X3T9.2/87-107 Date: July 16, 1987 To: X3T9.2 Membership From: John Lohmeyer, Chair X3T9.2 and Larry Lamers, General Technical Editor SCSI-2 Subject: SCSI-2 Working Group Minutes - July 13-15, 1987 An SCSI-2 working group meeting was held July 13-15 in Irvine, CA. We would like to thank Jeff Stai of Western Digital for hosting this meeting. The excellent facilities and the PC that he provided for use at the meeting (and at the following editing session) were much appreciated. It was announced that X3T9.2 had passed a motion at the last plenary meeting setting the goal of completing technical work on SCSI-2 by October with the intention of forwarding SCSI-2 at the February plenary meeting. This will require setting priorities on what is included in SCSI-2 and what is held over for SCSI-3. The following people attended: Name Status Organization ------------------------------ ------ ------------------------------ Mr. Robert N. Snively P Adaptec, Inc. Mr. Robert R. Williamsen A Adaptec, Inc. Mr. James J. Semenak P AT&T Information Systems Mr. Fernando Ramirez V Burndy Corp. Mr. Gene Milligan P CDC/MPI Mr. Gerald Houlder A CDC/MPI Mr. Dan Davies P Cipher Data Products, Inc. Ms. Donna Pope P Cipher/Optimem Mr. Peter Johansson P Congruent Software, Inc. Mr. Steve Goldman P DPT Mr. Paul Nitza P Emulex Corp. Ms. Kim V. Le A Emulex Corp. Mr. Randy Weber S Emulex Corp. Mr. I. Dal Allan P ENDL Mr. Robert Liu P Fujitsu America, Inc. Ms. Elen Hunt A Hewlett Packard Co. Mr. Dave McIntyre A Hewlett Packard Co. Mr. Gary R. Stephens A IBM Corp. Mr. Gerald Marazas A IBM Corp. Mr. William Homans P LMSI-CPI Mr. Paul Boulay O LMSI-OSI Mr. Skip Jones P Maxtor Corp. Mr. Paul Wassenberg P Micropolis Corp. Mr. Jim Patton N Micropolis Corp. Mr. Thomas Gettys A Miniscribe Corp. Mr. Daniel E. Moczarny P Mitsubishi Electronics Amer Mr. John Lohmeyer P NCR Corp. Mr. James McGrath P Quantum Corp. Name Status Organization ------------------------------ ------ ------------------------------ Mr. Ron Schlitzkus O Scientific Micro Systems Mr. Mike Eneboe S Scientific Micro Systems Mr. Lawrence J. Lamers A Sony Corp. of America Mr. Dennis Appleyard P Storage Technology Corp. Mr. D. W. Spence A Texas Instruments Mr. Ralph Schultz P UNISYS Mr. Peter Dougherty O UNISYS Mr. Ken Post P Western Digital Mr. Jeff Stai A Western Digital Status Key: P - Principal A - Alternate L - Liaison O - Observer S - Special Interest Group V - Visitor N - No Mail Alternate The following new documents were distributed at the meeting: Document Doc Date Author Description of Document ------------- -------- --------------- ---------------------------------------- X3T9.2/87-7 6/26/87 D. W. Spence Revised proposal for downloading Rev 3 microcode. X3T9.2/87-99 6/30/87 R. Schultz Terminator Power X3T9.2/87-100 7/1/87 R. Schultz Media Changer Issue X3T9.2/87-101 7/1/87 D. Skinner Tape buffer control (size and EOT handling) X3T9.2/87-102 7/1/87 D. Skinner Proposal on discarding VOLUME OVERFLOW data X3T9.2/87-103 7/8/87 T. Johnson NCR Printer Proposal for SCSI-2 X3T9.2/87-104 7/10/87 J. Lohmeyer SCSI-2 Working Group Notice -- Aug 31 - Sept 2 X3T9.2/87-106 6/30/87 L. Lamers Proposal for CD-ROM in SCSI-2 X3T9.2/87-108 7/10/87 G. Stephens Buffer Device Recovery after CHECK CONDITION X3T9.2/87-109 7/9/87 B. Spence Additional Concerns RE IMMEDIATE-BIT Operations X3T9.2/87-110 7/7/87 D. McIntyre SCSI-2 Group 1, 2, and 5 Command Descriptor Blocks X3T9.2/87-111 7/13/87 J. Stai WRITE AND VERIFY command new wording X3T9.2/87-112 7/14/87 J. Stai FORMAT UNIT command new wording X3T9.2/87-113 6/10/87 R. Weber Sequential-Access Device Recoverable Error Counters X3T9.2/87-114 7/10/87 R. Weber Corrections to Additional Sense Codes for Sequential-Access Devices X3T9.2/87-115 7/10/87 M. Eneboe Flexible Disk Drive I/F, Pin 1 Definition - MODE SENSE/SELECT, Page 5 X3T9.2/87-116 7/9/87 D. Anderson Passing Media Initialization Pattern during FORMAT