SORT BY:

LIST ORDER
THREAD
AUTHOR
SUBJECT


SEARCH

IPS HOME


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

    Re: iscsi: comments to iSCSI rev 6



    
    
    Santosh,
    
    This subject was discussed several times. The reason for using the LUN is
    to enable iSCSI proxies (or virtualizing devices) to operate as routers in
    the host to LUN direction and have them possibly insert the LUN field in
    the LUN to host direction.  It addition it enables tag allocation per LUN
    (not per target).
    
    Clearly the LUN could be any other routing tag but the LUN is already a
    well understood mechanism and will guarantee interoperability.
    
    I wonder if we should not make the LUN a mandatory field for NOP-out in
    general and enable "path checking" up to
    the LU when the using proxies.
    
    As for the tag name - it would be missleading as we do not require it to
    identify uniquely a task.
    
    Julo
    
    
    
    Santosh Rao <santoshr@cup.hp.com> on 04-05-2001 03:07:24
    
    Please respond to Santosh Rao <santoshr@cup.hp.com>
    
    To:   Matt Wakeley <matt_wakeley@agilent.com>, Julian
          Satran/Haifa/IBM@IBMIL
    cc:   IPS Reflector <ips@ece.cmu.edu>
    Subject:  Re: iscsi: comments to iSCSI rev 6
    
    
    
    
    Matt Wakeley wrote:
    
    > > Section 2.12.3: indicate that the LUN is copied from the NOP-IN.  This
    is
    > > much more clear than "the correct value for the task".
    > >
    > > +++ and open it up to strange thing like LU5 asking a nop-out to LU7 ?
    +++
    >
    > How will that occur?  The target sent the NOP-IN ping request.  The
    target
    > includes a LUN.  The only "correct value" that the initiator can return
    for
    > the LUN in the NOP-OUT ping response is the value the target sent in the
    > original request.
    
    Julian,
    
    Re-wording the LUN descriptions in NOP-IN & NOP-OUT to read that
    initiators MUST echo the LUN field from a received NOP-IN into a NOP-OUT
    is a more clear and un-ambiguous definition than the current text.
    
    That said, I still continue to have reservations about the need for a
    LUN field in a non-SCSI PDU. Fibre Channel does not use the LUN field in
    its ELS' & BLS'. LUN is a SCSI construct and it must not be used in
    non-scsi PDUs.
    
    I'm yet to hear strong reasons why LUN is required in the NOP-IN or
    NOP-OUT. The originator of the NOP operation must generate task tags
    independent of LUNs for non-SCSI PDUs. Such PDUs have no scsi context.
    
    As a side cosmetic comment, can we re-name "Target Transfer Tag" to
    "Target Task Tag" to have symmetry with the use of Initiator Task Tag ?
    
    - Santosh
     - santoshr.vcf
    
    
    
    


Home

Last updated: Tue Sep 04 01:04:40 2001
6315 messages in chronological order