Results of SPI-3 Letter Ballot

John Lohmeyer lohmeyer at
Fri Oct 29 09:27:34 PDT 1999

* From the T10 Reflector (t10 at, posted by:
* John Lohmeyer <lohmeyer at>
Voting Results on T10 Letter Ballot 99-015r0 on
Forwarding SPI-3 to first public review

Organization                      Name                 S Vote Add'l Info
--------------------------------- -------------------- - ---- ----------
Adaptec, Inc.                     lawrence lamers      P Yes  
Advansys                          Robert Frey          P YesC Cmnts 
AMP, Inc.                         Charles Brill        P Yes  
Amphenol Interconnect             BILL MABLE           P No   Cmnts 
Ancot Corp.                       Bart Raudebaugh      P Yes  
Andataco                          Gregg Neely          P Yes  
Apple Computer                                           DNV  
Berg Electronics                  Douglas Wagner       P Yes  
BREA Technologies, Inc.           Bill Galloway        P No   Cmnts 
Circuit Assembly Corp.            Ian Morrell          P No   Cmnts 
CMD Technology                    Edward Haske         P Yes  
Compaq Computer Corp.             Bill Ham             A No   Cmnts 
Crossroads Systems, Inc.          Neil T. Wanamaker    P YesC Cmnts 
Dallas Semiconductor              Charles Tashbook     P Yes  
Dell Computer                                            DNV  
Distributed Processing Tech.      Roger Cummings       P Yes  
ENDL                              I D Allan            P Yes  
Fujitsu                           Eugene Lew           P Yes  
General Dynamics                  Nathan Hastad        P Yes  
Hewlett Packard Co.               J. R. Sims, III      P Yes  
Hitachi Cable Manchester,Inc      Zane Daggett         P No   Cmnts 
Hitachi Storage Products          Anthony Yang         P Yes  
Honda Connectors                  Thomas J. Kulesza    P Yes  
IBM Corp.                         George Penokie       P Yes  
Iomega Corp.                      Tim Bradshaw         P Yes  
KnowledgeTek, Inc.                Dennis P. Moore      P Yes  
Linfinity Micro                   Louis Grantham       P Yes  
LSI Logic Corp.                   John Lohmeyer        P Yes  
Madison Cable Corp.               Jie Fan              P No   Cmnts 
Maxtor Corp.                      Pete McLean          P Yes  
Molex Inc.                        Joe Dambach          P Yes  
Mylex Corp.                                              DNV  
Ophidian Designs                  Edward A. Gardner    P Abs  IV Cmnts 
Panasonic Technologies, Inc       Han Zou              P Yes  
Philips Electronics               Bill McFerrin        P Yes  
QLogic Corp.                      Skip Jones           P Yes  
Quantum Corp.                     Mark Evans           P YesC Cmnts 
Seagate Technology                Gene E. Milligan     P No   IV Cmnts 
Storage Technology Corp.          Erich Oetting        P Yes  
Sun Microsystems Computer Co      Robert Snively       P Yes  
Texas Instruments                 Robert Morris        P Yes  
Toshiba America Elec. Comp.       Tasuku Kasebayashi   P Yes  
UNISYS Corporation                Ken Hallam           P Yes  
Unitrode Corporation              Paul D. Aloisi       P YesC Cmnts 
Western Digital Corporation       Jeffrey L. Williams  P Yes  

P       Voter indicated he/she is principal member
A       Voter indicated he/she is alternate member
O       Voter indicated he/she is observer member
?       Voter indicated he/she is not member or does not know status
YesC    Yes with comments vote
Abs     Abstain vote
DNV     Organization did not vote
IV      Individual vote (not organizational vote)
Cmnts   Comments were included with ballot
NoCmnts No comments were included with a vote that requires comments
DUP     Duplicate ballot (last ballot received from org. is counted)
PSWD    The password was not correct (vote not counted)
ORG?    Organization is not voting member of T10 (vote not counted)

Ballot totals:
 34 Yes
  7 No
  1 Abstain
  3 Organization(s) did not vote
 45 Total voting organizations
 12 Ballot(s) included comments

This 2/3rds majority ballot passed.


