A/B Port bit in Inquiry Data

Gerry_Houlder at notes.seagate.com Gerry_Houlder at notes.seagate.com
Mon Feb 2 12:31:48 PST 1998


* From the T10 (formerly SCSI) Reflector (t10 at symbios.com), posted by:
* Gerry_Houlder at notes.seagate.com
*
I also recall discussion of how to handle "more than two" ports. The
consensus was that the MultiP bit would indicate that "more than one" port
exists on that target. It was also thought to be inappropriate to reserve a
bunch of INQUIRY bits to identify up to 8 or 16 or whatever ports
(especially since no one in the room envisioned a target other than a RAID
controller that might have more than two ports).

The group recommended handling like this:
(a) check the MultiP bit to see if the target has more than one port.
(b) If it does, do we really care how many ports it has? Would you do
anything different if you knew it was a 4 port device versus an 8 port
device? The host computer probably can't get to all of the paths anyway,
some paths are only accessible to other host computers, or may be inactive
backups, or may be unused.
(c) If you care whether several different devices (in the host computer's
I/O port space) are really different paths to the same target, Use INQUIRY
VPD page to read the physical ID of the target. It was felt that the World
Wide Name Identifier in VPD page would allow the host to make the
associations needed.
(d) If the world wide name stuff doesn't suit your needs, then create a new
INQUIRY VPD page that does provide the port identification stuff you need.
If you can convince most of the T10 committee you need a new page and
convince at least one vendor you need the new page, then you will get it.

With the above algorithm in mind, the working group decided that the A/B
bit is not useful for new implementations (even 2 port implementations --
they can use the same algorithm described above for any multi-port device).
It was made vendor specific to protect vendors that have already
implemented 2 port devices using that bit. Seagate is a vendor with an
implementation using the A/B port bit.

The gist of all the above occurred when Charles Monia was trying to get a
multiple port device model into SAM-2 and was gathering opinions about what
to keep and what to throw away. Most of the documentation of these
discussions was lost when Charles moved on to other things ...

By the way, Ralph, your original reflector mail seemed to be a reply to
Larry (Lamers?) e-mail that didn't go out on the SCSI reflector. What was
the question you were trying to answer?


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




More information about the T10 mailing list