SORT BY:

LIST ORDER
THREAD
AUTHOR
SUBJECT


SEARCH

IPS HOME


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

    RE: iSCSI: padding on intermediate sequences



    Julian,
     
    Notice in the example I gave below, that the pad was only applied when the F bit was set. Section 9.7.7 says:
     

    The Data Segments of Data-in and Data-out PDUs SHOULD be filled to the

    integer number of 4 byte words (real payload) unless the F bit is set

    to 1.

     
    So my example is not forbidden. Can you please change section 9.7.7 to be more clear and forbid this behavior? I suggest something like:
     
    The Data Segments of Data-in and Data-out PDUs SHOULD be filled to the
    integer number of 4 byte words (real payload) unless the F bit is set
    to 1
    on the last Data PDU that concludes the entire requested
    data transfer for the task from the target's perspective.
     
    Eddy
    -----Original Message-----
    From: Julian Satran [mailto:Julian_Satran@il.ibm.com]
    Sent: Wednesday, March 27, 2002 1:42 AM
    To: Eddy Quicksall
    Cc: Sanjay Goyal
    Subject: RE: iSCSI: padding on intermediate sequences



    The spec clearly forbids this behavior - data pdus (except the last) have to filled (not padded) to a 4 byte boundary. That is what the text says.

    Regards,
    Julo


    Eddy Quicksall <Eddy_Quicksall@ivivity.com>

    26-03-02 18:52
    Please respond to Eddy Quicksall

           
            To:        Julian Satran/Haifa/IBM@IBMIL
            cc:        Sanjay Goyal <sanjay_goyal@ivivity.com>, Sukha Ghosh <sukha_ghosh@ivivity.com>
            Subject:        RE: iSCSI: padding on intermediate sequences

           


    Maybe I didn't say something right. Here is what I meant:
     
    I-T Read CDB of 1024 bytes
    T-I Data-In with 513 bytes and F bit set. So, pad is 3 bytes. I don't like pad here.
    T-I Data-In with 511 bytes and F bit set. So, pad is 1 byte. This is the only place I would like to see pad.
     
    Since the F bit ends a sequence and since a transfer can have several sequences, the spec allows each sequence to have pad. It is a problem for the HW guys to deal with pad except at the end of the whole transfer.
     
    Eddy
    -----Original Message-----
    From:
    Julian Satran [mailto:Julian_Satran@il.ibm.com]
    Sent:
    Tuesday, March 26, 2002 11:09 AM
    To:
    Eddy Quicksall
    Subject:
    Re: iSCSI: padding on intermediate sequences


    Eddy - it says real payload. They must be filled not padded. That is mandataed to avoid padding abuse.


    But I assume that I missread your question.


    Julo


    Eddy Quicksall <Eddy_Quicksall@ivivity.com>
    Sent by: owner-ips@ece.cmu.edu

    26-03-02 16:52
    Please respond to Eddy Quicksall

           
           To:        "ips@ece. cmu. edu (E-mail)" <ips@ece.cmu.edu>

           cc:        

           Subject:        iSCSI: padding on intermediate sequences


         



    Section 9.7.7 says:

     

    The Data Segments of Data-in and Data-out PDUs SHOULD be filled to the

    integer number of 4 byte words (real payload) unless the F bit is set

    to 1.

     

    This allows one to pad intermediate sequences of a transfer. But, there are reasons why padding on intermediate sequences within a transfer can make life difficult.

     

    Can we forbid padding on all but the last PDU for the transfer? If that is objectionable, could we forbid padding on all but the last PDU of a transfer if DataInOrder=yes.

     

    Eddy







Home

Last updated: Thu Mar 28 10:18:22 2002
9363 messages in chronological order