|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: iSCSI: Checking the I bitOK, Except for the fact that Eddy pointed out to me 3.2.1.1 does call it the I bit that is always set to 1 for responses (and Async Msgs by the way), I think my basic statement is still true -- It is not a reserved bit. So the UNH statements really does not apply. And if it is a bit that is always set to one but does not need to be checked, why is it set to one? Why not set it to zero and mark it reserved, so that the UNH statement does apply, and then we can use it for something else in the future. The bit seems redundant, why do we need it to be set? Does this have something to do about with expediting the handling of responses in a target which is working on extended copy commands? . . . John L. Hufferd Senior Technical Staff Member (STSM) IBM/SSG San Jose Ca Main Office (408) 256-0403, Tie: 276-0403, eFax: (408) 904-4688 Home Office (408) 997-6136, Cell: (408) 499-9702 Internet address: hufferd@us.ibm.com Julian Satran/Haifa/IBM@IBMIL@ece.cmu.edu on 12/14/2001 09:15:29 AM Sent by: owner-ips@ece.cmu.edu To: ips@ece.cmu.edu cc: Subject: Re: iSCSI: Checking the I bit I think that Eddy's suggestion to state that parties have to check only what they have to check is valuable and we may want to include a general statement about that. Julo John Hufferd Sent by: To: owner-ips@ece.c "Eddy mu.edu Quicksall" <Eddy@Quicksall 14-12-01 09:55 .com> cc: "ips@ece. cmu. edu \(E-mail\)" <ips@ece.cmu.ed u> Subject: Re: iSCSI: Checking the I bit Eddy, Technically, the In coming PDUs all have Byte 0, Bit 6, set to one. It is not identified as the I (Immediate) bit. And it is NOT reserved. So the Statement from the UNH Plugfest does not apply. I think your point is that if all the incoming PDUs have that bit set, why do we need to set the bit, and why do we need to check it. I think this bit has evolved over time, and perhaps up to now no one has noticed. If every incoming PDU has the bit set, we may not need the bit to be set, and perhaps it should be reserved, thereby not requiring the check. . . . John L. Hufferd Senior Technical Staff Member (STSM) IBM/SSG San Jose Ca Main Office (408) 256-0403, Tie: 276-0403, eFax: (408) 904-4688 Home Office (408) 997-6136, Cell: (408) 499-9702 Internet address: hufferd@us.ibm.com "Eddy Quicksall" <Eddy@Quicksall.com>@ece.cmu.edu on 12/13/2001 03:26:18 AM Sent by: owner-ips@ece.cmu.edu To: "ips@ece. cmu. edu \(E-mail\)" <ips@ece.cmu.edu> cc: Subject: iSCSI: Checking the I bit Is it necessary for the initiator to check the I bit in every response? If an initiator does not need it, then I don't want to take the extra time to check it. I think this is consistent with the thinking of all attendees of the UNH Plug Fest because the report from UNH IOL was that "all companies failed that test". I would like to propose adding some wording to 3.2.1.1 similar to "It is not necessary to check this bit for 1 if the implementation in the initiator does not need its use". Eddy_Quicksall@iVivity.com
Home Last updated: Tue Dec 18 10:18:01 2001 8125 messages in chronological order |