January 17, 1988 The SCSI-2 CAM (Common Access Method) Committee of January 12 was hosted at the Red Lion Inn, Costa Mesa, by Western Digital. The following attendees were present at the meeting: M. Aarons EMULEX J. McGrath QUANTUM D. Allan ENDL L. Mihalik NCR J. Armstrong NCR T. Murray NCR B. Babbitt BABBITT ENGINEERING S. Nester SMS P. Bair NCR L. Payne OPTOTECH C. Ballard BALLARD SYNERGY R. Perry SEAGATE S. Betts NCR J. Polonsky SMS B. Bonner WESTERN DIGITAL T. Putnam OPTOTECH M. Brownstein INFOWORLD M. Ramezani WESTERN DIGITAL F. Burgess SEAGATE D. Rawcliffe SEAGATE M. Dawson NCR T. Reichert INTERACTIVE SYSTEMS L. Edwards VERBATIM L. Robinson STORAGE DIMENSIONS M. Evans SYTRON Z. Sagi WESTERN DIGITAL C. Fleckenstein NCR J. Semenak AT&T J. Hamada MITSUBISHI ELECT S. Shah WESTERN DIGITAL L. Holmstrom IOMEGA T. Shea ADAPTEC W. Ihde LMS OSD D. Skinner NCR M. Jerbic HEWLETT PACKARD G. Slaughter SUN MICROSYSTEMS J. Katzung APPLE B. Snively ADAPTEC L. Lamers SONY J. Stedt PEER PROTOCOLS J. Lawlor AT&T J. Tennant SMS C. Linsley SMS J. Tierney WANGDAT J. Lohmeyer NCR A. Welsh COLUMBIA DATA P. Manley SYQUEST TECHNOLOGY B. Wheeler SYSGEN G. Marazas IBM E. Young ARCHIVE D. McDonald DILOG Presentations were made by: SUN MICROSYSTEMS Greg Slaughter BALLARD SYNERGY Clint Ballard NCR Jerry Armstrong The main focus of this meeting was to define the model and a glossary of terms. Jerry Armstrong had consolidated material from several people. By arrangements made via the SCSI BBS there was a pre-meeting get-together by those working on expanding the structure based concepts from December. The results of this consolidated the individual efforts and formed the basis for much of the material covered in the meeting. NOTE: The SCSI BBS (316-636-8700) is the clearing house for idea exchanges between members working on proposals. If you want to participate, dial in and go to Area 13 for the CAM information. Proposals etc will be in the Files section and short exchanges are in the Message section. Greg Slaughter described Sun Microsystems's approach toward writing SCSI drivers. There was an interesting interchange on the rationale for his approach which included no provision for AutoSense. Clint Ballard at Ballard Synergy described the syntax which had been defined at the pre-meeting at Western Digital the previous evening (see attached). Jerry Armstrong of NCR described the structure which had been agreed upon the previous evening (see attached). There was much interaction on the subject matter presented, and an expansion of the model to define the levels within the CAM layer. The name of the function is shown in the boxes and the relevant control blocks are named to the right. CPU & Operating System - - - - - - - - - - - - - - - - - - - - - - - - - - - - +-----------------------+ | OSD Driver | Operating System Dependent Control Block +-----------------------+ | Transport Module | Transport Control Block +-----------------------+ | SCSI Interface Module | Common Access Method Control Block +-----------------------+ | Protocol Hardware | Command Descriptor Block +-----------------------+ - The OSD is Operating System Dependent and will contain the references that are necessary to cross the border in and out of the CAM layer. - The Transport Module contains a truncated form of the CAM Control Block for functions which do not need a SCSI CDB e.g. CAM configuration. - The Interface Module manages the CAM Control Block and contains all the information necessary to dispatch a CDB on behalf of the requester. - The protocol hardware is self-explanatory. Proposals on how pointers should be handled are to be included in the OSD interface. In the control blocks a pointer is presently represented as being 4* bytes - the * indicates that we do not have a method yet, and the field size is subject to change. OS/2 may require both Physical and Virtual pointers to be carried and other environments may use segments with offsets. A lot of details have to be fleshed out, and a number of volunteers are going to prepare material which will be consolidated by Jerry Armstrong. In parallel, I will be working on a suitable format for the standardization of the CAM. We have to comply with many editing and style rules to publish standards. I expect to be able to integrate the proposals into a working document that is in a standards format after the February meeting. The volunteers are not necessarily going it alone, as in some cases, more than one was willing to work on the same subject. The volunteers are shown by subject matter, and the number shown in parentheses corresponds to the Open Issues on the last page of the Working Document Proposal. Target Functions Bob Snively (2, 3) Initialization Shishir Shah Queueing Jim McGrath (4) CHS Algorithm Judy Polonsky (7) OSDs: Apple II Jerry Katzung (5) DOS Ron Perry (5) Mac Jerry Katzung (5) Novell Lance Robinson (5) OS/2 Tracy Murray (5) Unix Joe Lawlor (5) The primary task at our February 22 meeting will be to critique all the proposals that are developed over the next few weeks, so there should be a lot of work to review in Austin. Now for the dates. The CAM meetings are going to be scheduled following SCSI meeting activities. The dates and known arrangements are shown below. Be sure to make your reservations as early as possible for a meeting. The plenary hosts will not necessarily be our CAM hosts, but every effort will be made to meet in the same hotel. At present we have no host for Austin. If there are no volunteers then we will run as no-host and charge a meeting fee of $10-20 to cover meeting room costs and coffee. 2/22/89 Austin Wyndham Southpark Hotel 512-448-2222 Mention X3T9/Motorola meeting (No CAM host yet) 3/ 8/89 San Jose tbd by Jim McGrath (Quantum) 4/26/89 St Petersburg Tradewinds Hotel 813-367-6461 Mention X3T9/AMP meeting (Columbia Data is host) 5/10/89 Wichita tbd by John Lohmeyer (NCR) 6/21/89 San Jose tbd by AMD I. Dal Allan Chairman