UNIT command X3T9.2/87-117 7/7/87 P. Johansson Proposed Addition to the INQUIRY Command X3T9.2/87-118 7/13/87 J. Schuessler Wide SCSI Cabling Proposal X3T9.2/87-119 7/9/87 B. Raudebaugh Additions to the Automatic Medium Changer proposal X3T9.2/87-120 7/13/87 J. Patton TRANSLATE command proposal All items from the agenda were discussed (not necessarily in this order) with the following results: 1. Review of SCSI-2 draft document section 8 (86-109 Rev1) Document X3T9.2/87-111: WRITE & VERIFY command -- the FUA bit was changed back to reserved, since the data must be written to the medium in order to verify it. The VERIFY command will also be modified. The implementor's note was modified to remove the first sentence and was moved to paragraph #1. Also added was a reference to the optional nature of MODE SELECT. Document X3T9.2/87-112: FORMAT UNIT command -- A disable saving parameters bit will be added in bit 2 of byte 1 of the FORMAT data header. An initialization pattern feature was also added (see agenda item #6). This document was accepted as ammended and will be incorporated in SCSI-2 R2. MODE SELECT document from the working group: Changeable pages should be referenced in MODE SELECT. The AWRE/ARRE bits need some work to reduce redundancy. The working group recommends that queuing of RESERVE commands be removed from SCSI-2. An IMMED bit was added to the FLUSH CACHE command. 2. Downloading microcode (87-7 Rev 2) [Bill Spence] {Plenary had already accepted this document with changes noted in the minutes.} Bill brought in Rev 3 incorporating the changes requested in the plenary meeting. This will be incorporated in SCSI-2 R2. 3. INQUIRY enhancements/vital product data (87-42) [Greg Floryance] Additions to INQUIRY Command (87-117) [Peter Johansson] Document 87-42: Deferred -- Greg was not present. Document 87-117: A new field, Serial Number, was added in bytes 56 to 67 of INQUIRY data. Gary Stephens will provide the wording to be incoporated in SCSI-2 R2. 4. READ LOG command (87-43) [Greg Floryance] Deferred -- Greg was not present. 5. Cabling requirements (87-44) [D. Kinsell] This document will be incorporated in SCSI-2 R2 and the appendix. 6. Certification Pattern (87-47) [Daniel Loski] (87-83) [Tony Kozlowski] Passing Media Initialization Pattern (87-116) [Dexter Anderson] Initially, there was a straw vote 8-4 against adding an initialization pattern. The sentiment was to drop the initialization/certification pattern topic. However there was some sentiment to add a method to have an incrementing LBA written in each logical block at format time. Later, during agenda item #1, we accepted a revised SMS proposal. 7. Asynchronous event notification (87-52) [Bob Snively] Bob proposed adding a bit to the SEND command to indicate that the format of the data is REQUEST SENSE data format. A revised proposal will be forthcoming. There was some concern that COPY managers would confuse a SEND command used for Asynchronous Event Notification with their normal WRITE, SEND, or PRINT command. After some discussion it was pointed out that devices which become initiators solely to perform the COPY command should not indicate that they are Processor-type devices in response to the INQUIRY command. Thus they should not be notified of asynchronous events. Bob will add some words to describe this. 8. Power-on protocols appendix (87-53 R1) [Bob Snively] This document will be incorporated in SCSI-2 R2. 9. Command Queuing (87-59 Rev 2) Bob had not completed a review against the IPI document (lifo vs fifo). The consensus was for lifo, since IPI uses lifo. The open items on command queuing from the last working group meeting are: 1) ABORT message definition Clears a LUN for an initiator; ABORT TAG clears a tag; CLEAR QUEUE clears the queue for the specified LUN for all initiators and generates an UNIT ATTENTION condition. 2) Queue full: messsage or status Changed to status -- allows command to be re-transmitted. 