Minutes of X3T10 SCSI Working Group 5/18/94
John Lohmeyer
jlohmeye at ncr-mpd.FtCollinsCO.NCR.COM
Tue May 24 12:51:15 PDT 1994
Accredited Standards Committee
X3, Information Processing Systems
Doc. No.: X3T10/94-107r0
Date: May 24, 1994
Project:
Ref. Doc.:
Reply to: J. Lohmeyer
To: Membership of X3T10
From: Weber/Lohmeyer
Subject: Minutes of X3T10 SCSI Working Group May 18, 1994
-----------------------------------------------------------------------------
Agenda
1. Opening Remarks
2. Attendance and Membership
3. Approval of Agenda
4. Physical Topics
4.1 Fast-20 Progress Report (94-98, -061r4) [Ham/Lohmeyer]
4.2 SPI Rev 12c Review [Lamers]
4.3 Proposed changes to SPI (94-038) [Aloisi]
5. Protocol Topics
5.1 SAM Forwarding Comments (Project 994D) [Monia]
5.2 Optional and Mandatory in SAM (94-081) [McGrath]
5.3 SBP Forwarding Comments (Project 992D) [Lamers/Roberts]
5.4 FCP Comments Resolution (Project 993D) [Snively]
5.5 Message Handling Chart for SIP (94-032) [Houlder]
6. Command Set Topics
6.1 SCSI-3 Download Microcode (94-80r1, 94-104r0) [McGrath, Cummings]
6.2 SCSI-3 Inquiry Command (94-079r1) [McGrath]
6.3 Data Recovery on Deferred Errors (94-067) [Houlder]
6.4 RAID 5 Support on SCSI Disk Drives (94-111) [Houlder]
6.5 Proposal on the READ POSITION command (93-187R0) [Lappin]
6.6 SPC Topics [Weber]
6.7 Placing the Processor Device Commands in SPC (94-018) [Weber]
7. Miscellaneous Topics
7.1 AIIM C21 Liaison [Podio]
8. Meeting Schedule
9. Adjournment
Results of Meeting
1. Opening Remarks
John Lohmeyer the Chair, called the meeting to order at 9:10 a.m., Wednesday,
May 18, 1994. He thanked Chuck Brill of AMP for hosting the meeting.
As is customary, the people attending introduced themselves. A copy of the
attendance list was circulated for attendance and corrections.
It was stated that the meeting had been authorized by X3T10 and would be
conducted under the X3 rules. Ad hoc meetings take no final actions, but
prepare recommendations for approval by the X3T10 task group. The voting
rules for the meeting are those of the parent committee, X3T10. These rules
are: one vote per company; and any participating company member may vote.
The minutes of this meeting will be posted to the SCSI BBS and the SCSI
Reflector and will be included in the next committee mailing.
2. Attendance and Membership
Attendance at working group meetings does not count toward minimum attendance
requirements for X3T10 membership. Working group meetings are open to any
person or company to attend and to express their opinion on the subjects
being discussed.
The following people attended the meeting:
Name S Organization Electronic Mail Address
---------------------- -- ------------------------- -------------------------
Mr. Norm Harris P Adaptec, Inc. nharris at adaptec.com
Mr. Lawrence J. Lamers A# Adaptec, Inc. ljlamers at adaptec.com
Mr. Fernando L. Podio L AIIM C21 Chair fernando at pegasus.nist.ncs
l.gov
Mr. Neil T. Wanamaker P Amdahl Corp. ntw20 at eng.amdahl.com
Mr. Michael Wingard O Amphenol Interconnect
Mr. Ken Scherzinger V Amphenol Spectra Strip
Mr. Jerry Fredin V AT&T Global Info. Jerry.Fredin at WichitaKS.NC
Solutions R.COM
Mr. John Lohmeyer P AT&T/ NCR john.lohmeyer at ftcollinsco
Microelectronics .ncr.com
Mr. Joe Stoupa O Burr-Brown Corp.
Mr. Bob Gannon O C&M Corp.
Mr. Ian Morrell A Circuit Assembly Corp.
Mr. Joe Chen P Cirrus Logic Inc. chen at cirrus.com
Mr. Peter Johansson O Congruent Software, Inc.