Comments attached to YesC ballot from Robert Frey of 

Comment #1: Recommend additional information units phase sequence figure

  The following phase diagrams are included in the current revision:
  13.1 Phase sequences for physical reconnection and selection using
       attention condition with information unit transfers disabled
  13.2 Phase sequences for selection without using attention condition
       with information unit transfers disabled
  13.3 Phase sequences for selection without using attention condition/
       physical reconnection with information unit transfers disabled

  I recommend another phase sequence figure: 

  13.4 Phase sequences for selection using attention condition with
       information unit transfers enabled

  Without a 13.4 phase sequence figure the text of section is
  not represented in SPI-3 Section 13. There is value to having Section
  13 include all prescribed/legal sequences, which is what I think it
  is attempting to do.
  Text from
    10.3.1 Selection Selection using attention condition Information unit transfers enabled
      If information unit transfers are enabled for the connecting initiator
      the target shall proceed to MESSAGE OUT phase.

Comment #2: Data group transfer wide residual reporting and restriction

  Proposal A:
    Add text to that for DT DATA IN phase will state the correct
    procedure for reporting wide residual and for DT DATA OUT phase require
    intermediate data fields to be an even number of bytes.

    Proposed wide residual reporting clarification text and restriction to be
    added to Data group transfers:
      During DT DATA IN phase if the number of bytes in a data field is not
      a multiple of two bytes, then after sending the pad and pCRC fields the
      target device shall change to MESSAGE IN phase and send an IGNORE WIDE
      RESIDUE 16.2.3 message with the NUMBER OF BYTES TO IGNORE field set to
      During DT DATA OUT phase if a target requests a pCRC field prior to
      the last data field of a command, the initiator shall transmit an even
      number of bytes in that data field.
    The first paragraph will lower the probability of incorrect initiator
    and target implementations.
    The second paragraph is needed because there is no mechanism defined for
    an initiator to indicate an intermediate wide residual to a target. A
    target is only able to infer a wide residual at the very end of a wide
    transfer transfer by recalling whether the command transfer length.

  Proposal B:
    Add a requirement that all intermediate data fields (pCRC interval)
    be a multiple of two bytes. This would be consistent with IU transfers
    ( which require the iuCRC interval to be an even number of bytes.

Comment #3: Typo

  4.7, paragraph two, 2nd sentence "2X depending on the whether" should
  be "2X depending on whether".


Comments attached to No ballot from BILL MABLE of 
Amphenol Interconnect:


I was unable to get onto the reflector but the reason we are voting no 
pertains to the comments submitted by Bill Ham pertaining to the techinical 
changes. I will try and get you the file name for the comments that were 
submitted by Bill Ham.



   Chair's Note: The referenced document is T10/99-317r0, Comments
                 against SPI-3 rev 10 section 6 and annex E


Comments attached to No ballot from Bill Galloway of 
BREA Technologies, Inc.:

Comments sent seperate

   Chair's Note: Bill Galloway's comments are contained in T10/99-320r0


Comments attached to No ballot from Ian Morrell of 
Circuit Assembly Corp.:

I incorporate by reference the "comments against SPI 3 rev 10, Section 6 and 


   Chair's Note: The referenced document is T10/99-317r0


Comments attached to No ballot from Bill Ham of 
Compaq Computer Corp.:

Comments are contained in the following two documents:

The first document is titled:

Comments against SPI3 rev 10 section 6 and annex E.doc

which has been mailed to you and posted on the web site as document number 

The second document is titled:

CPQ 1xx SPI-3 letter ballot comments.doc. 

which has been emailed to you today.

   Chair's Note: This document was numbered T10/99-321r0.

The technical comments in the first document and the first three comments in 
the second document need to be accepted in order to change the vote to "yes".

The numbering of the comments has a gap between 64 which is the last of the 
first document and 100 which is the first of the second document.

I hope that providing two documents is not a great inconvenience.

Cheers, Bill


Comments attached to YesC ballot from Neil T. Wanamaker of 
Crossroads Systems, Inc.:

(T), Mandating a check condition in these cases 
seems a retrograde step.


