Draft Minutes, Parallel SCSI WG - March 7, 2000
John Lohmeyer
lohmeyer at t10.org
Mon Mar 13 12:34:46 PST 2000
* From the T10 Reflector (t10 at t10.org), posted by:
* John Lohmeyer <lohmeyer at t10.org>
*
A PDF of these minutes is available at:
ftp://ftp.t10.org/t10/document.00/00-155r0.pdf
Accredited Standards Committee*
National Committee for Information Technology Standards (NCITS)
Doc. No.: T10/00-155r0
Date: March 13, 2000
Reply to: John Lohmeyer
To: T10 Membership
From: John Lohmeyer and Ralph Weber
Subject: Parallel SCSI Working Group Meeting -- March 7, 2000
Dallas, TX
Agenda
1. Opening Remarks
2. Approval of Agenda
3. Attendance and Membership
4. SPI-4 Physical Topics
4.1 SPI-4 clocking proposal (99-262) [Petty]
4.2 Proposal for Fast-160 to be included in SPI-4 (99-295) [Milligan]
4.3 Fast-160 Phase Encoded Data Enabling (99-323) [Moore]
4.4 Driver Precompensation with Receiver Filtering for Ultra4 (00-175)
[Smith]
4.5 ISI Measurements (99-337) [Bridgewater/Bastiani]
4.6 Issues with Implementing Transmitter Pre-Compensation (00-103) [Uber]
4.7 Ultra 320 Summary and Recommendations (00-106) [Evans]
4.8 Use of eye measurements (00-126) [Bastiani]
4.9 Ultra320 SCSI Calibration Protocol (00-133) [Leshay]
4.10 Proposal for training pattern to be included in SPI-4 (00-132)
[Evans]
4.11 Ultra320 SCSI vs. Ultra160 SCSI Eye Diagram Data (00-147) [Brown]
4.12 Requirements for Measuring Receive Signals in SPI-4 and beyond
(00-149) [Ham]
4.13 320 data transfer rates on 25 meter cables (00-153) [McGarrah]
4.14 640 data transfer rates, a first look (00-154) {McGarrah]
4.15 Receiver Input Voltage Budget for Eye Patterns (00-158) [Bridgewater]
4.16 Proposed Training for Skew Compensation (00-174) [Bastiani]
4.17 Proposal for LVD Bus Length (00-170) [Evans]
4.18 Receiver Equalization (00-168) [Brown]
4.19 Pre-Emphasis Experimental Data (99-167) [Gasparik]
4.20 Expander Topics [Ham]
5. SPI-4 Protocol Topics
5.1 SCSI out of band communications method (99-213) [Petty]
5.2 Margin Control (99-264) [Lamers]
5.3 Flow Control & Read Streaming (00-142) [Lamers]
5.4 PPR Message Enhancements (99-283) [Lamers]
6. Domain Validation Technical Report Topics
7. New Business
8. Meeting Schedule
9. Adjournment
Results of Meeting
1. Opening Remarks
John Lohmeyer, the T10 Chair, called the meeting to order at 9:00 a.m.,
Tuesday, March 7, 2000. He thanked Paul Aloisi of Texas Instruments for
hosting the meeting.
As is customary, the people attending introduced themselves and a copy of the
attendance list was circulated.
