Proposals posted: negotiation rewrite, WAKEUP cleanup, PR ignore target ports, etc.

Elliott, Robert Robert.Elliott at
Sat Jun 30 00:17:24 PDT 2001

* From the T10 Reflector (t10 at, posted by:
* "Elliott, Robert" <Robert.Elliott at>
These proposals have recently been posted to the T10 web site:

01-131r1 SPI-4 negotiation message rewrite
Includes negotiation sequence figures for initiator-originated and
target-originated WDTR and SDTR negotiations.  Please review these sequences
ahead of the July Parallel SCSI WG meeting. As suggested in the May WG I
made the initiator and target pessimistic when errors occur, generally
invalidating the transfer agreement if any problem occurs.

01-201r0 SPI-4 Add DISCONNECT OUT during IU mode to bus free list
Adds another case to the list of expected bus frees.

01-202r0 SPI-4 Obsolete AIP annex
Proposes that the asynchronous information protection (AIP) annex be dropped
|from SPI-4, since packetized protocol will be widely implemented in Ultra
320 devices and covers command and status information with CRC.

01-134r1 SAM-2, SPC-3, SPI-4, SBC-2 WAKEUP and reset cleanup
Proposes moving the power condition model into SPC-3 because it applies to
all device types.  Defines WAKEUP in SAM-2 with an opening statement that
protocols may or may not support it.  Introduces terms "wakeup" and "wakeup
event" to try to make all 4 affected standards use the same terminology.
Splits power condition into logical unit power condition and target port
power condition to clarify behavior for multi-LU and multiport devices.

01-099r2 SPC-3 Letting persistent reservations ignore target ports
Drops the "ignore target port" bit in rev 1.  Instead, proposes that
protocols with initiator port names that are defined to be world wide unique
are free to ignore the target port through which a reservation is made.
01-182r0 by Jim Hafner defines the "initiator port name" employed by this

01-198r0 SPC-3 CMDDT and variable length CDBs
This revision is the same as presented in the May CAP WG - I didn't have a
document number at that time and just obtained it.  It highlights a problem
with the INQUIRY CMDDT feature (which reports which bits are supported in a
CDB) and CDBs with service actions, including the variable length CDBs.  I
hope to provide a rev 1 with a proposed fix before the July CAP WG meeting.

Rob Elliott, Compaq Server Storage
Robert.Elliott at

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

More information about the T10 mailing list