Comment # 3 on bug 1179632 from Jeff Mahoney

I don���������t know if it���������s a regression.  I don���������t use these devices often and I
started using them this week to migrate an array to larger disks so they���������re
seeing heavy I/O activity. I don���������t expect the SCSI errors are necessarily part
of it.  I think it���������s the usb device disconnect in combination with the
usb-level errors that are causing a use-after-free.

This is also not the most recent kernel.  I���������ve installed the latest update,
though nothing stood out in the changelog that looked like a fix.  I���������ve also
adjusted the memory reserved for kdump to (hopefully) capture a dump next time.