usb/168743: panic: double fault caused by umass

[FreeBSD at nagilum.org]

I'm compiling a new kernel with those options as I type. Yes, this is  
very much reproduceable. If you look at the dmesg you will also see  
the previous reboot for the same reason but with a different device -  
a WLAN stick that also includes some mass storage for the windows  
driver and probably another for the mini-sd or whatever card fits in  
there..
The Pendrive I then used is a 64MB (yes, megabytes) so from that you  
can probably guess how ancient that thing is so it should not have any  
new and fancy options aka quirks ;).
What I haven't tried yet is simply loading the module while no device  
is attached.
I'll try that when I have the new kernel booted. It will take some  
time though since this is a 500MHz Geode..
I think I also saw a few other bugreports related to this issue maybe  
someone else already included a backtrace.
Thanks for looking into this,
Alex.


Zitat von Hans Petter Selasky <hselasky at c2i.net>:

> Hi,
>
> This sounds like a missing MODULE_DEPEND() issue or your pendrive is  
> returning
> an error code to some of the SCSI commands which is not handled properly.
>
>
> Could you add these options to the kernel config file:
>
> options         KDB                     # Kernel debugger related code
> options         KDB_TRACE               # Print a stack trace for a panic
>
> Then get full backtrace.
>
> Is this issue reproducable. I cannot reproduce over here using a 9-stable
> GENERIC kernel without "device umass".
>
> --HPS