Mr. Michael Bryan P Conner Peripherals Mike_Bryan+aDevelopment+a
Colorado%conner at mcimail.c
om
Mr. Louis Grantham P Dallas Semiconductor grantham at dalsemi.com
Mr. Charles Monia P Digital Equipment Corp. monia at starch.enet.dec.com
Dr. William Ham A# Digital Equipment Corp. ham at subsys.enet.dec.com
Mr. William Dallas A# Digital Equipment Corp. dallas at wasted.enet.dec.co
m
Mr. Ralph Weber A# Digital Equipment Corp. weber at star.enet.dec.com
Mr. Edward A. Gardner A Digital Equipment Corp. gardner at ssag.enet.dec.com
Mr. Kenneth J. Hallam A ENDL khallam at endlas.com
Mr. Edward Lappin P Exabyte Corp. tedl at exabyte.com
Mr. Gary R. Stephens P FSI Consulting Services 6363897 at mcimail.com
Mr. Robert Liu P Fujitsu Computer 74503.1610 at compuserve.com
Products,Am
Mr. Jeffrey L. P Hewlett Packard Co. jlw at hpdmd48.boi.hp.com
Williams
Mr. Paul Boulay A# Hitachi Computer Products p_boulay at hitachi.com
Ms. Nancy Cheng O Hitachi Computer Products n_cheng at hitachi.com
Dr. Sam Karunanithi P Hitachi Micro Systems, skarunan at hmsi.hitachi.com
Inc.
Mr. George Penokie P IBM Corp. gop at rchvmp3.vnet.ibm.com
Mr. Ken Cummings O IBM Corp. kcummings at vnet.ibm.com
Mr. Giles Frazier O IBM Corp. gfrazier at ausvm6.vnet.ibm.
com
Dr. Gerald Marazas A# IBM PC Company marazas at bcrvmpc2.vnet.ibm
.com
Mr. Dean Wallace P Linfinity Micro
Mr. Chuck Grant A Madison Cable Corp.
Mr. Robert J. O NCCOSC RDTE DIV 821 gallenbe at nosc.mil
Gallenberger
Mr. Clifford Carlson O NCCOSC RDTE DIV 821 wcarlson at nosc.mil
Mr. Skip Jones P QLogic Corp. sk_jones at qlc.com
Mr. Gerald Houlder A Seagate Technology Gerry_Houlder at notes.seaga
te.com
Mr. David Deming O Solution Technology
Mr. Erich Oetting P Storage Technology Corp. Erich_Oetting at Stortek.com
Mr. Robert N. Snively P Sun Microsystems, Inc. Bob.Snively at sun.com
Mr. Vit Novak A Sun Microsystems, Inc. vit.novak at sun.com
Mr. Paul D. Aloisi P Unitrode Integrated Aloisi at uicc.com
Circuits
Mr. Gerald Root V Voltex Interconnect
Systems
Mr. Tak Asami A Western Digital asami at dt.wdc.com
Corporation
45 People Present
Status Key: P - Principal
A, A# - Alternate
O - Observer
L - Liaison
S,V - Visitor
3. Approval of Agenda
The proposed agenda was approved.
4. Physical Topics
4.1 Fast-20 Progress Report (94-98, -061r4) [Ham/Lohmeyer]
Bill Ham reported on the results of a meeting in Milpitas. He indicated that
the group wanted to get the fast SCSI cable length up to 3 meters. He
reported that 16 devices represent too large a load on a 1.5 meter single
ended bus. 8 devices will work on a 1.5 meter bus, but the receiver
specifications will need some adjustment. Stub length plays only a minor
role. Bill suggested that the 3:1 stub length be changed to a 2:1. Also the
discussion suggested that the stub length be changed to a recommendation
(from a requirement).
Bill showed a series of single-ended waveform slides. Many of the waveforms
showed a pronounced shelf at the receiver. Bill was asked about the cause
of the shelf. He attributed it to impedance effects of the device loads.
Bill discussed the tradeoffs between cable length and receiver thresholds.
The data Bill presented led him to the following proposals regarding the
receiver threshold. On the assertion edge, 'may detect' changes to 1.4 volts
and 'must detect' changes to 1.1 volts. On the negation edge, 'may detect'
changes to 1.6 volts and 'must detect' changes to 1.9 volts. The hysteresis
should change to 200 millivolts minimum.
John pointed out that the chip folks were reluctant to change the threshold
values at the previous meeting. John also noted that the chip folks did not
have the data Bill just finished showing to this meeting. Most of the chip
folks were not present to comment on Bill's presentation.
Several other options were discussed. Bill was concerned about the closest
possible compatibility with existing hardware. Different slew rates and
different pin capacitance for device chips were discussed. Since there are
several options for compatible changes, the discussion was lengthy.
