern.ipc.semmap=4096
kern.ipc.shmall=524288
kern.ipc.shmseg=1024

# machine specific

snd_hda_load="yes"                      # dev01 audio
kate_load="yes"                         # dev01 temp sensors