2. Approval of Agenda
George Penokie led the group in sorting the agenda items based on the
following priorities:
1. Free running clock
2. Training patterns
3. Timings
4. Test configurations
The draft agenda was approved with the following changes:
4.16 Proposed Training for Skew Compensation (00-174) [Bastiani]
4.17 Proposal for LVD Bus Length (00-170) [Evans]
4.18 Receiver Equalization (00-168) [Brown]
4.19 Pre-Emphasis Experimental Data (99-167) [Gasparik]
4.20 Expander Topics [Ham]
During the course of the meeting, the following agenda items were
added/revised:
4.4 Driver Precompensation with Receiver Filtering for Ultra4 (00-175)
[Smith]
For future meetings, the agenda was reorganized as follows:
4. SPI-4 Physical Topics
4.1 Free Running Clock
4.1.1 SPI-4 clocking proposal (99-262) [Petty]
4.1.2 Fast-160 Phase Encoded Data Enabling (99-323) [Moore]
4.2 Training Patterns
4.2.1 Ultra320 SCSI Calibration Protocol (00-133) [Leshay]
4.2.2 Proposal for training pattern to be included in SPI-4 (00-132) [Evans]
4.2.3 Proposed Training for Skew Compensation (00-174) [Bastiani]
4.3 Timings
4.3.1 Proposal for Fast-160 to be included in SPI-4 (99-295) [Milligan]
4.4 Test Configurations
4.4.1 Use of eye measurements (00-126) [Bastiani]
4.4.2 Receiver Equalization (00-168) [Brown]
4.4.3 Requirements for Measuring Receive Signals in SPI-4 and beyond (00-149)
[Ham]
4.4.4 Ultra320 SCSI vs. Ultra160 SCSI Eye Diagram Data (00-169) [Brown]
4.5 Other SPI-4 Physical Topics
4.5.1 320 data transfer rates on 25 meter cables (00-153) [McGarrah]
4.5.2 640 data transfer rates, a first look (00-154) [McGarrah]
4.5.3 Pre-Emphasis Experimental Data (99-167) [Gasparik]
4.5.4 Driver Precompensation with Receiver Filtering for Ultra4 (00-175)
[Smith]
4.5.5 Expander Topics [Ham]
3. Attendance and Membership
Attendance at working group meetings does not count toward minimum attendance
requirements for T10 membership. Working group meetings are open to any
person or organization directly and materially affected by T10's scope of
work. The following people attended the meeting:
Name S Organization Electronic Mail Address
---------------------- -- ------------------------- -------------------------
Mr. Lawrence J. Lamers P Adaptec, Inc. ljlamers at ieee.org
Mr. Vincent Bastiani A# Adaptec, Inc. bastiani at corp.adaptec.com
Mr. Charles Brill P AMP, Inc. cebrill at ix.netcom.com
Mr. Douglas Wagner P Berg Electronics wagnerdl at bergelect.com
Mr. Bill Galloway P BREA Technologies, Inc. billg at breatech.com
Mr. Dennis Lang A Circuit Assembly Corp. dennisl at circuitassembly.
com
Mr. Edward Haske P CMD Technology haske at cmd.com
Mr. Robert C. Elliott P Compaq Computer Corp. Rob_Elliott at compuserve.
com
Dr. William Ham A Compaq Computer Corp. bill.ham at digital.com
Mr. John Tyndall A Crossroads Systems, Inc. jtyndall at crossroads.com
Mr. Rick DeMars V CTS rdemars at ctscorp.com
Mr. Charles Tashbook P Dallas Semiconductor charles.tashbook at dalsemi.
com
Mr. Bill Anderson O DDK Electronics bill_anderson@
ddkconnectors.com
Mr. William Dawkins O Dell Computer bill_dawkins at dell.com
Mr. Ralph O. Weber A ENDL Texas roweber at acm.org
Mr. Eugene Lew P Fujitsu elew at fcpa.fujitsu.com
Mr. Nathan Hastad P General Dynamics nathan.j.hastad at gd-is.com
Mr. Zane Daggett P Hitachi Cable Manchester, zdaggett at hcm.hitachi.com
Inc
Mr. George O. Penokie P IBM Corp. gop at us.ibm.com
Mr. John Lohmeyer P LSI Logic Corp. lohmeyer at t10.org
Mr. Frank Gasparik V LSI Logic Corp. frank.gasparik at lsil.com
Mr. William Petty V LSI Logic Corp. william.petty at lsil.com
Mr. Andrew Brown V LSI Logic Corp. andrew.brown at lsil.com
Mr. Brian Day V LSI Logic Corp. brian.day at lsil.com
Mr. Mark Strauss V Lucent Technologies msstrauss at lucent.com
Mr. W. Hunsicker V Lucent Technologies whunsicker at lucent.com
Mr. Makeshwar V Lucent Technologies makesh at lucent.com
Kothandaraman
Ms. Jie Fan P Madison Cable Corp. jfan at madisoncable.com
Mr. Jay Neer P Molex Inc. jneer at molex.com
Ms. Betty Akeredolu O Molex Inc. bakeredolu at molex.com
Mr. Galen Fromm V Molex Inc. gfromm at molex.com
Mr. Richard Moore A# QLogic Corp. r_moore at qlc.com
Mr. Ting Li Chan V QLogic Corp. t_chan at qlc.com
Mr. Richard L. V QLogic Corp.