Noting that chip changes had not been anticipated for this meeting, John
recommended that a decision wait for the June 2nd Fast-20 Study Group
meeting. Finally, Bill put in a plug for the very clean fast performance of
differential buses.
Bill turned to the name for the project. Bill noted that using megabits per
sec as a naming base produces identical names for different configurations.
John expressed a concern over an apparently continuous series of name
changes. In the end, the group elected to stay with "Fast-20".
Vit Novak presented a proposal for an annex in the Fast-20 proposal (94-
103r0). The proposed annex would contain informative data regarding Fast
SCSI Node Capacitance. The May 5th Fast-20 Study Group had agreed to specify
25 pF node capacitance provided a note or annex were included recommending
that lower capacitance parts be used in maximum configuration systems.
Vit's proposal reopened the capacitance debate because it contained a node
capacitance budget that totaled 20 pF. Jim McGrath wanted to stay with 25
pF because his market rarely includes more than a couple devices per bus and
there is a cost penalty for lower capacitance devices. Ed Gardner and Larry
Lamers proposed 15 pF devices. John suggested deferring the food fight to
the June 2nd meeting.
John suggested adding a recommendation that synchronous transfer parameters
be negotiated lower if fast parameters produce errors. This produced a
discussion of how initiators know what to do and how to get this information
to host software developers.
Norm Harris and Paul Aloisi discussed maximum voltage levels in an active
negation. Norm proposed reducing the maximum from 3.7 volts to 3.5 or 3.24
volts. John recalled that the 3.7 volt number had been a "top-of-the-head"
suggestion from Mark Knecht and probably needed refinement. This topic was
also deferred to the June 2nd meeting.
The group discussed specific wording changes in 94-061r4. Larry Lamers plans
to issue a rev 5 for the next mailing.
The location of terminators (in the device or not) was discussed at length.
A straw vote slightly favored (10 for, 8 opposed) striking the sentence "SCSI
devices shall not include termination."
Stub length was discussed as follows. Changing from 1:3 to 1:2 and from a
requirement to a suggestion were the issues. The ensuing straw poll produced
the following result
1) 5 - leave it alone
2) 7 - shall to should
3) 5 - shall to should & 1:3 to 1:2
After that, the consensus favored changing from shall to should.
4.2 SPI Rev 12c Review [Lamers]
Larry reported that in SPI 12c he included SCAM (93-109r5) into the document
as a normative annex. That caused changes in the lettering of the annexes.
Larry also added information accepted at earlier meetings; for example, Gene
Milligan's comments on connectors, ISO editor's comments, and "release"
instead of "negate."
Larry made some changes to make SPI more consistent with SIP. Also, Larry
expressed discomfort with some terminology aspects of the SCAM annex. John
noted that Ed Gardner's SCAM clarifications from November (93-173) needs to
be included in the SCAM annex.
Bill Ham raised concerns about the differential capacitance data in the
informative annex. Larry suggested making a public review comment on the
matter.
Larry noted that the revised SPI will need a second public review.
4.3 Proposed changes to SPI (94-038) [Aloisi]
Paul Aloisi presented two options for handling hot-plugging devices with
TermPwr. The options were not supplying TermPwr to hot-pluggable devices or
limiting TermPwr bypass capacitors to 10 microfarads for hot-plugging
devices. Bob Snively recommended adding the following: "Power cycling
includes TermPwr, including terminators." Paul will bring a specific
proposal to the Plenary.
Paul also raised a concern about the Voh definitions in SPI. This issue was
brought up at the Fast-20 study group meeting, but it really relates to
making SCSI compatible with 3.3 volt devices. Paul also raised concerns
about SCSI termination in a 3.3 volt environment. Paul expressed concern
about SCSI with PCMCIA (where 3.3 volt is about to become the only legal
voltage).
5. Protocol Topics
5.1 SAM Forwarding Comments (Project 994D) [Monia]
In Charles' review SAM comments, the difference between HEAD OF QUEUE and
HEAD OF TASK SET was the first discussion topic. HEAD OF QUEUE is the only
usage of the word queue. However, HEAD OF QUEUE is a proper name. So, there
is no need for defining "queue" as a term.
The definition and usage of the word "response" was discussed next. Many
present brought similar usages of "response" and "confirmation" from other
communications definitions. Work progressed through the Hewlett Packard
comments. Significant discussion was produced by the definition of "task
ended." The events following an abort task operation greatly complicated
reaching a consensus definition of the end of a task.
After much discussion, Charles agreed to define task management functions
as being confirmed but not necessarily with the benefit of a response. In
this context, definitions for "confirmed" and "response" must be taken from
figure 18 in revision 13 of the SAM.