Comments attached to No ballot from Zane Daggett of 
Hitachi Cable Manchester,Inc:

61 comments have been submitted in a document uploaded to the ftp site 
entitled "comments against SPI-3 rev 10 section 6 and annex".  This 
document should serve as my reasoning for voting no on this ballot.  

If this method is not acceptable please contact me at 603-661-3972.  I did 
not paste these comments here because they include many graphics.


   Chair's Note:  The referenced document is T10/99-317r0.


Comments attached to No ballot from Jie Fan of 
Madison Cable Corp.:

The current cable performance requirements and it's related Annex should be 
replaced with a newer version of 99-111r7 (for cable requirements) and 98-
219r6 (for Annex)-- Jie Fan


Comments attached to Abs ballot from Edward A. Gardner of 
Ophidian Designs:

My reasons for abstaining are that I don't feel technically 
qualified to competently evaluate SPI-3.  


Comments attached to YesC ballot from Mark Evans of 
Quantum Corp.:

Quantum's comments for the letter ballot to forward SPI-3, rev 10

Quantum # 1. Global
The terms "clause" and "subclause" are used in many places to reference 
clauses or subclauses elsewhere in the standard.  However, these terms are not

used consistently.  Sometimes just the clause or subclause number to be 
referenced is used.  Sometimes "clause" is used where a subclause is 
referenced.  I think that the terms should be deleted, leaving only the number

for reference (e.g., "See x.y for...")

Quantum # 2. Global
The terms "annex X" and "Annex X" are used inconsistently throughout the 
standard.  I think "Annex X" is correct, but, one way or the other, they 
should, at least, be consistent.

Quantum # 3. Global
The terms "cross talk", "cross-talk", and "crosstalk" are used inconsistently 
throughout the standard.  My dictionary says it should be "crosstalk", Word's 
spellchecker likes" cross talk" and "cross-talk", but, one way or the other, 
they should, at least, be consistent.

Quantum # 4. Global
The term "REQ(ACK)" - and sometimes "REQ (ACK)" - are used in many places 
where I think the meaning is "REQ or ACK" or "REQ and ACK".  I think the 
single term should be replaced by the correct three-word phrase or defined in 
3.1.  If defined in 3.1, it would seem to me that then any "REQ or ACK" or 
"REQ and ACK" in the document should be replaced by "REQ(ACK)".

Quantum # 5. Global

There are many instances where text in a list should be indented.  These 
include:  page 1 (PDF page 23), 1 Scope, the list after paragraph 1; page 15 
(PDF page 37), 4.3 Physical topologies...; page 21 (PDF page 43), 4.9 
Protocol, the list at the bottom of the page; page 23 (PDF page 45), 5.1 SCSI 
parallel..., the list near the bottom of the clause; page 46 (PDF page 68), 
two places; etc.

Quantum # 6. Global
The symbols "+" and "-" are used in several places and should be replaced by 
the words "plus" and "minus" because these symbols are defined as "add" and 
subtract" in subclause 3.2 Symbols and abbreviations.  I think it also might 
work to put quotation marks around the symbols in this instance, but that 
would probably also require additional entries in 3.2 - which is also okay by 
me.  (For examples see page 17 (PDF page 39), 4.4 Bus loading, paragraph 1, 
and page 52 (PDF page 74), 6.7.1 LVD stub length and spacing).

Quantum # 7. Global
The term DATA BUS is used in many places in the document with inconsistent 
meaning.  The definition in 3.2 says "data bus [no caps]..." is an "...8-bit 
or 16-bit bus."  However, on page 79 (PDF page 101), in 8.2 Signal 
descriptions, I/O:, for example, we find "DATA BUS [the capitalization of this

term in the document is inconsistent, as well]".  In this particular 
description I think we mean "...the data bus [as defined in 3.2], P_CRCA, and 
P1 (if present for a 16-bit DATA BUS)..." or "...DB(7-0, P_CRCA) or DB(7-0, 
P_CRCA, P1)..."  On page 87 (PDF page 109), in 9.2.13 Data release delay, I 
think the same thing is true.  I'm not sure exactly how this should be 
rectified, but a search needs to be performed on all forms of "data bus" and a

correct definition included for each case.

