ZBC - possible late letter ballot comment

Black, David david.black at emc.com
Tue Mar 3 19:05:51 PST 2015


Formatted message: <a href="http://www.t10.org/cgi-bin/ac.pl?t=r&f=r150303d_f.htm">HTML-formatted message</a>

> I can't argue why that subtle difference is essential in two different
paragraphs though, it does seem possible to combine
> the thoughts into a more general statement that if any addressed LBAs are
in a sequential write required zone and any
> of the other addressed LBAs are in a different zone then it is a boundary
violation and termination etc is in order.
+1, e.g.:
If the device server processes a write command with an ending LBA that is not
in
the same zone as the starting LBA and that command specifies an LBA in a
sequential
write required zone, then the device server shall terminate the command with
CHECK
CONDITION status, with the sense key set to ILLEGAL REQUEST, and the
additional
sense code set to WRITE BOUNDARY VIOLATION.
Thanks,
--David
From: owner-t10 at t10.org [mailto:owner-t10 at t10.org] On Behalf Of Ballard,
Curtis C (HP Storage)
Sent: Tuesday, March 03, 2015 7:35 PM
To: Joe Breher; Gerry Houlder
Cc: T10 Reflector; Curtis Stevens; Paul Suhler
Subject: RE: ZBC - possible late letter ballot comment
'an ending LBA that is not in the same sequential write required zone as the
starting LBA'
Highlighting 'same' and 'in . . . sequential write required zone'. Using
'same' implies that it is in a sequential write required zone and the
question is whether that sequential write required zone is the same one.
Maybe it is too subtle but I believe the intent that was discussed when this
text was worked on was "in a sequential write required zone but not the same
sequential write required zone'. The second statement is explicitly not in a
sequential write required zone at the start.  I think the difference is that
the cases are starting from two different zone types.
I can't argue why that subtle difference is essential in two different
paragraphs though, it does seem possible to combine the thoughts into a more
general statement that if any addressed LBAs are in a sequential write
required zone and any of the other addressed LBAs are in a different zone
then it is a boundary violation and termination etc is in order.
>From a programming perspective though I can see the requirement resulting in
two different checks that roughly align with the existing text. I think
'starting in a sequential required zone' and 'starting in a zone that is not
a sequential required zone' could be different processing paths.
Curtis Ballard
Hewlett-Packard Company
+1 970 898 3013 / Tel
Curtis.Ballard at hp.com / Email
Fort Collins, CO
From: Joe Breher [mailto:Joe.Breher at hgst.com]
Sent: Tuesday, March 3, 2015 4:54 PM
To: Gerry Houlder
Cc: Ballard, Curtis C (HP Storage); T10 Reflector; Curtis Stevens; Paul
Suhler
Subject: Re: ZBC - possible late letter ballot comment
Well, I disagree with both of you. :)
By 'an ending LBA that is not in the same sequential write required zone as
the starting LBA', this covers any and every case where more than one zone is
specified. There is no way to have an ending LBA in the same zone as the
starting LBA, while some of the specified LBAs are in another zone.
The second one is only applicable to cases where LBAs of multiple zones are
specified. As such, covered by the first.
If it's any consolation, I stared at it for a while -- then consulted with
others to reinforce my sense of unease -- before pulling the trigger.
If I'm wrong, I'm still not seeing it - help me out.
Joe Breher
Storage Architecture Technologist
Standards Setting Organization
San Jose Research Center
HGST, a Western Digital company
(478) 2-Breher
(478) 227-3437
This e-mail may contain confidential or legally privileged information of
HGST. If you are not the intended recipient, please notify us immediately by
responding to this e-mail and then deleting it from your system.
On Mar 3, 2015, at 12:25 PM, Gerry Houlder
<gerry.houlder at seagate.com> wrote:
I see Joe's point, but agree with Curtis that one requirement is "going into
a SWR zone from any kind of zone" and the other is "going out of a SWR zone
to any kind of zone". The specification is that any write command that starts
in one zone and ends in a different zone is prohibited, as long as as least
one of those zones is a sequential write required  zone. I agree with Curtis
that these requirements could be combined into a single requirement.
On Tue, Mar 3, 2015 at 12:43 PM, Ballard, Curtis C (HP Storage)
<curtis.ballard at hp.com> wrote:
I don't believe the second paragraph is a subset.  The first paragraph starts
in a sequential write required zone and the second starts in a zone that is
not sequential write required.	Since that starting conditions are different
that doesn't look to me like a subset relationship.
The first paragraph covers the case where the write starts in a sequential
write required zone and tries to cross out of that zone.  In the second
paragraph the write starts in a zone that is not sequential write required
and tries to cross into the sequential write required zone.
It seems like they could be closer together or possible combined into a
single thought but I believe there is a difference.
Curtis Ballard
Hewlett Packard
From: owner-t10 at t10.org<mailto:owner-t10 at t10.org>
[mailto:owner-t10 at t10.org<mailto:owner-t10 at t10.org>] On Behalf Of Joe Breher
Sent: Tuesday, March 3, 2015 10:43 AM
To: T10 Reflector; Curtis Stevens
Cc: Paul Suhler
Subject: ZBC - possible late letter ballot comment
We seem to have some duplicated material in zbc-r02, which has escaped all of
our otherwise vigilant eyes.
Within 4.3.3.4.1, we have both of the following paragraphs:
If the device server processes a write command with an ending LBA that is not
in the same sequential write
required zone as the starting LBA, then the device server shall terminate the
command with CHECK
CONDITION status, with the sense key set to ILLEGAL REQUEST, and the
additional sense code set to WRITE
BOUNDARY VIOLATION.
and
If the device server processes a write command with the starting LBA that is
not in a sequential write required
zone and specifies an LBA that is in a sequential write required zone, then
the device server shall terminate the
command with CHECK CONDITION status, with the sense key set to ILLEGAL
REQUEST, and the additional
sense code set to WRITE BOUNDARY VIOLATION.
The criteria in the latter paragraph is a proper subset of the criteria in
the first paragraph. Accordingly, the latter paragraph is redundant, and
should be stricken in entirety.
If the editor is amenable, I would like to enter this as a late letter ballot
comment.
Upon internal discussion, our esteemed Dr. Suhler points out that ZACr01 has
a similar issue. Expect a corresponding T13 discussion.
Joe Breher
Storage Architecture Technologist
Standards Setting Organization
San Jose Research Center
HGST, a Western Digital company
(478) 2-Breher
(478) 227-3437<tel:%28478%29%20227-3437>
This e-mail may contain confidential or legally privileged information of
HGST. If you are not the intended recipient, please notify us immediately by
responding to this e-mail and then deleting it from your system.



More information about the T10 mailing list