The requirements for targets supporting the ACA bit in the CDB surfaced for
discussion based on Charles' desire for a recorded opinion from the group.
Most of the speakers said that ACA=1 must be supported and ACA=0 can
optionally be unsupported. George Penokie noted that the ACA bit is
overloaded because it is thought to control a compatibility operation and
to control automatic clearing the ACA condition. This produced a lengthy
discussion of whether a second control bit is needed.
The discussion was inconclusive in several different ways. Charles requested
a Plenary vote on whether ACA=1 must be supported by SCSI-3 devices.
George Penokie presented a discussion regarding the ACA bit in the CDB. The
presentation began with a quiz about the SCSI-2. The question was "According
to the SCSI-2 standard, what action is the target supposed to take if it
received a queued command during a contingent allegiance condition?" George
proposed at least a half dozen actions, many of which are done and several
of which are of debatable validity.
George's point was that having the ACA bit set to zero does not provide a
defined basis for specific requirements text in the SAM standard. Referenc-
ing the SCSI-2 standard is the only operable definition for the SAM. The
Hewlett Packard comment #50 brought this discussion to the floor. After some
discussion, Charles agreed to make SAM state that ACA=0 means operate like
SCSI-2 (without any attempt to translate or codify SCSI-2 operation in such
cases).
Most of the remaining HP comments were resolved without significant
discussion. Next, Charles turned to resolving the IBM comments. Most of the
IBM comments were resolved with minimal discussion. Then, the group began
discussing the usefulness of the state definitions and tables as a definition
of the queuing model behavior.
Jim McGrath complained the state-based model cannot easily be translated into
requirements on a target implementation. Jim wished that the translation be
performed as part of the SAM document, not left as an exercise for the
reader. Comments to the effect of, "I cannot understand this," were made
frequently (relative to each of the queuing models). Charles said that the
queuing model (X3T9.2/92-141r8) cannot be inserted directly in SAM because
it defines observable states on the parallel bus. SAM must also work for the
serial interfaces.
No consensus was reached when the discussion was closed. Jim McGrath summar-
ized the areas of disagreement as revolving around suspended data, which he
always felt was a bad idea. Jeff Williams summarized the problems as being
the complexity and testability of the overall queuing model.
5.2 Optional and Mandatory in SAM (94-081) [McGrath]
The group agreed that there are several different types of requirements
expressed in the SAM: mandatory or optional for the protocols, mandatory or
options for targets, and the matrix of combinations of these. Jim has
proposed some terms that work toward explicitly naming each (or most) of the
requirements types. Charles was generally in agreement with the group.
However, the previous issue of Fibre Channel host adapters not supporting
linked commands (in Charles' mind) still bothered Charles.
The need for a upper-layer protocol service to return supported options
information was discussed. After the usual amount of vocal disagreement,
Charles agreed to put words suggesting that hosts provided methods for
determining what options are supported.
5.3 SBP Forwarding Comments (Project 992D) [Lamers/Roberts]
Responses with comments on the SBP (revision 16) letter ballot have been
received from:
o Giles Frazier, IBM Corporation
o Charles Monia, DEC
o Scott Smyers, Apple Computer
o Jim McGrath, Quantum Computer
Gerald Marazas has discussed these comments with each of the originating
parties above, and secured tentative resolution, mutually agreeable, on each
comment item. There remain no known open (unresolved) technical issues being
raised against SBP. The resolution is listed as tentative pending
examination next week by these parties of the specific text changes relative
to the SBP, revision 16, document.
The comments cited above will be documented along with proposed disposition
in a SBP resolution document. Additionally, revision 17 of SBP will be
produced consistent with the proposed disposition. Both the resolution
document and the SBP revision 17 document will be made available to the X3T10
secretary in time for the June mailing deadline. Gerald Marazas will request
a meeting vote at the July X3T10 meeting for acceptance of both the proposed
disposition and the revision 17 of SBP.
5.4 FCP Comments Resolution (Project 993D) [Snively]
Bob Snively presented FCP revision 008a to the group. Bob also distributed
94-109r0, the document listing the comments resolution for all review
comments on FCP 008.
5.5 Message Handling Chart for SIP (94-032) [Houlder]
Gerry Holder led a discussion of his message handling chart. Questions were
raised about the single-choice nature of the document. Jeff Williams noted
that the standard allows for several message/circumstances implementations.
The document, as last revised, seemed to be ready for a Plenary vote.