3) Single queue tag messsage Simplified queue tag structure, message codes 2Xh are two bytes long, instead of extended messages. This is a new message structure. 4) Linked command execution Executed with a single queue tag as a mono-atomic unit. 5) Need for TAG message to follow IDENTIFY message or CDB Some question of performance on existing silicon; After discussion, agreed that TAG should follow IDENTIFY. 6) Data integrity This is a complex system issue; Do we need extra protection? All system manufacturer's are requested to review and advise. This an open issue and it must be resolved prior to incorporation into SCSI-2. 7) Sense data can be lost Errors will stop execution from queue until REQUEST SENSE. 8) Error handling How are non-disk devices affected? 9) Ambiguous TAG message for REQUEST SENSE Resolved by #7 10) Lifo/Fifo Lifo was favored by IPI O/S people. 11) Ordering of commands Associated with data integrity issue. 10. SCSI level/version issue (87-15, 87-60) [Semenak, Stai, Boulay, Lohmeyer, Snively] Bob has action item to do write-up. There is some consensus that a new command is the best approach to establishing a version level. Bob will generate a proposal based on this approach. 11. SCSI-2 Broadcast (87-61) [Paul Boulay] Paul has withdrawn this proposal from SCSI-2 consideration. 12. READ UPDATED BLOCKS command (87-63) [Skip Jones] This will be incorporated in SCSI-2 R2 modified to make multiple-block transfers as optional. 13. Tape glossary terms (87-65 Rev 1) [Bill Spence] Withdrawn. This document is going to be merged with the addendum of document 87-96. Bill Spence and Gary Stephens will revise the addendum. 14. Wide SCSI (87-36,-69) [Steve Goldman] Wide SCSI Cabling proposal (87-118) [Jim Schuessler] Steve lead a dinner meeting to develop a proposal to solve the odd-byte transfer problem. Connector issues are on the agenda for Colorado Springs. 87-118 was added to the plenary agenda. The feeling of the dinner meeting was to adopt a message to deal with the odd- byte transfer problem -- Steve will provide a written proposal. Steve is looking into changing some of the pin assignments to accomadate a smaller "B" connector for 16-bit systems. 15. Optical Disk 12-byte command set (87-72) [John Hoy] This will not be incorporated in SCSI-2 R2 because it duplicates all of the 10-byte command descriptions, creating a massive document. The editors are investigating ways to include this information without creating so much paper. 16. Media Changers (87-74) [Skip Jones] Additions to Automatic Media Changers (87-119) [Bart Raudebaugh] The working group is in favor of incorporating the proposal as a new section in SCSI-2 R2. Paul Boulay will add a caveat to cover apathy and indifference propounded within the working group. 17. Autosense (87-76 Rev 1) [Paul Nitza] Some clarification is needed on data pointers and additional explanation, especially on handling soft reset, is to be done by Jeff Stai. This will be incorporated in SCSI-2 R2. 18. INQUIRY Peripheral Device Type (87-77 Rev 1) [Paul Nitza] The first page is accepted and will be incorporated in SCSI-2 R2. The second page is to be re-considered for a MODE SENSE page. Paul will generate another proposal for the options reporting. 19. Direct-access tape device-type code (87-84) [Tony Kozlowski] The working group recommends that this proposal be rejected. 20. Immed bit in WRITE FILEMARKS & VERIFY commands (87-88) [Bill Spence] {Plenary accepted -88 with minor changes} Additional Concerns RE: IMMED bit operations (87-109) [Bill Spence] 87-109 has been accepted for incoporation into SCSI-2 R2. The implementor's note on the REWIND command needs some wordsmithing to reflect the old way SCSI-1 devices may have operated. 21. Potential Flaw Management Problem (87-89) [Jim Patton] Translate Command Proposal (87-120) [Jim Patton] Paul Nitza will do a proposal for paging for SEND/RECEIVE DIAGNOSTICS and the Translate page will be included in the proposal. 