Known defects in SMC
Erich Oetting x2178
eros at stamps.stortek.com
Fri Dec 17 07:53:16 PST 1999
* From the T10 Reflector (t10 at t10.org), posted by:
* Erich Oetting x2178 <eros at stamps.stortek.com>
> I have been making a light review of the tape and medium changer standards.
Thanks, it's always nice to have input like this.
> The following two entries are incorrect in Table 2. The op codes are
> duplicated with other legitimate entries. The text of the standard seems
> be correct (Interpretation Precedence: text before tables).
> RELEASE ELEMENT (6) 16h17h O 6.6
> RELEASE ELEMENT (10) 56h57h O 6.7
> RESERVE ELEMENT (6) 16h O 6.9
> RESERVE ELEMENT (10) 56h O 6.10
This problem was noted earlier. The RELEASE ELEMENT (6) and RELEASE ELEMENT
opcodes are correct (17h and 57h) in ANSI NCITS 314-1998.
> Since the LOG SELECT command is called out in this draft, the LOG SENSE
> command must also be called out. SPC says that if LOG SELECT is
> LOG SENSE shall also be implemented.
> LOG SENSE support is mandatory in FC-TP, but the SMC standard does not
> it. This must be fixed. The entries should be:
> LOG SELECT 4Ch O SPC-2
> LOG SENSE 4Dh O SPC-2
You are correct here. No idea why this was not discovered earlier.
> The FC-TP profile calls out the REPORT LUNS command as mandatory. This
> does not appear in Table 2. It must be added.
> REPORT LUNS A0h O SPC-2
The REPORT LUNS command does need to be added to the table. This was spotted
earlier, but after SMC was forwarded to ANSI.
> I found another big error in the SMC document. The document calls out
> ERO command as allowed for SMC device servers in Table 2 as optional and
> refers the reader to SPC (line below taken from Table 2).
> REZERO UNIT 01h O SPC
> I have checked both the latest SPC and SPC-2 drafts. No such command
> description appears in either document.
> Either this command is unnecessary for SMC, and the entry removed, or there
> is a big omission in SPC.
> I suspect that the REZERO command description, if needed, must be added to
> SMC. There are REZERO commands lurking around in other command set
> left over form SCSI-2 and still implemented by some device servers.
> With the INITIALIZE ELEMENT STATUS command in SMC, the REZERO command may
> be needed; it may simply be removed from the table. However, such a
> is a technical change that requires some approval (hence the copy to John
The REZERO UNIT command is obsolete and should be removed from the table.
> Also, the FC-TP document seems to require that you reference SPC-2 and
> to permit FC-TP to conform to the base standards. That document also adds
> new mode page requirements and commands not called out, etc.
SPC-2 and FCP-2 did not exist as standards when SMC was forwarded to ANSI.
SMC-2 will refer to these standards. I hope to have a draft SMC-2
the next T10 mailing.
> A thorough review needs to be performed since the base standards have moved
> and the FC-TP profile appears to require certain behaviors that are found
> only in the next generation standards. FC-TP cannot be approved when the
> base standards do not support the features it requires. FC-TP is not
> supposed to define new behavior.
> It appears that there has been no review of SMC vs. the other standards and
> the requirements of FC-TP. FC-TP cannot be approved with SMC in this state.
Thanks for your concern. I will prepare a new revision of an earlier
listing known defects in SMC making note of these additional problems.
Reply to erich_oetting at stortek.com
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at t10.org
More information about the T10