Jockey's log confirms that the program simply triggers the problem when unbinding and then rebinding the driver (hence the driver_bind() call in the backtrace), which in turn calls driver_probe_device(), etc. up to local_pci_probe() which is what fails.
To sum it up, it's not a bug in Jockey. The broadcom driver is the problem here.
Jockey's log confirms that the program simply triggers the problem when unbinding and then rebinding the driver (hence the driver_bind() call in the backtrace), which in turn calls driver_ probe_device( ), etc. up to local_pci_probe() which is what fails.
To sum it up, it's not a bug in Jockey. The broadcom driver is the problem here.