SORT BY:

LIST ORDER
THREAD
AUTHOR
SUBJECT


SEARCH

IPS HOME


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

    RE: iSCSI Plugfest at UNH (misc issues) (fwd)



    >For convience of programming at the initiator, it would be best to let the 
    >initiator decide if it wants to use the ITT or not.
    
    Actually, adding this check should be quite trivial for initiators.
    Something like:
    
    if (!ulp_scsi_command->is_tagged) {
      iscsi_scsi_cmd.itt = 0xffffffff;
    } else {
      iscsi_scsi_cmd.itt = iscsi_itt++;
    }
    
    Linux initiators could do something like:
    /* in queuecommand() */
    
    if (!SCpnt->tag) {
      iscsi_scsi_cmd.itt = 0xffffffff;
    } else {
      iscsi_scsi_cmd.itt = iscsi_itt++;
    }
    
    This doesn't seem to be an unreasonable request if it significantly
    simplifies target code.
    
    >What is the problem from the target's standpoint if there is a value in the 
    >ITT for an untagged request?
    >
    >Eddy
    >
    >
    >----Original Message Follows----
    >From: Mike Brown <mbrown@cs.uml.edu>
    >To: rod.harrison@windriver.com
    >CC: ips@ece.cmu.edu, sandeepj@research.bell-labs.com
    >Subject: RE: iSCSI Plugfest at UNH (misc issues) (fwd)
    >Date: Mon, 23 Jul 2001 14:21:59 -0400 (EDT)
    >
    >Rod,
    >
    > >
    > >	The initiator needs the initiator task tag in order to
    > >associate inbound PDUs with the correct command. For
    > >untagged commands the ATTR field of the SCSI Command PDU
    > >should be set to 0, Untagged.
    >
    >The initiator associates inbound PDU's with the itt.  I agree with
    >Sandeep.  If the SCSI Command is untagged, then the itt should
    >be explicitly set to 0xffffffff in the iSCSI PDU since the initiator can
    >still use the itt to associate inboud iSCSI PDU's with a current SCSI
    >command because SAM-2 sec 4.9.1 states:
    >
    >"An untagged task does not include a tag in any of its component
    >definitions, thus restricting the number of concurrent untagged tasks in a
    >single task set to _ONE_ per initiator"
    >
    >If there are multiple outstanding tasks, (one untagged task and multiple
    >tagged tasks), each task will still have a unique itt, and therefore the
    >initiator can associate inbound iSCSI PDU's with a particular SCSI
    >command.
    >
    > >
    > >	- Rod
    > >
    > >-----Original Message-----
    > >From: owner-ips@ece.cmu.edu [mailto:owner-ips@ece.cmu.edu]On
    > >Behalf Of
    > >Sandeep Joshi
    > >Sent: Monday, July 23, 2001 10:31 AM
    > >To: ips@ece.cmu.edu
    > >Subject: Re: iSCSI Plugfest at UNH (misc issues)
    > >
    > >
    > >
    > >
    > >Hi Julian,
    > >
    > >Some additional items from the plugfest to resolve..
    > >
    > >1) Sec 2.3.1 : Untagged tasks.
    > >   If the task attribute in the SCSI Command is "untagged"
    > >   then the initiator task tag should be ignored, correct ?
    > >   Can the draft explicitly state that the initiator must
    > >   set the Initiator Task Tag to 0x'ffffffff if its an
    > >   untagged task.
    > >
    > >   Do we assume that the initiator ULP will not issue more
    > >than
    > >   one untagged tasks simultaneously ?
    > >
    > >2) ExpCmdSN
    > >   What are the semantics of something like this.. the
    > >   target does not increase the expCmdSN but keeps sending
    > >   the status (SCSI response) for commands after the
    > >expCmdSN ?
    > >   Clearly, the target has received and executed the
    > >commands
    > >   (in cmdSN order).
    > >
    > >   Hence, the expCmdSN in a SCSI response must not be less
    > >   than the cmdSN of the original command (for which
    > >response
    > >   is being received).
    > >
    > >   This seems like a valid property which could be mandated
    > >   and checked, to allow efficient initiator
    > >implementations.
    > >   The target eventually suffers but the standard could
    > >   prevent such targets from being deployed :-)
    > >
    > >3) Sending status in read PDU
    > >   Since quite a few initiators did not support this,
    > >   this feature had to be enabled or disabled frequently at
    > >   the target.
    > >
    > >   Either we should make it mandatory to implement (at
    > >initiator)
    > >   or we should add a key "SendStatusInReadPDU=yes/no".  The
    > >first
    > >   option seems simpler.
    > >
    > >thanks
    > >-Sandeep
    > >
    >
    >
    >-Michael F. Brown, UMass Lowell Computer Science
    >
    >phone:  (978) 934-5354
    >email:  mbrown@cs.uml.edu
    >
    >"If a driver behaves incorrectly, the Driver Verifier will crash the system
    >  immediately. In this way, the Driver Verifier prevents the driver bug from
    >  being masked by further processing. The result is a faster, easier 
    >debugging
    >  process."          -Windows 2000 Driver Development Kit Release Notes
    >
    >
    >_________________________________________________________________
    >Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp
    >
    
    
    -Michael F. Brown, UMass Lowell Computer Science
    
    phone:  (978) 934-5354
    email:  mbrown@cs.uml.edu
    
    "Black holes are where god divided by 0."
    
    


Home

Last updated: Tue Sep 04 01:04:14 2001
6315 messages in chronological order