On Tue, 2010-04-27 at 22:50 -0700, Greg KH wrote: Hi Greg, > As for the whole idea of the extra device (which it doesn't look like > you ever initialize anywhere), it's not good to have one sitting in the > middle of the device chain that isn't owned by a bus somehow. That just > looks wierd and can cause problems with udev rules. > > why is this really needed at all? Can't you just export the port number > in the device as an attribute instead? You're right, please disregard this. We got fooled by the incrementing scsi host id in the device path. The misleading part (for me) is that even if a device sysfs devpath *seems* to represent its topology, some parts of the path are not "persistent" (across simple rmmod/modprobe) I just read udev path_id source code and saw the trick for the scsi host number (which would have been broken with this patch). Since the ata code always creates scsi hosts in ata port order, we can assume that lowest scsi host number is the first ata port, so the patch is not needed. -- Maxime -- To unsubscribe from this list: send the line "unsubscribe linux-ide" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html