Proposed ammendment for 2/4/7 and 2/4/8 concerns

David Burg daviburg at windows.microsoft.com
Mon Dec 4 04:25:37 PST 2006


* From the T10 Reflector (t10 at t10.org), posted by:
* David Burg <daviburg at windows.microsoft.com>
*
Dear Katata-san,
Let me clarify that Microsoft is not only looking to have busy sense code on
all command to avoid host timeout, but Microsoft is also looking into this to
increase the system responsiveness.
Despite new buses been point-to-point, there will still be a host software
waiting on the I/O acceptance by the device. The busy sense code importantly
allows the host software to ca-ncel it request when the device is busy and
the user is not willing to wait any longer. This has an important impact on
the perception of quality from the user (because he drives the experience
rather than the computer drives him).
If listed error codes are fatal, which I would like to be proven, where are
then listed the non-fatal error codes??
(Henry may further response on other points that you brought up thank to his
wider knowledge of the other Windows storage components).
Best regards,
David Burg.
PS.: Thanks to Kohda-san for letting me know that the majordomo server has
bounce my initial message.
-----Original Message-----
From: owner-mtfuji5 at avc-pioneer.com [mailto:owner-mtfuji5 at avc-pioneer.com] On
Behalf Of keiji_katata at post.pioneer.co.jp
Sent: Sunday, December 03, 2006 5:32 PM
To: mtfuji5 at avc-pioneer.com
Cc: T10
Subject: RE: Proposed ammendment for 2/4/7 and 2/4/8 concerns
Henry san, I think you are wrong on this writing.
>The specification already allows the return of 2/4/7 and 2/4/8 error
>codes in all the conditions described.
A listed error code will be reported by a drive at a fatal error. But it may
not
be for time-out.
As I wrote in my document, this is not allowed to avoid the host side
time-out.
For example ATA/ATAPI does not allow to have a time-out on the drive side.
Your
"2/4/7 and 2/4/8 error" is a kind of time-out of the drive side. It may cause
some trouble. Drive side Time-out may cause boot failer from ODD etc.
2/4/8 on CD/DVD-R/RW Write command was necessary to eliminate (stop) that ODD
on
the slave occupied the Parallel ATA cable and CPU could not access HDD on the
master. At the DAO writing, a write command occupied the ATA cable during
writing Lead-in area. User saw the frozen screen at the time. During writing
User data area of CD/DVD, low speed write operation of ODD caused delay of
the
data encoding of the high speed CPU on HDD. 2/4/8 was not used to avoid OS
time-out.
But now this is not necessary because SATA, USB, iSCSI are point to point
connection. So a drive can occupy the port (cable) when it is necessary. So
Pioneer drive designed for Serial connection does not use 2/4/8 error on
Write
command.
Henry san, your proposal is very new item. As far as I know that only Windows
ODD driver may have the short time-out (8-10 sec.). No other system including
Windows other device drivers (ex. HDD) may not have such short time-out
accoring
to my experience.
Best regards,
Keiji Katata
PIONEER CORP.
Henry Gabryjelski <Henry.Gabryjelski at microsoft.com>@avc-pioneer.com on
2006/12/02 10:00:13
mtfuji5 at avc-pioneer.com$B$KJV?.$7$F$/$@$5$$(J
$BAw?.<T(J:	  owner-mtfuji5 at avc-pioneer.com
$B08 at h(J:  <mtfuji5 at avc-pioneer.com>, T10 <T10 at t10.org>
cc:
bcc:
$B7oL>(J:  RE: Proposed ammendment for 2/4/7 and 2/4/8 concerns
Hello Sergio,
This proposal is not requiring any change in the specification. The
specification already allows the return of 2/4/7 and 2/4/8 error codes
in all the conditions described.  The "proposal" is instead a way to
_AVOID_ changing the specifications.  Pioneer's previous proposal was to
add a mode page to enable/disable 2/4/7 and 2/4/8 error codes to avoid
problems with a single OEM's drive validation suites.  So there is not
any new behavior for the majority of drives.
This issue was first discussed in the October 2006 Mt. Fuji meeting,
again at the November meeting, and I believe is scheduled to be voted
upon in the December meeting.  I do hope you will be able to join us at
this meeting.  I would be happy to discuss not only the root problems
which initiated these discussions, but also the benefits that any hosts
will receive from a responsive drive with you during the meeting breaks.
Thank you,
.
-----Original Message-----
From: owner-mtfuji5 at avc-pioneer.com
[mailto:owner-mtfuji5 at avc-pioneer.com] On Behalf Of Sergio Henriques
Sent: Friday, December 01, 2006 4:25 PM
To: mtfuji5 at avc-pioneer.com; T10
Subject: Re: Proposed ammendment for 2/4/7 and 2/4/8 concerns
Dear Mt. Fuji attendees,
This does not seem right to me.  I would expect this new behavior
(that benefits only a host that cannot reprogram its timeout values
to meet new drive technologies) should be OFF by default, not ON by
default.
Sergio Henriques.
Apple Computer, Inc.
On Dec 1, 2006, at 8:26 AM, Henry Gabryjelski wrote:
> I am sorry, I did not attach the file.  Please find the file
> attached this time.
>
>
>
> Thank you,
>
>
>
> Henry Gabryjelski
>
> Microsoft Corporation
>
>
>
> From: Henry Gabryjelski
> Sent: Friday, December 01, 2006 8:20 AM
> To: 'mtfuji5 at avc-pioneer.com'
> Cc: T10
> Subject: Proposed ammendment for 2/4/7 and 2/4/8 concerns
>
>
>
> Hello Mt. Fuji people,
>
>
>
> Here is the detail of the proposed amendments to address Pioneer's
> concerns regarding the use of 2/4/7 and 2/4/8 errors for the
> December 12, 2006 meeting.
>
>
>
> Thank you,
>
>
>
> Henry Gabryjelski
>
> Microsoft Corporation
>
> <Fuji 2-4-7 Proposal.zip>
*
* 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 mailing list