[MtFuji] MMC/Fuji: Proper way for device to fail a read/write streaming command?

keiji_katata at post.pioneer.co.jp keiji_katata at post.pioneer.co.jp
Tue May 12 23:24:40 PDT 2009


* From the T10 Reflector (t10 at t10.org), posted by:
* keiji_katata at post.pioneer.co.jp
*
PS.
I'm very sorry that a "not" was lost.
I think that the Streaming model does not cover data writing with a specific
bits rate.
Best regards,
Keiji Katata
PIONEER CORP.
Keiji Katata
2009/05/13 15:21
$B08 at h(B:  mtfuji5 at avc-pioneer.com
cc:    T10 Reflector <t10 at t10.org>
$B7oL>(B:  Re: [MtFuji] MMC/Fuji: Proper way for device to fail a
read/write
       streaming command?  (Document link: Keiji Katata)
Hi David,
Here are questions.
>commits with streaming to deliver a given throughput, respond promptly, etc.
Could you tell me your test environment?
>some device also return this sense information to streaming write.
Could you tell me when the drive reported 2/4/8? And why did you think that
this
is not desired?
I think that these information may be important to discuss your writing.
Here are comments.
I think that drive can report 2/4/8 on a write command with Streaming=1.
2/4/8
means that a drive buffer is full to receive the write command or a drive is
performing a higher priority job than the write command with Streaming=1 e.g.
OPC for the zone. When this OPC was failed the write command will be failed
with
Media error or some fatal error.
Streaming model of MMC/Fuji is designed for two purposes.
The 1st purposes that a system can allocate Continuos Data Area (CDA) on a
disc
zone. Streaming=1 write stops any reallocation of the data sector for the
CDA.
The 2nd purposes that a system can read a CDA with a desired bits rate and a
seek latency.
I think that the Streaming model does cover data writing with a specific bits
rate.
Best regards,
Keiji Katata
PIONEER CORP.
David Burg <daviburg at windows.microsoft.com>@avc-pioneer.com on 2009/05/12
05:43:59
mtfuji5 at avc-pioneer.com$B$KJV?.$7$F$/$@$5$$(B
$BAw?.<T(B:	  owner-mtfuji5 at avc-pioneer.com
$B08 at h(B: T10 Reflector <t10 at t10.org>, "mtfuji5 at avc-pioneer.com"
<mtfuji5 at avc-pioneer.com>
cc:
$B7oL>(B: [MtFuji] MMC/Fuji: Proper way for device to fail a read/write
streaming
command?
Hello,
While testing drives for Windows logo, we found that some write commands are
failing with sense information 02/04/08.
That is absolutely fine and actually desired with non-streaming writes.
However, some device also return this sense information to streaming write.
We
are both wondering if this is legitimate ? as the drive commits with
streaming
to deliver a given throughput, respond promptly, etc. and if the sense code
02/04/08 is the right one, or if it should be $B!H(BUNIT
ATTENTION/INSUFFICIENT TIME
FOR OPERATION$B!I(B or something else.
Is a drive allowed to fail to complete a streaming read or write within a
short
time, and if it is allowed, how the drive fail the command exactly?
With regards,
David.
*
* 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