Response to SPI-4 proposal 01-251

John Lohmeyer lohmeyer at t10.org
Fri Sep 7 07:02:05 PDT 2001


* From the T10 Reflector (t10 at t10.org), posted by:
* John Lohmeyer <lohmeyer at t10.org>
*
Richard,

My apologies for taking so long to respond to your emails regarding my 01-251
proposal.  I was on vacation until yesterday.

I had a strong sense of deja vue regarding this issue and did some digging
through the X3T9.2 archives.  I found a very similar 1992 request for
interpretation on SCSI-2. In X3T9.2/92-212, Mr. Junzo Neriishi of Toshiba
Corp. submitted a request for interpretation regarding synchronous data
transfers.  They had run into a bus hang condition because the target chip
did not continue to generate REQ pulses up to the REQ/ACK offset and the
initiator chip did not send ACK pulses for all received REQ pulses.

Sound familiar?

X3T9.2 concluded that:

"While no one at the meeting could point to specific violations of SCSI-2 in
either of the implementations you describe, both fall short of the intent of
the synchronous data transfer protocol. A change to either implementation
could both fix the problem and improve data transfer rates."

Since SCSI-2 was in the approval stage, X3T9.2 could not easily fix SCSI-2 so
it was agreed to fix the problem in SPI.  X3T9.2/92-235 was accepted for
inclusion in SPI.  It essentially made both of these initiator and the target
designs illegal in SPI by requiring targets to generate ACK pulses until the
REQ/ACK offset was reached and by requiring initiators to continue sending
ACK pulses until the number of ACK pulses reached the number of REQ pulses
received.

I think the words Bill Galloway quoted stemmed from the 92-235
proposal.  These SPI words got lost when we moved to SPI-2 because it was
easier for the technical editor to go back to SCSI-2 for base wording as he
merged SPI, SIP, and Fast-20.

I feel that it is irresponsible for T10 to ignore this issue.  We must take
some action to close the loophole.

I am sorry you will not be at the meeting next week.  Since your company is
the host, is it possible to have someone cover for you at the meeting?

Regards,
John

PS:  I found a paper copy of X3T9.2/92-212 will attempt to scan it and post
it.  Both 92-235 and 92-242 are on the T10 web site at the following URLs:

    ftp://ftp.t10.org/x3t9.2/document.92/92-235r0.pdf

    ftp://ftp.t10.org/x3t9.2/document.92/92-242r0.pdf


--
John Lohmeyer                  Email: lohmeyer at t10.org
LSI Logic Corp.                Voice: +1-719-533-7560
4420 ArrowsWest Dr.              Fax: +1-719-533-7183
Colo Spgs, CO 80907 

*
* 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