Rananiec
Mr. Mark Evans P Quantum Corp. mark.evans at quantum.com
Mr. Russ Brown V Quantum Corp. russ.brown at quantum.com
Mr. Bruce Leshay V Quantum Corp. bleshay at tdh.qntm.com
Mr. Gene Milligan P Seagate Technology Gene_Milligan at notes.
seagate.com
Mr. Gerald Houlder A Seagate Technology gerry_houlder at notes.
seagate.com
Mr. Daniel (Dan) F. O Seagate Technology daniel_f_smith at notes.
Smith seagate.com
Mr. Mayank R. Patel V Seagate Technology mayank_r_patel at notes.
seagate.com
Mr. Erich Oetting P Storage Technology Corp. erich_oetting at stortek.com
Mr. Paul D. Aloisi P Texas Instruments Paul_Aloisi at ti.com
Mr. Donald R. Getty A# Texas Instruments donald_getty at ti.com
Mr. Mike Kosco V Texas Instruments mike at mvbuilders.com
Mr. John Wilson V Texas Instruments jwilson at asic.sc.ti.com
Mr. Kenneth J. Hallam P UNISYS Corporation ken.hallam at unisys.com
47 People Present
Status Key: P - Principal
A,A# - Alternate
O - Observer
L - Liaison
V - Visitor
4. SPI-4 Physical Topics
4.1 SPI-4 clocking proposal (99-262) [Petty]
Bill Petty reviewed the proposal to run P1 as an ST type signal (99-262r0).
The P1 is not trained or deskewed and has the same timing budget as SPI-3.
Bruce Leshay questioned how an expander could detect the beginning of
training. Bill noted that the choices of training methods are strongly tied
to the clocking mechanism.
4.2 Proposal for Fast-160 to be included in SPI-4 (99-295) [Milligan]
Gene Milligan reviewed the draft in progress of 99-295 (r2 is completed and r3
is in progress). He noted that agreement had been reached with Bill Petty on
the error tolerances timing values for Fast-160. The group discussed the
table of error tolerance values in detail and agreed on some corrections to
its contents.
The group discussed the degree to which data can be transmitted at rates below
the negotiated rate. The problem is that the current and previous versions of
SCSI use the negotiated data transmission rate as a maximum -- transmission at
slower rates are perfectly legal. However, transmission at rates other than
negotiated rate cause problems for the timing budget compensation functions
(skew compensation, adaptive equalization, etc.) that rely on training
operations because the training results are only good for the transfer rate at
which the training was performed. Everybody had been assuming that transfers
would occur at the negotiated rate, which would also be the rate at which the
training was performed. The potential for slower transfer rates raised
questions about such assumptions. One agreement was that, if slower transfer
rates are used, the training must be done at the slower rate and redone
anytime the rate is changed.
4.3 Fast-160 Phase Encoded Data Enabling (99-323) [Moore]
Richard Moore asked that this item be removed from future agendas.
4.4 Driver Precompensation with Receiver Filtering for Ultra4 (00-175)
[Smith]
Dan Smith presented data and computed filtering results for a configuration
where precompensation is combined with receiver filtering (which is different
|from receiver equalization). Many questions were raised about the
appropriateness of the Chebyshev filter used to generate the presented data.
