Additional important comment on SES

Bob Snively bob.snively at Eng.Sun.COM
Thu Oct 31 13:37:22 PST 1996


* From the SCSI Reflector (scsi at symbios.com), posted by:
* bob.snively at Eng.Sun.COM (Bob Snively)
*

Jay Elrod of Seagate has pointed out some discrepancies between
SFF-8067 and SES revision 7.0 with respect to the behavior of
a disk drive providing interface services for an enclosure services
process.  I hope I am paraphrasing our teleconference correctly when I
say:



A)	Comment:

	SFF-8067 is woefully out of date, only mentioning 4 of the
	pages defined for SES. 

	Solution: 

	This should be brought into compliance.  I agree.

	This requires an update of SFF-8067.

B)	Comment:

	As presently worded, there is a strong implication in SFF-8067
	that the disk drive should consider direction of transfer for each
	page and provide check conditions for each page.  Jay points out
	that this may be unknowable by a disk drive as future revisions
	to SES occur.  

	Solution:

	Jay suggests that the disk drive simply pass through the requests.
	For SEND DIAGNOSTIC commands that were invalid or of an invalid
	length, the enclosure services processor would be required to
	deposit the arriving stream of bits in the bit bucket.  An
	INVOP condition would be established for later presentation.
	For RECEIVE DIAGNOSTIC RESULTS commands that were invalid or
	of an invalid length, the enclosure services processor would
	be required to create a page header that indicated an additional
	data length of zero and the drive would be required to 
	deal with this by posting back just the received page header.
	Again, an INVOP condition would be established for later
	presentation by the enclosure services device.  I agree.

	Note that enclosure services devices do not have any of these
	problems and would simply generate an appropriate check
	condition when such invalid requests were made.

	This requires clarification of both SFF-8067 and may require
	a few additional words in selected paragraphs of SES.

C)	Comment:

	As presently worded, all unused pages are reserved for general
	SCSI use.  Jay strongly recommends that all the pages 1-F
	be reserved for enclosure services functions.  That way, he
	can create a drive that simply passes any of these pages
	through for the enclosure services device to throw away or
	use.  Any new pages introduced into SES or SES-2 would not
	require the drives to be modified, but could be simply
	passed on through to the enclosure services device (at least
	until we had more than 15 pages defined).  While such 
	expansion is not likely at any time soon, his point is valid
	and these "spare pins" should be provided.  Any pages not 
	valid would be treated as in B.

	Solution:

	The clarification of SFF-8067 would include this change as
	an extension to B above.

	Small wording changes might be required to section 4.1.5 of SES.

	Section 6.1 would have new lines installed in the diagnostic
	page code table for reserved enclosure services pages.
	This would include pages 08-0F.  In addition, the reverse
	directions for single-direction pages would be clearly
	labeled as reserved.  The text would be modified to indicate
	how those additional pages were to be handled, probably
	referencing the modified section 4.1.5.

I believe that these should be added to the technical corrections of
SES, as extensions to the following problem numbers:

	SU.21 (referencing section 4.1.5, clarifying INVOP)
	SU.18/29 (referencing section 4.1.1 and 4.1.2, clarifying model)
	SU.131 (referencing 6.1.7, but corrected in section 4.1.6)
	SY.43 (referencing the reserved values in the diagnostic page codes)

	but documented as separate comments identified during the
	editorial process for convenience.

Is that okay with everybody?
	

	
*
* For SCSI Reflector information, send a message with
* 'info scsi' (no quotes) in the message body to majordomo at symbios.com




More information about the T10 mailing list