Quantum # 8. Global
On page 86 (PDF page 108) 9.2.1 Arbitration delay, the phrase "arbitration has

been won" is used.  I can't find where in the draft standard that "winning 
arbitration" or "losing arbitration" are defined.  I think the readers should 
know what these means, but I also think that these concepts are such critical 
elements of parallel SCSI that the conditions should be specifically defined 
in this document.

Quantum # 9. Global
The names of the timing values listed in tables 30, 31, and 32 are used 
inconsistently throughout the standard.  Sometimes the first letter of each 
word is capitalized and sometimes not.  Since these are specific, defined 
things, I recommend that the first letter of each word be capitalized (e.g., 
Bus Settle Delay) wherever used.  I think this would be more clear.

Quantum # 10. Global
On page 103 (PDF page 125), Selection without using attention 
condition, paragraphs 2 and 3, are the words " 
initiator...waits...before...looking..."  I think this would be better stated 
as something like, "...waits before enabling detection..."  I think a global 
search should be performed on "looking" and, where used as above, it should be


Quantum # 11. Global
On page 103 (PDF page 125), Selection without using attention 
condition In this paragraph 4, are the words "a target...shall determine that 
it is selected..."  I think this should be something like, "A target shall be 
selected when..."  I think a global search should be performed on "determine" 
and, where used as above, it should be changed.

Quantum # 12. Global
The phrases "bad parity" and "parity error" are used in many places in the 
document.  I searched on "parity" and found no specific definition for these 
conditions in the document.  Though the definition of these phrases should be 
intuitive, I think that this is such a critical element of parallel SCSI that 
the condition should be specifically defined in the document.

Quantum # 13. page 1 (PDF page 23), 1 Scope, paragraph 1:
There should be a colon after the "are" at the end of the first paragraph.

Quantum # 14. page 3 (PDF page 25), 2.1 Normative references, paragraph 1:
In the first sentence "...though reference in the text..." should be changed 

to "...though referenced in the text..."

Quantum # 15. page 4 (PDF page 26), 2.3 References under development, Note 1:
Does Global Engineering have copies of draft standards?  I would have thought 
that folks would be referred to the T10 web site for these.

Quantum # 16. page 8 (PDF page 30), 3.1.68 physical reconnect:
In the second sentence, "A target does a physical reconnect..." should be 
replaced with something like, "A target initiates a physical reconnect..."

Quantum # 17. page 8 (PDF page 30), 3.1.79, SCSI device:
"...connect the drivers..." should be replaced with, "...connect its 

Quantum # 18. page 9 (PDF page 31), 3.2 Symbols and abbreviations,
For QAS I think, "Quick Arbitrate and Selection" should be either, "Quick 
Arbitrate and Select" or, "Quick Arbitration and Selection".

Quantum # 19. page 10 (PDF page 32), 3.3.2 invalid:
I know it's this way in other standards, but I would change the last words 
from, " error." to, "as an error."  (see also page 11 (PDF page 33), 
3.3.8 reserved.)

Quantum # 20. page 10 (PDF page 32), 3.3.5 may not:
"A keyword..." should be changed to "Keywords..." or "A key phrase".

Quantum # 21. pages 17 and 18 (PDF pages 39 and 40), 4.7 Data transfers, 
paragraph 2:
The following sentence is cumbersome, "As a result the REQ(ACK) signals rising

edge to rising edge time varies by 2X depending on the whether ST or DT 
transfers are enabled, however the data's transfer rate remains the same."  I 
think this should be changed to something like, "As a result, the time from 
rising edge to rising edge for REQ and ACK signals for the same transfer rate 
is twice as long for a DT transfer as it is for an ST transfer."

Quantum # 22. page 19 (PDF page 41), 4.8 Data transfer modes, paragraph 2:
The second sentence is in error, "The 8-bit information transfer mode is used 
for all information transfers except DATA phases."  I know this was an 
oversight on the editor's part, but additional information is transferred on 
the upper eight bits if AIP is in effect.  The sentence could be changed to, 
"The 8-bit information transfer mode is used for all information transfers 
except DATA phases (except when an alternate error detection scheme for 
asynchronous information phases is in effect - see Annex M)."

