Hi Ulrich, > In my primitive point of view iSCSI is just "another type of cable", making me wonder: > Is iSCSI allowed to reorder the requests at all? Shouldn't the block layer or initiator do > so, or the target doing out-of order processing (tagged queueing)? iSCSI RFC does not require to serialize a commands flow. It's just an "iSCSI user" feature - to send some set of SCSI commands in an unbreakable batch to a device. But, as far as I understood, the problem, Mike described, is not a reorder but an increasing of time between full data transmission of the commands from the batch. > I mean: If there is a problem that occurs even without using iSCSI, should iSCSI try to fix it? Since that is software iSCSI specific issue it could be fixed/improved in software. How it's handled in HW offloaded implementation is unknown for me. BR, Dmitry
