Kernel panic; fatal trap 12; on task 22, USB0: was Re: Moused issues?

[Joe Altman]

NB: I've copied -usb because it looks to me like it's definitely USB
that's implicated; but I still need a clue for getting a crash dump;
-questions seems the place to ask for that.

On Sun, Sep 23, 2007 at 07:57:41PM -0400, Joe Altman wrote:
> 
> uname for the machine on which it fails:
> 
> 6.2-STABLE FreeBSD 6.2-STABLE #0: Sun Sep  2 17:30:39 EDT 2007 i386
> 
> Is it possible to get more info on this for debugging short of putting
> debugging in the kernel, and configuring a dump device; rebuilding
> world, and hoping my machine does not become unusable?

I'm still seeing the USB associated crashes, with sources updated
several times and world remade between Sept. 23 and Oct. 6, 11 AM EDT.

The above uname shows the only kernel I can use.

I've attempted to obtain a dump using these instructions:

http://www.freebsd.org/doc/en_US.ISO8859-1/books/developers-handbook/kerneldebug.html#KERNELDEBUG-OBTAIN

And here, I think, are the relevent details.

In /etc/rc.conf, I have:

##Crash
dumpdev="AUTO"
dumpdir="/var/tmp/crash"

/var/tmp/crash exists with appropriate permissions; and swapinfo shows
2 Gig of space:

 ~ $: swapinfo
Device          1M-blocks     Used    Avail Capacity
/dev/ad0s1b          2048        0     2048     0%

AIUI, dumpdev="AUTO" dictates the use of /dev/ad0s1b.

/var/tmp has this much free space:

df -m
/dev/ad0s1f      3962   618  3027    17%    /var/tmp

However, my kernel tosses this message to the console:

Fatal trap 12: page fault while in kernel mode

<snip>

and indicates a problem with task 22, USB[1] 0. Prior to problem with
task 22, USB 0; it was the same fatal trap, task 25, USB 1.

Then it indicates that there is no dump device. It seems to me that
all I should have to do is define the dumpdev in rc.conf to obtain a
dump; and in my case, since /var is smaller than RAM and swap, define
/var/tmp/crash.

So it looks to me as if I am experiencing what is described here:

"...a kernel is crashing before dumpon(8) can be executed." taken from
the kerneldebug.html page.

Or am I missing something in the configuration of the dump device? If
not, is my only option to put a dump directive into my kernel config?
If so, what is the proper syntax?

device dump  

or something else?

Also, I decided it was easier for my to snap a picture than
transcribe the screen; it's here if anyone is interested:

chthonic.com/crash-crash-crash

I'm limping along, I think; and would appreciate some clues, please.

One more data point: I'm using an IBM PIII laptop; and I do not
experience any crash on that; and my AMD dual core does not crash
either; all three machines use, AFAICT, the same USB code.

Thanks.

[1] The USB controller is: <Intel 82801EB (ICH5) USB controller and
<Intel 82801EB/R (ICH5) USB 2.0