Dan agreed that a better filter than the one selected could be used, and
stated that he would work on that issue in preparation for the next meeting.
Bill Ham wondered if precompensation would be necessary if a good enough
filter were used, and explained how Dan's data seemed to demonstrate this
point.
4.5 ISI Measurements (99-337) [Bridgewater/Bastiani]
Vince Bastiani asked that discussion of this topic be deferred to the next
meeting.
4.6 Issues with Implementing Transmitter Pre-Compensation (00-103) [Uber]
Mark Evans asked that this item be removed from future agendas.
4.7 Ultra 320 Summary and Recommendations (00-106) [Evans]
Mark Evans requested that this item be removed from future agendas
4.8 Use of eye measurements (00-126) [Bastiani]
Vince Bastiani asked that discussion of this topic be deferred to the next
meeting.
4.9 Ultra320 SCSI Calibration Protocol (00-133) [Leshay]
Mark Evans presented a list of options for the frequency of training. The
group discussed the effects of the options on expanders and performance. The
length of time required to do training weighed on the choices and was also
discussed.
The three choices as presented for straw poll voting were:
o remember (i.e., train once for data in and once for data out after each
PPR negotiation),
o train before each data phase,
o train for I_T nexus once for data in and once for data out.
On a straw poll vote of 18:2, the group decided not to include training at the
beginning of each phase.
The group then unanimously agreed to select only one of the remaining two
choices. The straw poll was 13 for remember and 7 for train for I_T nexus?
George stated that the vote was not conclusive and recommended that both
choices remain in the proposal.
Bruce Leshay presented a detailed proposal for handling the transmission of a
training pattern. One key feature of the proposal was the use of the SEL
signal to identify a training a sequence.
George Penokie asked Bill Petty if he wanted to continue with his competing
training pattern proposal. Bill's concern was that Bruce's proposal had the
deskew training running at full speed. Bill Petty discussed this issue with
Bruce. Bruce agreed that the fact that SCSI allows transmitting at rates
slower than the negotiated rate had not been accounted for in the proposal
(see also item 4.2).
Another problem with training patterns was the time they consume. Some argued
that the `training time' would be so short as to be unnoticeable. George
Penokie was very concerned about the overall performance impact of training
times. This issue caused difficulties because being able to use longer
training patterns (with longer training times) would solve some problems. To
assist in the decision making, George Penokie accepted an action item to build
a model that shows the performance impact of the training times on packetized
performance.
4.10 Proposal for training pattern to be included in SPI-4 (00-132) [Evans]
The group agreed to review 00-132, which casts 00-133 (discussed in item 4.9)
into standards wording, outside the meeting.
4.11 Ultra320 SCSI vs. Ultra160 SCSI Eye Diagram Data (00-169) [Brown]
Russ Brown presented data on receiver equalization.
4.12 Requirements for Measuring Receive Signals in SPI-4 and beyond (00-149)
[Ham]
Bill Ham presented a description showing the need for SPI-4 to include a
transformation function that describes the effects of a receiver equalization
or filtering mechanism so that signals at the receiver connector can be
transformed to the signals processed by the receiver. He reminded the group
that SCSI is obliged to specify signals at the connector, but the quality of
the signals processed by the receiver is what really matters.
Bill Galloway noted that specifying a better filter loosens requirements on
the other components of the system, while, specifying a 'sloppy' filter puts
more work on the other system components (e.g., cables, connectors, and
transmitters).
4.13 320 data transfer rates on 25 meter cables (00-153) [McGarrah]
Russ Brown presented data showing the value of receiver equalization on 25 m
cables operating a 320 MB/sec transfer rates, using experimental data and the
computed effects of equalization.
4.14 640 data transfer rates, a first look (00-154) {McGarrah]
Russ Brown presented data showing the value of receiver equalization on 25 m
cables operating a 640 MB/sec transfer rates, using experimental data and the
computed effects of equalization.
