Fuji6 rev100 DRAFT

Henry Gabryjelski Henry.Gabryjelski at microsoft.com
Tue Dec 27 08:32:47 PST 2005


* From the T10 Reflector (t10 at t10.org), posted by:
* "Henry Gabryjelski" <Henry.Gabryjelski at microsoft.com>
*
Ai-san,

I am concerned about the below suggestions for START STOP unit command
processing as related to hybrid discs.  This method means that software
which sends a layer change request (because previously it detected the
logical unit supports the hybrid disc feature) cannot get an error when
the command is a mistake.

I think it is a better (and more easily tested) design to do the
following:

1) If hybrid disc feature is not _supported_ (i.e. does not exist in
feature list), then the bits are ignored.
2) If the hybrid disc feature is supported by the logical unit
(regardless of current bit setting), then the bits are always confirmed
by the logical unit.

This closely matches the processing of other commands:  Since the fields
are (in reality) not defined for logical units without the feature,
those LU can ignore the fields.  But those LU which have the feature
acknowledge that these bits are defined.  Therefore, they cannot ignore
the values.  To make the change as suggested creates a one-time
exception to how fields in command blocks are interpreted, which I think
will cause long-term confusion.

Therefore, I would like to urge that the model section 7.3 (p322) be
modified to match the command section 17.43 (p709), in order to simplify
command processing and ability to test the command for the host side
also.

Thanks,
.


-----Original Message-----
From: owner-mtfuji5 at avc-pioneer.com
[mailto:owner-mtfuji5 at avc-pioneer.com] On Behalf Of Takaharu Ai
Sent: Tuesday, December 27, 2005 12:34 AM
To: mtfuji5 at avc-pioneer.com
Cc: t10 at t10.org
Subject: Re: Fuji6 rev100 DRAFT

Hello Kohda-san and Mt.Fuji members,

We have not finished reviewing whole the document but I would like to
inform you the current comments on Mt.Fuji6 Rev1.0 Draft


1. ASC/ASCQ mistakes
In the description of Format Code=21h, 22h and 23h of SEND DISC
STRUCTURE command, Mt.Fuji requests to set ASC/ASCQ to INVALID FIELD IN
PARAMETER LIST for the CHECK CONDITION status when the parameters have
already been set. But in this situation, the drive returns CHECK
CONDITION status after checking the CDB and never receive any
parameters.
For this type of the condition, INVALID FIELD IN CDB is the correct
ASC/ASCQ.

Same mistake is also at SEND OPC INFORMATION command on page 699.
Mt.Fuji requests to set INVALID FIELD IN PARAMETER LIST whenParameter
List Length field is not set to zero. But it must be INVALID FIELD IN
CDB.


2. Inconsistency among Hybrid disc model and its command
In the Hybrid disc model section 7.3 on page 322, the drive behavior
upon receiving the START STOP UNIT command to request the change of
Format-layer is as follows;
    If the logical unit receives this command when the Hybrid disc
    Feature is not current, the logical unit ignores the FL bit and the
    Destination Format-layer # field, and consequently reacts as if the
    command was the "LOAD" command, i.e. LoEj bit set to one and Start
    bit set to one.
But in START STOP UNIT command section 17.43 on page 709, it is as
follows;
    If the Hybrid disc Feature exists but is not current and either the
    FL bit or the Destination Format-layer # field is not set to zero,
    the logical unit shall terminate the command with CHECK CONDITION
    status, 5/24/00 INVALID FIELD IN CDB.

In my memory, we agreed the drive behavior and its commands in October
meeting according to my proposal explained in the distributed document
named "Command proposal for Hybrid disc-2.pdf". It said;
    When Hybrid Disc Feature is not current, both Layer bit and
    Destination Format-layer # field shall be ignored.
The name of Layer bit was changed to FL bit in November meeting.
So, the description in the Hybrid disc model section is correct and
17.43
must be changed to as follows;
    If the Hybrid disc Feature exists but is not current, both the FL
    bit and the Destination Format-layer # field shall be ignored and
    treated as zero.
This is the mistake on my final proposal documents.


