Serial support of RESERVE(6)/RELEASE(6)

dallas at dallas at
Mon Jun 17 12:45:27 PDT 1996

* From the SCSI Reflector, posted by:
* dallas at


The solutions proposed does solve the problem in 
an across the board consistent manner.

Solution 1 requires the peripheral drivers to have
knowledge of the inter-connect.

Solution 2 does not solve the problem.

A solution would be to obsolete RESERVE and
RELEASE (6) for SCSI-3 and mandate the 10 byte
versions for SCSI-3 devices.  This solution is 
only a part solution since it does not address
those devices that have soft addressing 
(FCP, SCAM, etc).

Persistent Reserve could be a solution for SCSI-3
devices where the RESERVE and RELEASE commands are
obsoleted for SCSI-3 and Persistent Reserve mandated. 

>From my view point any solution that does not require
the peripheral drivers to have specific knowledge
of an inter-connect is good.  The first solution you 
proposed  would (in the worst case) require the peripheral 
drivers to have code for each inter-connect for SCSI 
reservations (e.g. FCP, SSA, Narrow SPI, Wide SPI, etc).

The generic solution of changing SPC for SCSI-3, only
requires the drivers to have generic SCSI-2 and SCSI-3
knowledge (e.g., if SCSI-2 device do this versions 
specifics and if SCSI-3 do that versions specifics.


------- Replied-To Message

Date: Mon, 17 Jun 96 14:38:07 EDT
From: at (MAIL-11 Daemon)
  To: scsi at Symbios.COM
Subj: Serial support of RESERVE(6)/RELEASE(6)

* From the SCSI Reflector, posted by:
* scheible at VNET.IBM.COM
I have placed this note on the SCSI and SSA reflectors.  Sorry for any

   How can serial interfaces support 3 bit addresses?

   Currently SPC and SBC state that RESERVE(6) and RELEASE(6) are
mandatory (for block devices).  However, these commands only have a
3 bit address.  This means that S3P(SSA), FCP (FC), SBP (1394) and
even 16 bit SCSI cannot address all possible devices (only 8).
Fortunately, RESERVE(10) and RELEASE(10) fix this problem by using
a 64 bit address.

   SSA uses an 8 byte Unique ID (IEEE address) as the device ID for
SAM.  FCP maps a port address into the 8 byte field.  I did a quick
scan of FCP and SBP and could not find any mention on how to address
devices in these commands.  Can anyone help on this issue?

   I can see several solutions:

1) Make the command support transport layer specific, then the serial
   interfaces can support RESERVE(10)/RELEASE(10) and not support the
   6 byte versions.  This requires X3T10 support.

2) Take a literal translation (like 16 bit SCSI) and only support
   devices 0000000000000000h thru 0000000000000007h.  In SSA, since
   the high bits are a company ID, this would not allow most (all?)
   companies from use this command.  In 16 bit SCSI, you can only
   Reserve/Release the low 8 devices.

   Anyone have any input?

John Scheible
scheible at

% ====== Internet headers and postmarks (see DECWRL::GATEWAY.DOC) ======
% Received: from by (5.65/rmc-22feb94) id
   AA16226; Mon, 17 Jun 96 14:24:35 -040
% Received: from by (8.7.5/UNX 1.2/1.0/WV
  ) id OAA20571; Mon, 17 Jun 1996 14:17:42 -0400 (EDT
% Received: (from root at localhost) by ( id MAA0
  3833; Mon, 17 Jun 1996 12:10:18 -0600
% Received: from by via
   smap (V1.3) id smaa03703; Mon Jun 17 12:08:52 199
% Received: (from majordom at localhost) by Symbios.COM ( id MAA1813
  4 for scsi-outgoing; Mon, 17 Jun 1996 12:08:19 -0600
% Received: from ([]) by Symbios.COM (
  .6.6) with ESMTP id MAA18129 for <scsi at>; Mon, 17 Jun 1996 12:08:1
  3 -0600
% From: scheible at
% Received: (from root at localhost) by ( id MAA0
  3616 for <scsi at>; Mon, 17 Jun 1996 12:08:14 -0600
% Message-Id: <199606171808.MAA03616 at>
% Received: from by via smap (V1.
  3) id sma003567; Mon Jun 17 12:08:02 199
% Received: from SJEVM5 by VNET.IBM.COM (IBM VM SMTP V2R3) with BSMTP id 7312; 
  Mon, 17 Jun 96 14:07:45 ED
% Date: Mon, 17 Jun 96 11:04:25 PDT
% To: scsi at Symbios.COM
% Subject: Serial support of RESERVE(6)/RELEASE(6)
% Sender: owner-scsi at Symbios.COM
% Precedence: bulk

------- End of Replied-To Message

More information about the T10 mailing list