SORT BY:

LIST ORDER
THREAD
AUTHOR
SUBJECT


SEARCH

IPS HOME


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

    Re: iscsi : MaximumBurstSize & FirstBurstSize values.



    
    ----- Original Message -----
    From: "Santosh Rao" <santoshr@cup.hp.com>
    To: "IPS Reflector" <ips@ece.cmu.edu>
    Sent: Friday, September 21, 2001 11:20 PM
    Subject: Re: iscsi : MaximumBurstSize & FirstBurstSize values.
    
    
    > Robert Snively wrote:
    >
    > > I am sorry, but I don't think this is a necessary or appropriate
    >
    > > limitation.  FirstBurstSize is a PDU size, and its characteristics
    > > are determined by the unsolicited data buffer capability of the
    > > target.  MaximumBurstSize is a PDU size, and its characteristics
    > > are determined by the solicited data buffer size and the buffer
    > > size increments available to the initiator.  These may meet that
    > > rule, but they may also not meet that rule.  That is why there
    > > are two values.
    > >
    > > It may be important to clarify that FirstBurstSize overrides
    > > MaximumBurstSize for the first burst.
    >
    > Bob,
    
    Although I agree with Bob here but isnt it worth thinking, why should the
    firstburstsize be greater than MaximumBurstSize. I think in all practical
    implemetation firstburstsize will be lesser than MaximumBurstSize.
    
    Considering bob's explanation, It will not be a good idea to put any
    restrictions with respect to size but  just to avoid some confusion it might
    be better to rename FirstBurstSize as UnsolicitedBurstSize and
    MaximumBurstSize as MaximumSolicitedBurstSize.
    
    >
    > Your solution works fine for me.
    >
    > The current draft wording is not clear on how the initiator must send
    > un-solicited data when MaximumBurstSize is < FirstBurstSize, since the
    > meaning of the "F" bit in the data-out sequence is currently over-loaded
    to
    > imply end-of-sequence as well as end-of-unsolicited data. Sequence length
    > could be construed as less than length of un-solicited data.
    >
    > Regards,
    > Santosh
    >
    >
    >
    > > > -----Original Message-----
    > > > From: Santosh Rao [mailto:santoshr@cup.hp.com]
    > > > Sent: Friday, September 21, 2001 11:43 AM
    > > > To: IPS Reflector
    > > > Subject: iscsi : MaximumBurstSize & FirstBurstSize values.
    > > >
    > > >
    > > > All,
    > > >
    > > > I believe the FirstBurstSize setting has to be <= MaximumBurstSize ? I
    > > > suggest that this be explicitly called out in the definition of
    > > > FirstBurstSize.
    > > >
    > > > If this rule is not followed, an initiator would need to break the
    > > > un-solicited data it sends into more than 1 data-out sequence, and in
    > > > setting the F bit to indicate end-of-sequence, it can confuse
    > > > the target
    > > > into thinking its end of un-solicited data.
    > > >
    > > >
    > > > The definitions currently read :
    > > >
    > > > "4.1.1 MaximumBurstSize Field (16 bit)
    > > >
    > > > This field defines the maximum SCSI data payload in an iSCSI
    > > > sequence (a
    > > > sequence of Data-In or Data-Out PDUs ending with a Data-In or Data-Out
    > > > PDU with the F bit set to one) in units of 512 bytes. The      default
    > > > value assumed for a new iSCSI session (I_T nexus) is 512 units.
    > > >
    > > > 4.1.3 FirstBurstSize Field (16 bit)
    > > >
    > > > This field defines the maximum amount of unsolicited data an iSCSI
    > > > initiator may to send to the target in units of 512 bytes. The default
    > > > value assumed for a new iSCSI session (I_T nexus) is 128 units. "
    > > >
    > > > Regards,
    > > > Santosh
    > > >
    >
    
    


Home

Last updated: Sun Sep 23 20:17:21 2001
6682 messages in chronological order