3. Event class to be reported when Format-layer is changed
In the Hybrid disc model section 7.3 on page 323, Class 3 event is
requested to be reported in the text and in the figure. But those must
be Class 4, Media Event.
This is also the mistake on my proposal documents.


4. Editorial matching with INCITS (SBC2r16, SPC3r23)
Followings may also be necessary to be matched to INCITS documents. But
before that, I have some questions;
  - Why do you try to do this at this moment without any consensus of
    Mt.Fuji members?
  - Which documents are referred? I refer the above versions', but it is
    only the draft, not the INCITS standards.

FORMAT UNIT command
  - LUN -> FMTPINFO bit, RTO_REQ bit and LINGLIST bit
  - Format Code field -> DEFECT LIST FORMAT field
  - Reserved -> Vendor specific
  - Interleave Value -> Obsolete
PREVENT/ALLOW MEDIUM REMOVAL command
  - PREVENT/ALLOW MEDIUM REMOVAL command -> PREVENT ALLOW MEDIUM REMOVAL
    command
READ (10)/(12) command
  - LUN (Obsolete) field -> RDPROTECT field
  - Reserved -> FUA_NV bit
  - RelAdr -> Obsolete
  - Reserved -> GROUP NUMBER field
READ CAPACITY command
  - RelAdr -> Obsolete
START/STOP UNIT command
  - START/STOP UNIT command -> START STOP UNIT command
  - Codes for Power conditions -> START_VALID, ACTIVE, LU_CONTROL,
    FORCE_IDEL_0 and FORCE_STANDBY_0 are added, Code=5h is Obsolete
SYNCHRONIZR CACHE command
  - SYNCHRONIZE CACHE command -> SYNCHRONIZE CACHE (10) command
  - Reserved -> SYNC_NV bit
  - RelAdr -> Obsolete
  - Reserved -> GROUP NUMBER field
VERIFY (10) command
  - RelAdr -> Obsolete
  - Reserved -> GROUP NUMBER
WRITE (10)/(12) command
  - LUN (Obsolete) field -> WRPROTECT field
  - Reserved -> FUA_NV bit
  - RelAdr -> Obsolete
  - Reserved -> GROUP NUMBER
WRITE AND VERIFY (10) command
  - LUN (Obsolete) field -> WRPROTECT field
  - RelAdr -> Obsolete
  - Reserved -> GROUP NUMBER

Read/Write Error Recovery mode page
  - Read/Write Error Recovery -> Read-Write Error Recovery
  - Correction Span field, Head Offset count field, Data Strobe Offset
    Count field -> Obsolete
Informational Exception Control mode page
  - Reserved -> EBF
Power Condition mode page
  - Idle Timer -> IDLE CONDITION TIMER
  - Standby Timer -> STANDBY CONDITION TIMER

ASC/ASCQ
  - 11/07 RE-SYNCHRONIZATION ERROR -> 11/07 DATA RE-SYNCHRONIZATION
ERROR
  - 3/14/00 RECORDED ENTITY NOT FOUND Write -> 8/14/00 RECORDED ENTITY
    NOT FOUND Read
  - We shall not define Sense Key for the ASC/ASCQ which is not
    specified to be used for Mt.Fuji devices.

Best Regards,

Harry Ai
VEBU
Panasonic AVC Networks Company
Matsushita/Panasonic
Osaka, Japan



---------------- Start of the original message ----------------
>From: takeshi_kohda at post.pioneer.co.jp
>To: mtfuji5 at avc-pioneer.com
>Cc: t10 at t10.org
>Date: Tue, 20 Dec 2005 14:45:22 +0900
>Subject: Fuji6 rev100 DRAFT
>
>Dear all,
>
>After the publication of rev 0.99, I have reviewed through the document
and
>corrected some inconsistency between Fuji and MMC, SPC.
>The updated document is available on the fuji ftp site.
>
>ftp://ftp.avc-pioneer.com/Mtfuji_6/Spec/Fuji6r100DRAFT_diff.pdf
>or
>ftp://ftp.avc-pioneer.com/Mtfuji_6/Spec/Fuji6r100DRAFT_diff.zip
>
>Please check the section 1.9 Change history and colored portions in
this
>document.
>
>Best regards,
>---
>Takeshi Kohda
>

----------------- End of the original message -----------------


*
* 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