22. MODE SELECT Error Recovery Parameter Combinations (87-91) [Paul Nitza] This re-write will be incorporated into the SCSI-2 R2 document. The tables will be re-organized to incorporate the individual error recovery parameter bit definitions. The individual bit definitions will be deleted from the text. 23. SCSI I/F Redefined similarly to IPI-1 (87-92) [Gary Stephens] Tabled until SCSI-3. 24. Sequential Access Device Position Definition (87-93) [Gary Stephens] This document is accepted in principle to be added at the end of section 9. There are some coordination issues between the command descriptions and this new section that will be identified for now by a square bracket note. 25. REASSIGN BLOCKS revision (87-94) [Dave McIntyre] This document points out a few errors in the current revision -- it will be included in Rev. 2. 26. MODE SELECT changes that cause UNIT ATTENTION for Tape (87-95) [Gary Stephens] Document 87-95 is withdrawn. Dal Allan accepted an action item to draft alternative wording to cover UNIT ATTENTION reporting conditions. 27. Revised Tape Model (87-96) [Gary Stephens] Document 87-96 is adopted for incorporation in SCSI-2 R2 as an appendix. 28. Clarification of Additional Sense codes for BLANK CHECK (87-97) [Dan Davies] This document is accepted for incorporation in SCSI-2 R2 as amended (add "unexpected"). 29. Proposed Buffered Mode for Deferred Error Handling (87-98) [Dan Davies] The document will be incorporated in SCSI-2 R2, but there is additional work on error recovery to be done. 30. Terminator Power (87-99) [Ralph Schultz] Document 87-99 was withdrawn. Bill Spence will develop a proposal to deal with powering local internal terminators. 31. Media Changer Issue (87-100) [Ralph Schultz] Additions to the Automatic Medium Changer Proposal (87-119) [Bart Raudebaugh] See item 16. 87-119 was discussed, but no action was taken. 32. Tape Buffer Control (size and EOT handling) (87-101) [Dave Skinner] Accepted with minor modifications -- Table 9-20, Byte 10 Bit 3 was designated the Synchronize at Early Warning (SEW) bit. 33. Proposal on discarding volume overflow data (87-102) [Dave Skinner] 87-102 was accepted to be incorporated in SCSI-2 R2 in the REWIND and LOAD/UNLOAD commands. Gary Stephens has an action item to extend the concept to other tape commands where applicable. 34. NCR Printer Proposal (87-103) [Terry Johnson] This proposal was introduced for discussion at the next plenary meeting. 35. Buffered Device Recovery after Check Condition (87-108) [Gary Stevens] There was general agreement on the need for extended contingent allegiance. There is need for work on some areas -- Randy Weber accepted an action item to develop wording for tape commands. This will be incorporated in a future revision. 36. Group 1,2,5 CDB's (87-110) [Dave MacIntyre] The group was evenly divided on whether or not the LUN field should be retained, thus it should probably be retained in SCSI-2. Perhaps we can eliminate (reclaim) this field in SCSI-3. The consensus was that group two commands should be defined as ten bytes in length. 37. Sequential Access Recoverable Error Counter (87-113) [Randy Weber] This document is closely tied to document 87-43 R1. The 87-113 document is accepted subject to the review of document 87-43 R1. 38. Corrections to Additional Sense Codes for Sequential Access Devices (87- 114) [Randy Weber] This document is accepted for incorporation in SCSI-2 R2. The specific code will be assigned by the editor. 39. Flex Interface Pin Definition for MODE SELECT (87-115) [Mike Eneboe] This proposal has some problems with signal definitions. Gene Milligan, Mike Eneboe, and Terry Johnson are assigned an action item to resolve this issue. 40. Smaller connector discussion. Added to the agenda for the Colorado Springs meeting. cc: Attendees who are not P, A, or L members.