SORT BY:

LIST ORDER
THREAD
AUTHOR
SUBJECT


SEARCH

IPS HOME


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

    Re: iscsi : Fragmentation & Reassembly issues in iSCSI.



    Matt Wakeley wrote:
    
    > Ok,
    >
    > There is a DataPDULength which specifies max size of a *DATA* PDU,
    
    The DataPDULength is nebulous in its definition. It is'nt clear from its
    definition if it is the size of only a [scsi] *DATA* PDU, or it  specifies the
    max size of the data portion of all command and data PDUs. Also, what is the
    exact definition of "data payload" of command PDUs ? Does this include or
    exclude the additional header spanning beyond 48 bytes ?
    
    
    > and a
    > PingMaxReplyLength field which indicates how much buffer space an
    > implementation is willing to dedicate to Ping messages.
    
    >
    > Text strings that will not fit in a single Text command such that it's <=
    > DataPDULength can be sent as multiple text commands.
    
    If the login key negotiation can be performed as multiple text commands, how
    does the target know the initiator has sent its last text command of login keys
    to be negotiated ?
    
    The login phase is negotiated as follows :
    ==============================
    Initiator                                        Target
    ==============================
    1) Login    ----------->
                      <---------- 1) Login response of "accept login"
                                                    with F=0
    2) Text    ------------>
                  <-----------  2) Text Response
                :
                :
               <------------  3) Login Response of "accept login"
                                                   with F=1
    
    If the Text phase can consist of multiple text commands, how does a target know
    when to send in the login accept response with F=1 ? The target needs an
    indication that the initiator has sent in the last text command of the login
    phase and the "F" bit in the login command can provide that.
    
    >   Forget all this F bit stuff.
    
    If the login phase can comprise multiple text commands, the target needs a way
    to know the initiator has sent its final text command and the "F" bit provides
    this.
    
    OTOH, one could take the stance that the total number of login keys cannot
    exceed DataPDULength (and TotalText shall be >= DataPDULength), in which case
    the login phase cannot span beyond 1 text command.
    
    Regards,
     Santosh
    
    begin:vcard 
    n:Rao;Santosh 
    tel;work:408-447-3751
    x-mozilla-html:FALSE
    org:Hewlett Packard, Cupertino.;SISL
    adr:;;19420, Homestead Road, M\S 43LN,	;Cupertino.;CA.;95014.;USA.
    version:2.1
    email;internet:santoshr@cup.hp.com
    title:Software Design Engineer
    x-mozilla-cpt:;21088
    fn:Santosh Rao
    end:vcard
    


Home

Last updated: Tue Sep 04 01:05:45 2001
6315 messages in chronological order