SORT BY:

LIST ORDER
THREAD
AUTHOR
SUBJECT


SEARCH

IPS HOME


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

    Re: Question about unsolicited data in 12-95




    thanks - it would help to outline the changes you suggest especially when (to a tired reader) they look tiny.

    As for you question - your second answer is the correct one - Julo


    Bill Studenmund <wrstuden@wasabisystems.com>
    Sent by: owner-ips@ece.cmu.edu

    05/31/2002 06:37 AM
    Please respond to Bill Studenmund

           
            To:        <ips@ece.cmu.edu>
            cc:        
            Subject:        Question about unsolicited data in 12-95

           


    I have a question about unsolicited data, and if I understand it
    correctly, I have some suggestions for text corrections.

    My concusion stems around if you can send both immediate data and an
    unsolicited burst of PDUs (assuming negotiations aprove both).

    I think the answer is yes, as in 11.12 (talking about ImmediateData) there
    is:

        If ImmediateData is set to Yes and InitialR2T is set to No, then the
        initiator MAY send unsolicited immediate data and/or one unsolicited
        burst of Data-OUT PDUs.

    The and/or indicates to me that in that case the initiator can send
    immediate data, an unsolicited burst, or do both.

    If the above is not correct, please disregard the rest of the note.

    I had a question, when doing both immediate and an unsolicited burst of
    Data-OUT PDUs,  about if FirstBurstSize included the immediate data or
    not. The text in section 11.10, talking about the InitialR2T key,
    indicates that FirstBurstSize does include the immediate data.

    If all the above is correct, I'd like to suggest the following changes:

    Paragraph starting at top of page 41:

        rate iSCSI data PDUs. An initiator may send unsolicited data as imme-
        diate up to the negotiated maximum PDU size or in a separate PDU
        sequence (up to the FirstBurstSize). All subsequent data MUST be
        solicited. The maximum size of an individual data PDU or the immedi-
        ate-part of the first unsolicited burst MAY be negotiated at login.

    I suggest we change the first full sentance to:

        rate iSCSI data PDUs. An initiator may send unsolicited data as imme-
        diate up to the negotiated maximum PDU size, in a separate PDU
        sequence (up to the FirstBurstSize), or both. All subsequent data
        MUST be solicited. ...

    Third full paragraph of page 41:

        An iSCSI initiator MAY choose to send no unsolicited data, only imme-
        diate data or FirstBurstSize bytes of unsolicited data with a com-
        mand. If any non-immediate unsolicited data are sent, the total
        unsolicited data MUST be either the negotiated amount or all the data
        if the total amount is less than the negotiated amount for unsolic-
        ited data.

    I suggest something like:

        An iSCSI initiator MAY choose to send no unsolicited data, only imme-
        diate data, a burst of unsolicited Data-OUT commands, or both.
        If any non-immediate unsolicited data are sent, the total
        unsolicited data MUST be either the negotiated amount or all the data
        if the total amount is less than the negotiated amount for unsolic-
        ited data.

    I also have a question about a paragraph on p136, in section 9.3.4:

        If the Expected Data Transfer Length is higher than the FirstBurst-
        Size (the negotiated maximum amount of unsolicited data the target
        will accept), the initiator MUST send the maximum size of unsolic-
        ited data OR ONLY the immediate data.

    Does that mean that if the Expected Data Transfer Length > FirstBurstSize,
    I can't opt to send no unsolicited data? Or does it mean I can: send none,
    send 1 immediate PDU, or send FirstBurstSize (optionally with some of that
    as immediate data)?

    Take care,

    Bill





Home

Last updated: Fri May 31 14:18:37 2002
10434 messages in chronological order