RE: LUN field in R2T
Paul Koning <[email protected]> Fri, 4 Jan 2008 12:48:17 -0500
| Newsgroups | gmane.ietf.ips |
|---|---|
| Message-ID | <[email protected]> |
>>>>> "Sears" == Sears Bill <[email protected]> writes: Sears> Sounds like whatever the intent of the field was that the spec Sears> might be somewhat ambiguous. I'll err on the side of caution Sears> and make sure the LUN field is copied from the R2T into the Sears> Data-Outs. Personally in neither our initiator or target Sears> implementations do I even look at that LUN field in the R2T or Sears> Data-Out since the ITT and TTT are sufficient for tracking the Sears> IO. However I can conceed that some target implementations Sears> may benefit from having the LUN field in the Data-Out. >> Further, I'm not sure that 10.7.4 is clear on this point: >> "If the Target Transfer Tag is provided, then the LUN field >> MUST hold a valid value and be consistent with whatever was >> specified with the command;" >> One could argue that 'command' is generally used in the >> context of a SCSI command, not an R2T. To reiterate - It's >> reasonable to say that an initiator should copy the data, but >> I don't think that a target can depend on that behavior. It sure looks like the spec is contradictory. So the safe implementation approach would be: 1. In the initiator, copy LUN from R2T to Data Out, without looking at it. 2. In the target, copy LUN from the request to R2T (don't make up your own value). paul _______________________________________________ Ips mailing list [email protected] https://www1.ietf.org/mailman/listinfo/ips