Quantum # 23. page 42 (PDF page 64), 5.4.1 SE assignments, Table 5 - SE 
contact assignments - nonshielded alternative 4 connector, Note 3:
This note says, "The pins identified as being short and long only applies to 
the host connector and not the connector on the SCSI device.  All pins on the 
SCSI device connector are the same length."  I think the concept of "host" and

"device" used here are a carry overs from previous standards.  In figure 15 in

this standard (PDF page 29) the terms "device side" and "cable/backplane side"

are used.  I think these terms are more accurate and the note should use them,

as well.  (see also page 45 (PDF page 67), 5.4.2 Differential assignments, 
Table 8 - LVD/MSE contact assignments - nonshielded alternative 4 connector, 
Note 2:)

Quantum # 24. page 46 (PDF page 68), 6.1 SCSI bus interconnect overview, 
paragraph 3:
"The function of the interconnect is to:" should be changed to, "The functions

of the interconnects are to:"

Quantum # 25. page 50 (PDF page 72), Crosstalk, paragraph 2:
The terms "DATA" and "PARITY" are introduced in this paragraph with no 
previous explanation.  The previous instance of "DATA" used by itself in this 
standard referred to a phase.  I think most of us understand what is meant 
here, but that it would be more consistent (and correct) to say something 
like, "...DB(7-0, P_CRCA) or DB(7-0, P_CRCA, P1), and REQ or ACK pairs."

Quantum # 26. page 59 (PDF page 81), Table 19 - SE input voltage 
"VIL", VIH", "IIL", and "IIH" need to have their second two letters be 
subscripts in three places in this table.

Quantum # 27. beginning on page 64 (PDF page 86), 7.3.2 LVD driver 
characteristics and on several pages following:

I think the terms "+signal" and "-signal", and "+Signal" and "-Signal"
be replaced with the terms "+SIGNAL" and "-SIGNAL" to be consistent.  (See 
also page 80 (PDF page 102), 7.3.2 LVD signals, several places: The document 
should be searched for other occurrences of this.)

Quantum # 28. beginning on page 64 (PDF page 86), 7.3.2 LVD driver 
characteristics and on several pages following:
The terms "source x" are used in several places in these clauses.  This might 
be clearer if they were capitalized to be "SOURCE x".

Quantum # 29. page 68 (PDF page 90), Management of LVD release 
glitches, paragraph 1:
The term "bus settle delay" (that I think should be "Bus Settle Delay") is 
used for the first time in the document without definition or reference.  I 
would add "(see 9)" after the first time the term is used.  Since this is also

true for "system deskew delay" (that I think should be "System Deskew Delay") 
on page 69 (PDF page 91), Table 24 - Glitch management requirements and other 
timings, I would recommend that some words be added early in clause 7 that say

something like, "For specific timing definitions see 9."

Quantum # 30. page 69 (PDF page 91), Table 24 - Glitch management 
requirements..., Note:
The phrase, "BUS FREE phase starts a Bus Settle Delay after..." should be 
changed to "A BUS FREE phase starts one Bus Settle Delay after..."

Quantum # 31. page 75 (PDF page 97), 7.4.2 LVD/MSE multimode transceiver..., 
paragraph 4:
The phrase, "A LVD/MSE multimode SCSI device..." should be replaced by, "An 
LVD/MSE multimode SCSI device..."  Strangely enough, I searched the whole 
document, and this is the only occurrence of this.

Quantum # 32. page 76 (PDF page 98), 7.4.3 Transceiver ground drivers, 
paragraph 1:
The phrase, "...a MSE driver..." should be replaced by, "An MSE driver..."  
Once again I searched the whole document, and this is the only occurrence of 

Quantum # 33. page 78 (PDF page 100), 8.2 Signal descriptions, C/D:
I don't think the second "CONTROL" should be all caps.

Quantum # 34. page 79 (PDF page 101), 8.2 Signal descriptions, P_CRCA (data 
group transfer enabled):
The term "ULP" is used for the first time here without definition.  There are 
several ways to resolve this, but I think the best would be to add, "(ULP)" 
after, "upper level protocol" in 3.1.98.

