On Sep 26, 2006, at 10:43 PM, Blair Sadewitz wrote:

>
> I would submit a PR for this but I have been unable to capture the  
> contents of the panic message; it has happened twice.  If it  
> happens again, I will try to have a pen and paper handy before the  
> "crash dump megabyte countdown" consumes the screen.  What's  
> happening is just what the subject says; sometimes, as I noted  
> previously, I must prod the USB subsystem/kue driver by hotplugging  
> the adapter one or two times in order to get the firmware to load  
> properly and/or the device probed.
>
> How does one capture the panic message aside from writing it down,  
> anyway?  I will look into this when I have time but a suggestion  
> from someone more experienced than I would be helpful.
>

I don't know if it makes it into the message buffer in this case, try  
dmesg -x? If it doesn't work, try:

compile the kernel with -g and boot it
sysctl -w ddb.onpanic=1
cause the panic
bt /l
reboot 0x104

After reboot you can use gdb on the core or dmesg -x to extract the  
panic messages from the core file.