FINISH ZONE question

Curtis Stevens curtis.stevens at wdc.com
Tue Jun 9 13:11:28 PDT 2015


Formatted message: <a href="http://www.t10.org/cgi-bin/ac.pl?t=r&f=r1506094_f.htm">HTML-formatted message</a>

Agreed, I have been doing some explaining lately and wanted to see what
others pointed to.  I am looking at this text in the FINISH ZONE command…
The device server returns the initialization pattern for all unwritten LBAs
in this zone (see 4.3.3.3 and 4.3.3.4.2) in response to a read operation. The
device server may write the initialization pattern to the media for unwritten
LBAs.
We have taken pains to refer to unwritten LBA’s, this issue seems to be
that nothing is said here about the written LBAs after the write pointer… 
I can go on explaining, just wondering if others are having issues with a
lack of clarity.
-------------------------------------------------
Curtis E. Stevens
Director, Standards & Features Technology
3355 Michelson Dr. #100
Office: 1-1041
Irvine, Ca. 92612
Phone: 949-672-7933
Cell: 949-307-5050
E-Mail: Curtis.Stevens at WDC.com
Remember, you may only be blamed for something if you are actually doing
something.
From: Brad Besmer [mailto:brad.besmer at avagotech.com]
Sent: Tuesday, June 9, 2015 12:46 PM
To: Neil Wanamaker; Curtis Stevens; T10 at t10.org
Subject: RE: FINISH ZONE question
Curtis,
As currently specified in zbc-r03:
>> One implementation would leave the write pointer at LBA 10.
This would be compliant. In 4.3.3.3 Sequential write preferred zone model:
Figure 8 shows examples of write commands that do not start at the write
pointer (i.e., a non-sequential write operation). The ending value of the
write pointer for these examples is outside the scope of this standard.
>> At this point, if the host issues a FINISH ZONE, is LBA 50 required to be
preserved?
Yes, as indicated by 4.3.3.1<http://4.3.3.1&gt;: Write pointer zones overview
A read operation on an LBA that has been written since the last reset write
pointer operation in that write pointer zone shall return the last written
data.
Are you perhaps suggesting a potential change?
Brad
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 Neil
Wanamaker
Sent: Tuesday, June 09, 2015 1:22 PM
To: Curtis Stevens; T10 at t10.org<mailto:T10 at t10.org>
Subject: RE: FINISH ZONE question
Curtis,
The description of the FINISH ZONE makes no reference to the current value of
the write pointer, but rather to the data returned when UNWRITTEN (my
emphasis) LBAs are read, and to the ending zone condition.
If a drive is not required to preserve data written to it, it has little
value as a storage device.
Neil Wanamaker
Principal Engineer, Enterprise Storage Division
PMC-Sierra
1380 Bordeaux Dr., Sunnyvale, CA 94089
408.239.8277
neil.wanamaker at pmcs.com
From: owner-t10 at t10.org<mailto:owner-t10 at t10.org> [mailto:owner-t10 at t10.org]
On Behalf Of Curtis Stevens
Sent: Tuesday, June 09, 2015 11:27 AM
To: T10 at t10.org<mailto:T10 at t10.org>
Subject: FINISH ZONE question
In a sequential write preferred zone, if there are 100 LBAs, and there is a
write to LBA’s 0-9, and then a write to LBA 50.  One implementation would
leave the write pointer at LBA 10.
At this point, if the host issues a FINISH ZONE, is LBA 50 required to be
preserved?
-------------------------------------------------
Curtis E. Stevens
Director, Standards & Features Technology
3355 Michelson Dr. #100
Office: 1-1041
Irvine, Ca. 92612
Phone: 949-672-7933
Cell: 949-307-5050
E-Mail: Curtis.Stevens at WDC.com
Remember, you may only be blamed for something if you are actually doing
something.



More information about the T10 mailing list