SORT BY:

LIST ORDER
THREAD
AUTHOR
SUBJECT


SEARCH

IPS HOME


    [Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

    RE: iSCSI: FirstBurstSize and unsolicited data




    Pat,

    That piece is a leftover (together with the piece found by Ron) from the time neither FirstBurstSize nor R2T where mandatory to fullfill.
    FirstBurstSize was particularly nasty because if not done properly a target starting to send R2T in advance to compensate for latency
    might end-up with less data than expected .

    Now the only condition we have to take care is "Incorrect amount of data" and I renamed the condition to that.

    The text reads also:

    The target reports the "Incorrect amount of data" condition if dur-ing data output the total data length to output is greater than First-BurstSize, but the initiator sent an amount different than FirstBurstSize of unsolicited non-immediate data or the amount of data sent as a reply to an R2T does not match the amount requested.

    Julo


    pat_thaler@agilent.com
    Sent by: owner-ips@ece.cmu.edu

    06/15/2002 12:15 AM
    Please respond to pat_thaler

           
            To:        ips@ece.cmu.edu
            cc:        
            Subject:        RE: iSCSI: FirstBurstSize and unsolicited data

           



    RE:
    >
    > The target reports the "Not enough unsolicited data" condition only
    > if it does not support output (write) operations in which the total
    > data length is greater than FirstBurstSize, but the initiator sent
    > less than FirstBurstSize amount of unsolicited data, and out-of-order
    > R2Ts cannot be used.
    >
    > </quote>

    This text does seem strange. Why does it say "only if it does not support output (write) operations in which" the initiator sent less data than the standard requires?
    Why would the target support that given that the initiator is required to send FirstBurstSize of unsolicited data if it sends non-immediate data PDUs (and total data length is greater than FirstBurstSize). Even out-of-order R2Ts are enabled, the target shouldn't be required to send an R2T for data that should have been sent unsolicited. This will complicate implementations.

    The text also doesn't deal with cases where the Initiator sent only Immediate data.

    Also it doesn't deal with cases where the transfer is less than First Burst size and the initiator sent non-immediate unsolicited data with a length less than the required amount.

    I think the text should be

    "The target reports the "Not enough unsolicited data" condition if the Initiator sent  non-Immediate unsolicited data with a totla unsolicited data length less than the smaller of FirstBurstSize and Expected Data Transfer Length."

    It also isn't clear why this error has an iSCSI condition code while other similar errors do not. In particular, it is just as possible that an Initiator sends less data in response to an explicit R2T. When it sends less unsolicited data than it should, it is violating an implicit R2T. Why is violation of an implicit R2T different from violation of an explicit R2T? And why is sending less data than expected flagged with a Sense Data error but sending more data than expected is not?

    Regards,
    Pat




Home

Last updated: Fri Jun 21 02:18:59 2002
10912 messages in chronological order