Re: iSCSI: Inquiry, Mode Sense, Read Capacity

[Sandeep Joshi <sandeepj@research.bell-labs.com>]

Just a small correction on this topic..
  > There is no Target analog to the Initiator's ability to send
  > Immediate Data in the same PDU as the SCSI command.

There is an analog, but its part of the Data-IN PDU
See Section 3.7.
  > Status can accompany the last Data-in PDU if the command did not end 
  > with an exception.  Presence of status (and of a residual count) is 
  > signaled though the S flag bit.

-Sandeep

Black_David@emc.com wrote:
> 
> The governing text is the following from p.62 of -08 on the contents of
> the Response PDU:
> 
>      3.4.6 Data Segment - Sense and Response Data Segment
> 
>         iSCSI targets MUST support and enable autosense.  If Status is CHECK
> 
>         CONDITION (0x02), then the Data Segment contains sense data for the
>         failed command.
> 
>         For some iSCSI responses, the response data segment MAY contain some
> 
>         response related information, (e.g., for a target failure it may
>         contain a vendor specific detailed description of the failure).
> 
> Julian may be able to shed more light on the intended meaning of
> "phase collapse" for the response case, but the above text makes
> it quite clear that data transferred by the action of a SCSI command
> (including Read, Inquiry, Mode Sense, and Read Capacity) cannot be
> put in the Response PDU.  A target can send the Response PDU immediately
> following the last Data-In PDU for the associated command, which is
> similar to the ability of an Initiator to send unsolicited Data-Out
> PDUs.  There is no Target analog to the Initiator's ability to send
> Immediate Data in the same PDU as the SCSI command.
> 
> Thanks,
> --David
> 
> > -----Original Message-----
> > From: Lee Xing [mailto:lxing@crossroads.com]
> > Sent: Wednesday, October 03, 2001 2:00 PM
> > To: Black_David@emc.com; ips@ece.cmu.edu
> > Subject: RE: iSCSI: Inquiry, Mode Sense, Read Capacity
> >
> >
> > David,
> >
> > Thanks for the info.  It helps.
> > Please see my new question below.
> >
> >
> > > Q2:
> > > There are 10 target opcodes on page 43 of v.07, and two of
> > > them are specified as:
> > > (-snip)
> >
> > See previous answer.  Data-in transfers data for any SCSI command
> > that has data to transfer.  This includes all three of your examples.
> > "READ" actually includes any SCSI command that transfers "data" (as
> > that term is defined by SCSI) from Target to Initiator.  This includes
> > INQUIRY, MODE SENSE, and READ CAPACITY.
> >
> >
> >
> > New Question:
> >
> > on page 23 of v.08, it says "For performance reasons, iSCSI allows
> > a "phase-collapse".  A command and its associated data may be
> > shipped together from initiator to target and data and responses
> > may be shipped together from targets."
> >
> > The question is if "Data-in transfers data for any SCSI command
> > that has data to transfer" (see the answer to Q2) is a true
> > statement, should we clear the above sentence (on page 23 of v.08)
> > a bit?
> >
> > Thanks.
> >
> >
> >
> > Lee
> > Crossroads Systems, Inc.
> >