6. Command Set Topics
6.1 SCSI-3 Download Microcode (94-80r1, 94-104r0) [McGrath, Cummings]
Jim McGrath described the two proposals for using WRITE BUFFER to download
microcode. Jim noted that Ken Cumming's proposal uses new WRITE BUFFER mode
values to uniquely identify the new function. On the negative side, Jim
noted that Ken's proposal only supports download microcode and save. Jim
would like to add a download (and not save) mode.
The mechanics of downloading microcode were discussed. Jim and Ken wanted
to verify the completeness of the proposal, by virtue of holding a large
public discussion. Ken took the action item to make a revised 94-104r0.
John noted that final action on Ken's proposal will be placed on the July
Plenary agenda.
6.2 SCSI-3 Inquiry Command (94-079r1) [McGrath]
Jim proposed defining vital product data (VPD) pages for each standard
document, in which details of supported features can be returned. Jeff
Williams supported the need for attributes reporting, but said that the
proposal is overly complex and detailed.
After a lengthy discussion of the details, Jim listed the options as:
1) design an extension of the INQUIRY data that reports supported options
(like but not necessarily the same as 94-079r1), 2) choose to freeze options
at their current definitions, 3) hack and whack on the current INQUIRY data.
Jim wanted to hold a plenary vote to select from among the choices.
6.3 Data Recovery on Deferred Errors (94-067) [Houlder]
Gerry Houlder presented the customer requirement regarding cached data that
compelled him to propose the caching feature described in 94-067r1. Ed
Gardner asked if the proposal for a mandatory (or optional) write caching
feature. Giles was concerned about non-sequential disk write processing in
the target, which prevent returning unwritten information in a simple
LBA/size pair. Gerry received lots of advice on implementation details.
When asked for his perception of the resolution of the proposal, Gerry said
that he will proceed toward getting the document approved for inclusion in
the direct-access device commands document.
6.4 RAID 5 Support on SCSI Disk Drives (94-111) [Houlder]
This item was deferred to the July working group. Gerry suggested scheduling
an ad hoc meeting in June.
6.5 Proposal on the READ POSITION command (93-187R0) [Lappin]
Edward Lappin presented a proposal for increasing the amount of position data
reported by the READ POSITION command. When activated, Ted's version of the
READ POSITION command would report file number and set number in addition to
block number. Bill Dallas started a chorus of concern about the four-byte
size of the block number field. So, Ted was pressed to increase all
parameter data field sizes to eight-bytes. There were several other detailed
comments. Ted agreed to revised the proposal and return to the working
group.
6.6 SPC Topics [Weber]
Owing to the late hour, Ralph simply waved a couple of foils at the overhead,
mumbled something about wanting time on the plenary agenda, and called for
adjournment.
6.7 Placing the Processor Device Commands in SPC (94-018) [Weber]
Ralph noted that this is a proposal made at the January plenary meeting and
that a plenary vote would be appropriate this month. There was some
discussion about how much of the processor device model belongs in SPC.
However, not enough people were participating to make the discussion
worthwhile.
7. Miscellaneous Topics
7.1 AIIM C21 Liaison [Podio]
Fernando Podio presented ANSI/AIIM MS59-199x, a Media Error Monitoring and
Reporting Technique for Verification of Information Stored on Optical Digital
Data Disk. Fernando noted that the AIIM committee has made some changes
based on comments from the March X3T10 meeting. Particularly, Fernando
described the use of log pages for delivery of media condition data. The
working group pointed Fernando toward using existing mode pages for bad
sector recovery and toward defining log pages for early warning functions.
Fernando was advised to prepare a changes proposal for mode and log pages
that will be applied to the SBC. Fernando sought advice about the specific
details of defining for defining log page parameters.
8. Meeting Schedule
The next working group meetings will be July 18-22 1994 at the Sheraton Tara
Wayfarer Inn (603-622-3766) in Bedford, NH hosted by Digital. The room rates
are $93.00 plus tax. The reservation deadline for these rates is June 29,
1994. The group name is ANSI or X3T10. The host contact is Ralph Weber at
TEL: 603-881-1497, FAX: 603-881-0189, (Weber at star.enet.dec.com).
9. Adjournment
The meeting was adjourned at 8:45 p.m. on Wednesday May 18, 1994.
--
John Lohmeyer E-Mail: John.Lohmeyer at FtCollinsCO.NCR.COM
NCR Microelectronics Voice: 719-573-3362
1635 Aeroplaza Dr. Fax: 719-597-8225
Colo Spgs, CO 80916 SCSI BBS: 719-574-0424 300--14400 baud
More information about the T10
mailing list