Quantum # 35. page 98 (PDF page 120), 10 SCSI bus phases, paragraph 1:
It says here, "There are eight distinct phases..."  (and then the distinct 
phases are listed).  However, later in this clause we find the "NORMAL 
ARBITRATION phase", "QAS phase", "DT DATA IN phase", "DT DATA OUT phase", "ST 
DATA IN phase", and "DT DATA OUT" phase.  Are these indistinct phases?  
Somehow I think this needs to be resolved and reconciled in throughout 

Quantum # 36. beginning at page 98 (PDF page 120):
There are many places where the phrase, "...a [xxx time]..." where [xxx time] 
is Bus Clear Delay, Bus Settle Delay, etc.  I think these should be changed to

"…one [xxx time]…" to be more precise.

Quantum # 37. page 98 (PDF page 120), 10.1.1 Unexpected and expected bus free:
I think it is confusing to mix these conditions in a single subclause.  It 
would be easier for me to read and understand if all of the text related to an

expected bus free condition was in 10.1.1 and the text about unexpected bus 
free was in a new subclause, 10.1.2.  This would also, then, make two 
subclauses instead of one nested under 10.1 which is a more common style.

Quantum # 38. page 99 (PDF page 121), 10.2 Arbitration, paragraph 2:
There is a semicolon in the first sentence that I think should be a colon.

Quantum # 39. page 99 (PDF page 121), 10.2.1 NORMAL ARBITRATION phase, first 
Item (c) says, "...the SCSI device shall not arbitrate (i.e. assert the BSY 
signal and its SCSI ID) if more than a bus set delay has passed since the BUS 
FREE phase was last observed."  However, Note 23 immediately following says, 

"There is no maximum delay before asserting the BSY signal and the SCSI ID 
following the bus free delay in step (b)."  I may be missing something, but 
this looks like a contradiction to me.

Quantum # 40. page 108 (PDF page 130), Information unit transfer, 
first item in the third numbered list on the page:
I think that, "Shall after detecting a REQ transition;" should be changed to, 
"Shall wait until detecting a REQ transition;"  (See also the first item in 
the third numbered list on page 110 (PDF page 132)).

Quantum # 41. page 111 (PDF page 131), Data Group Pad field and 
pCRC field transfer to target:
The first sentence of the first paragraph is not a complete sentence (i.e., 
"If the I/O signal is false (transfer to the target) and the initiator 
determines the data field transfer is complete by detecting an assertion of 
the P_CRCA signal.")  I think the "and" in the sentence should be removed.

Quantum # 42. page 143 (PDF page 165), Two-byte messages:
The first sentence of this says, "Two-byte messages consist of two consecutive

bytes transferred during two consecutive MESSAGE IN phases or two consecutive 
MESSAGE OUT phases." but 16.1.1 says, "...a message shall not be split between

multiple MESSAGE phases."  I think the sentence in should be changed 
to, "Two-byte messages consist of two consecutive bytes transferred during a 
MESSAGE IN phase or a MESSAGE OUT phase."  (See also the first sentence in Extended messages.)

Quantum # 43. Annex E:
All of the figures in this Annex (except E.6) are at least somewhat blurry 
(they look even worse in my printed version of the PDF).  Is there anything 
that can be done about this?

Quantum # 44. Annex M, M.1, paragraphs 1 and 2:
Paragraphs 1 and 2 are identical.  Even though I like to see my words in 
print, one of these paragraphs should be deleted.


Comments attached to No ballot from Gene E. Milligan of 
Seagate Technology:

As you probably can see from the introduction below I am prepared to change my

vote to Yes when T10 has approved the resolution of comments. This however is 
not a conditional vote that I think is prohibited.

Chair's Note:  Gene Milligan's comments are contained in T10/99-322r0


Comments attached to YesC ballot from Paul D. Aloisi of 
Unitrode Corporation:

The cable information in section 6 needs to be updated to 99-111r7 and the 
annex e to the latest version of 98-219

******************** End of Ballot Report ********************

John Lohmeyer                  Email: lohmeyer at
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

More information about the T10 mailing list