Minutes of CAM-2 Study Group Meeting -- January 10, 1994

Ralph Weber -- VMS -- ZKO3-4/U14 weber at star.enet.dec.com
Thu Jan 13 06:55:10 PST 1994

Accredited Standards Committee                     X3T10/94-015R0
X3, Information Processing Systems

To:       Membership of X3T9.2 and X3T10

From:     Ralph O. Weber

Date:     January 12, 1994

Subject:  Minutes of CAM-2 Study Group Meeting
          January 10, 1994

1.0 Opening Remarks
2.0 Attendance and Introductions
3.0 Call for items to be placed on the current agenda
4.0 Review of CAM Revision 5

1.0  Opening Remarks

Ralph Weber, Digital Equipment Corp. called the meeting to order
at 9:15 am, Monday, January 10, 1994.  Ralph announced that Bill
Dallas, Editor of the CAM documents, could not attend due to
pressing requirements at his office.

The attendance list was circulated.  As is customary, the people
attending meeting introduced themselves.  Ralph thanked Rich
Bailey and Vitro for hosting this meeting.

2.0  Attendance

The following people attended the meeting:

Name                     Organization
Mr. Neil Wanamaker       AMDAHL
Mr. Jerrie Allen         AMDHAL
Mr. Gary Porter          Ancot
Mr. Ralph Weber          Digital Equipment Corp.
Mr. Gary Stephens        FSI Consulting Services
Ms. Carol Meister        Jaycor
Mr. Charles Binford      NCR
Mr. Sam Karunanithi      Hitachi
Mr. Bob Snively          Sun Microsystems

9 People Present

3.0  Call for items to be placed on the current agenda

The agenda was approved as presented.

4.0  Review of CAM Revision 5

The group undertook a line-by-line review of CAM Revision 5
starting at the title page.  Numerous editorial nits were noted. 
For example, page viii is headed "Table" not "Tables."  The next
several paragraphs review the more complex editorial changes.

The first paragraph of Clause 4.1 was rewritten to read:

     "A model of the CAM usage environment is illustrated in
     Figure 1.  Multiple applications are shown accessing a
     variety of SCSI devices.  Several device drivers, both
     peripheral drivers and SIMs, are present to support the
     peripherals on the system."

The following sentences were added at the end of Clause 4.1:

     "The choice of XPT and SIM packaging is an operating system
     dependency.  Clause 8 defines this dependency for certain
     named software and hardware platforms."

The group could find no real use for item i in the list of SIM
functions (Clause 4.4).  Item i reads: "Mechanisms to accept the
selecting and sensing of the SCSI HBA functions supported."  The
group recommended dropping it from Revision 7, unless a justifi-
cation for keeping it is presented before then.  A footnote to
that effect will be added to Revision 6.

The following definition for OSD was added to Clause 5.1:

     "OSD (operating system dependent):  This term describes a
     capability, method of operation, or feature that depends on
     the specific operating system on which CAM is implemented."

Clauses 6.2.5 and 6.3 were restructured and made to be a single
clause (Clause 6.3, Architectural Considerations).  Clause 7 was
renamed to "Principles of Operation."

Clause 7.3.2 was completely rewritten to read:

     "A peripheral driver is notified of the completion of a
     queued CCB using a callback routine.  The SIM/XPT calls the
     peripheral driver's callback routine when execution of the
     queued CCB completes.  The queued CCB includes a pointer to
     the peripheral driver's callback routine (in the Callback on
     completion field).

     "The peripheral driver callback routine is used much like a
     hardware interrupt handler.  The callback routine has the
     same privileges and restrictions as an interrupt handler.

     "The address of the specific CCB that completed is passed to
     the peripheral driver's callback routine."

In addition to the numerous editorial problems, a few larger
issues were raised.  The remainder of these minutes describe
the larger issues.

Because of the wording in some conformance statements, questions
were raised about the XPT to SIM interface.  If an SIM from one
source is expected to interoperate with an XPT from another
source, should not the CAM standard define the XPT to SIM
interface?  (Maybe such documentation already is present and the
group simply failed to find it.)  If the XPT to SIM interface is
not part of CAM, should some of the conformance statements be

The group strongly recommends moving the current Clause 5
(Definitions and conventions) to Clause 3.  This will result in
renumbering of Clauses 3, 4, and 5.  But, after that the Clause
numbers will remain the same.  N.B. This is only a recommendation
to the CAM editor.  The CAM editor was not present to comment on
this recommendation.

Many of those present also feel that an acronyms sub-clause
should be added to the definitions and conventions clause.  This
would gut the current definitions clause (because that clause is
mostly acronyms).  Again, this is only a recommendation to the
CAM editor.

The usage of the word "nexus" is inconsistent.  The usage in the
CAM document is not consistent with the definition (currently in
Clause 5).  Furthermore, the definition of "nexus" in CAM
conflicts with its definition in SCSI-2.  At a minimum, "nexus"
must be used consistently throughout CAM.

The group recommends that the SCSI-2 definition of "nexus" be
adopted for CAM.  A different word must be found to replace the
other uses of "nexus" in CAM.  If this recommendation is adopted,
then "nexus" should be removed from the definitions clause. 
SCSI-2 is a normative reference document for CAM.  So, all the
SCSI-2 definitions automatically apply in CAM.

More information about the T10 mailing list