4.15 Receiver Input Voltage Budget for Eye Patterns (00-158) [Bridgewater]
Vince Bastiani asked that discussion of this topic be deferred to the next
meeting.
4.16 Proposed Training for Skew Compensation (00-174) [Bastiani]
Vince Bastiani asked that discussion of this topic be deferred to the next
meeting.
4.17 Proposal for LVD Bus Length (00-170) [Evans]
Mark Evans presented a proposal that Fast-160 allow 25 m cables for
point-to-point applications and 12 m cables for multi-drop applications
(00-170r0). He noted that items 4.13 and 4.14 present data to justify the
proposal. Mark noted that this proposal uses the same cable lengths as
Fast-80.
Bill Ham commented that the cable working group may request changes in this
proposal as a result of the work in cable performance specifications and in
cable testing.
Bill also requested that note 1 be removed from the table and Mark agreed.
In the absence of any objections, the group unanimously recommended that
00-170r1 (r0 as modified) be approved for inclusion in SPI-4. George noted
that this is the first proposal to be approved for inclusion in SPI-4.
4.18 Receiver Equalization (00-168) [Brown]
Russ Brown presented key concepts and properties of a receiver equalization
method.
4.19 Pre-Emphasis Experimental Data (99-167) [Gasparik]
Frank Gasparik presented experimental data showing the helpful effects of
transmitter precompensation.
4.20 Expander Topics [Ham]
Bill Ham presented a list of expander topics that he believes must be added to
SPI-4:
o definition of simple expander port requirements (may use content from
EPI)
o inclusion of domain validation level 1 required for all initiator ports
o inclusion of minimal target requirements to support the first two
5. SPI-4 Protocol Topics
5.1 SCSI out of band communications method (99-213) [Petty]
John Lohmeyer noted that Bill has presented the proposal a couple of times and
has received no feedback. He stated that the proposal has been tested in the
lab. Bill Ham asked that `out-of-band' be dropped from the topic title
because it caused people to misinterpret the intent of the proposal.
In the absence of Bill Petty, John led the group in a discussion of the key
benefits of having such a protocol (i.e., that no SCSI IDs are required to
communicate with devices such as simple expanders, terminators, etc. Larry
Lamers announced that SDV (Domain Validation Technical Report) contains
expander communication. The group agreed to put this item under the Domain
Validation topic for future meetings.
5.2 Margin Control (99-264) [Lamers]
George Penokie stated a concern about use of messages to do margin control.
Larry Lamers asked that any interested parties send comments on this proposal
and that this topic be deferred to the next meeting.
5.3 Flow Control & Read Streaming (00-142) [Lamers]
Larry Lamers asked that any interested parties send comments on this proposal
and that this topic be deferred to the next meeting.
5.4 PPR Message Enhancements (99-283) [Lamers]
Larry Lamers asked that any interested parties send comments on this proposal
and that this topic be deferred to the next meeting.
6. Domain Validation Technical Report Topics
Larry Lamers asked that this topic be changed to "Expanders and Domain
Validation".
John Lohmeyer reported that SDV revision 0 was posted on the T10 web site as
of this morning.
7. New Business
Bill Ham announced the SCSI Physical Test group meeting to be held the next
morning (Wednesday, 3/8) and gave a brief overview of the expected content of
that meeting.
8. Meeting Schedule
The next meeting of the Parallel SCSI Working Group will be Monday, March 27,
2000 commencing at 1:00 p.m. until 6 p.m. and continuing Tuesday, March 28,
2000 at 9:00 a.m. until 5:00 p.m. at the Embassy Suites (408-942-0400) in
Milpitas, CA hosted by Adaptec.
The subsequent meetings of this group are April 26 starting at 1:00 p.m. and
continuing to April 27 in Colorado Springs, CO and May 16 starting at 9 a.m.
in Nashua, NH.
9. Adjournment
The meeting was adjourned at 5:45 p.m. on Tuesday March 7, 2000.
--
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