Load Unload Hold bit vs. Control Mode page Autoload Mode

Paul.A.Suhler at seagate.com Paul.A.Suhler at seagate.com
Wed Oct 2 09:11:02 PDT 2002


* From the T10 Reflector (t10 at t10.org), posted by:
* Paul.A.Suhler at seagate.com
*
I agree with Paul.  Autoload and execution of a LOAD UNLOAD command are
separate things.  In the case that he cites, I would expect a drive to
finish the autoload and then execute the command.

The reason for the text linking the Autoload field in the mode page and the
Hold field in the command was to ensure a symmetry of behavior:  if a tape
could be loaded to the hold point by autoload, then it could also get there
by executing a command.  That said, the wording was too broad, as it
required supporting the Hold bit in the command if the No Load (2) value
was supported for Autoload but not the Load To Hold (1).  I was probably
trying to get implementers to buy the whole thing when I first proposed
this (T10/99-263), and then later thought better of it during SSC-2 review
(Seagate 21).

I don't remember now why we decided to remove the entire paragraph.

Cheers,

Paul Suhler
Seagate Removable Storage Solutions

----- Forwarded by Paul A Suhler/Seagate on 10/02/2002 08:47 AM -----

* From the T10 Reflector (t10 at t10.org), posted by:
* Paul Entzel <Paul.Entzel at quantum.com>
*
Kevin,
My interpretation of the AUTOLOAD MODE field is that if affects the
operation of the device for an action that does not involve a SCSI command
to that device.  I came to this conclusion based on the description of
AUTOLOAD MODE from SPC-3 "The AUTOLOAD MODE field specifies the action to
be
taken by a removable medium device server when a medium is inserted."  No
mention of a SCSI command there.

Perhaps a better question is, can this "action to be taken by a removable
medium device server" be over-ridden by a SCSI command, for instance the
LOAD UNLOAD command.  In other words, if the AUTOLOAD mode says to load the
medium for full access, but while this process is happening a LOAD UNLOAD
command is received with the HOLD bit set, should the device server abort
the current load process and perform the SCSI command?  My opinion is that
it should not.  Other actions by a device server involve enabling a
command,
and only one command can be enabled at a time.  The device server should
finish the implicit load operation initiated by inserting the medium before
enabling SCSI commands that also affect the medium.

The paragraph that was removed does not answer this question either.  All
that sentence did was require support for the HOLD bit if the AUTOLOAD MODE
field is changeable or non-zero.

Paul Entzel
Quantum

-----Original Message-----
From: Kevin D Butt [mailto:kdbutt at us.ibm.com]

All,

I have a question regarding the removal from the Load Unload command the
paragraph
*******
If the AUTOLOAD MODE field in the control mode page (see SPC-2) is
changeable or a MODE SENSE
command reports a value in the AUTOLOAD MODE field other than zero, the
device server shall support a
value of one for the HOLD bit.
*******
|from SSC-2.  Was this intentional?  We need to state the precedence of the
AUTOLOAD MODE vs. the LOAD UNLOAD HOLD bit.  If the mode page setting is
changeable do we ignore the HOLD bit?  Does the HOLD bit take precedence
over the mode page setting? etc.?

Regards,

Kevin D. Butt
Fibre Channel & SCSI Architect
IBM Tape Microcode,
6TYA, 9000 S. Rita Rd., Tucson, AZ  85744
Tie-line 321; Office: 520-799-5280, Lab: 799-2869, Fax: 799-4138, Email:
kdbutt at us.ibm.com

*
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at t10.org


*
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at t10.org




More information about the T10 mailing list