One of the things I discovered while debugging the code is that there's a race condition setting the per connection DDB phy mask for a wide port. If the wide port forms after discover, then the DDB phy mask is set to a single port and always remains so. If the wide port forms before discovery, then the DDB phy mask is correctly set to all phys in the port. When the wide port forms after discovery, what happens is that the DDB sequencer code adds the phy in asd_update_port_links. However, the initial per connection DDBs never get updated ... is this correct? (I can't really tell, because the DDB code is deep magic within the driver). The driver doesn't even have a port_deform callback, so it never acts on phys being removed from the port, so I think someone needs to add phy removal from a port functionality to the driver (unless it's lurking somewhere I didn't find). Thanks, James - : send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
