Subject: RE: FCP-4: Sequence count to zero on retries
Date: Thu, 31 Aug 2006 10:37:30 +0100
From: "Banther, Michael" <michael.banther@hp.com>
To: "David Peterson" <David.Peterson@mcdata.com>
Cc: <t10@t10.org>
X-Message-Number: 7201
Formatted message: HTML-formatted message
Attachment #1: banther_michael.vcf

Hi Dave,
HP tape drives can set SEQ_CNT to zero for a sequence retry.
Regards,
Michael Banther
Hewlett-Packard Ltd.
+44 117 312 9503
________________________________
From: David Peterson [mailto:David.Peterson@mcdata.com] 
Sent: 30 August 2006 15:21
To: David Peterson; Banther, Michael; Erich Oetting; Greg Wheeless;
kdbutt@us.ibm.com; paul.entzel@quantum.com; paul.suhler@quantum.com
Subject: RE: FCP-4: Sequence count to zero on retries
I'm still need a response from HP. Quantum, and Sun...Dave
(no disclaimer)
________________________________
	From: David Peterson 
	Sent: Friday, August 25, 2006 1:52 PM
	To: 'BANTHER,MICHAEL (HP-UnitedKingdom,ex2)'; 'Erich Oetting';
'Greg Wheeless'; 'Kevin Butt (kdbutt@us.ibm.com)'; 'Paul Entzel
(paul.entzel@quantum.com)'; 'Paul Suhler (paul.suhler@quantum.com)'
	Subject: FCP-4: Sequence count to zero on retries
	Howdy,
	Forgot to mention the other action item I had for FCP-4. From
the minutes:
	5.2 Continuously Increasing Sequence Count
	It was requested at a prior meeting that FCP-4 require use of
Continuously Increasing Sequence Count for unacknowledged classes of
service (i.e., Class 3).
	Dave reported status on his action item to prepare a proposal.
He has studied the issue and is not yet convinced that the value
justifies the possible impact on implementations. He will be posting the
question to the T10 and T11 reflectors.
	It was noted that this requirement had been rejected
historically, but that more recent FC-4 protocols require it.
	Dave also directed the group to an inconsistency in FCP on the
maintenance of the Sequence Count in the event of Sequence retry with
CISC. He asked the community whether the Sequence Count "may" or "shall"
restart at zero for the retry. See FCP-4 subclauses 8.2, 12.4.1.6, and
12.4.1.7 for the inconsistent statements.
	ACTION: Dave Peterson to contact tape vendors to see whether
they are able to force the Sequence Count to zero on Sequence retries.
	So again, what say ye?
	From FCP-4 Rev 00:
	5.2 Continuously Increasing Sequence Count
	For unacknowledged classes, the Sequence Count for a
retransmitted FCP_DATA IU shall start at zero, even if continuously
increasing sequence count is being used.
	12.4.1.6 FCP_DATA IU recovery - write operations
	For unacknowledged classes, the SEQ_CNT field value may start at
zero, even if continuously increasing sequence count is being
used.12.4.1.7 FCP_DATA IU recovery - read operations
	For unacknowledged classes, the SEQ_CNT field value may start at
zero, even if continuously increasing sequence count is being used.
	Also, does anyone, other than HP who has already replied
indicating support, have an opinion whether continuously increasing
sequence be required?
	Thanks...Dave
	(no disclaimer)