reading blocks prior to a medium error

Ralph Weber roweber at IEEE.org
Tue Dec 28 13:46:40 PST 2010


* From the T10 Reflector (t10 at t10.org), posted by:
* Ralph Weber <roweber at ieee.org>
*
Regrettably, SAM-5 cannot be the only source for the commonsense
justification in the case of the RECOVERED ERROR sense key.
Level I -- SAM-5 as previously quoted:
"While some valid data may be present for other values of status,
the application client should rely on additional information from
the logical unit (e.g., sense data) to determine the state of the
buffer contents."
At this point, we must leave SAM-5 and find a description of the
sense data (specifically the sense key) since that is what is
being questioned by the proposed example.
Level II -- SPC-4 sense key definitions table, to whit:
"RECOVERED ERROR: Indicates that the command completed successfully,
with some recovery action performed by the device server. ..."
Viola! Commonsense prevails in SCSI (if one only digs deep enough).
All the best,
.Ralph
On 12/28/2010 10:16 AM, Douglas Gilbert wrote:
> * From the T10 Reflector (t10 at t10.org), posted by:
> * Douglas Gilbert<dgilbert at interlog.com>
> *
> George Penokie wrote:
>   >  This whole issue is about UNRECOVERED read errors
>   >  not RECOVERED read errors. All the read data is
>   >  valid when there is a RECOVERED read error. The
>   >  wording in SAM that states " While some valid data
>   >  may be present for other values of status, the
>   >  application client should rely on additional
>   >  information from the logical unit (e.g., sense data)
>   >  to determine the state of the buffer contents." allows
>   >  this.
>
> I am now taking issue with the two adjoining sentences in
> SAM-5 section 5.1 describing the Data-in buffer. In this
> thread they have now be used (when the status is
> CHECK CONDITION) to
>     a) justify that in the case of blocks prior to a
>	 medium error (i.e. an unrecovered error) that
>	 the initiator "_shall_ treat the buffer contents
>	 as invalid";
>     b) and when the sense key is recovered error, the same
>	 data-in buffer is valid because the following
>	 sentence contains some weasel words.
>
> Since sense data is defined in SPC-4, would it not be better
> that SAM-5 said nothing on the issue of validity of the
> data-in buffer? Alternatively it could just say it was
> valid when the status was good or condition met and refer
> to SPC-4 for other cases.
>
> IMO George Penokie's issue on this subject, raised at a
> recent T10 meeting, is not resolved by reference to that
> data-in buffer description in SAM-5.
>
> Douglas Gilbert
>
> *
> * For T10 Reflector information, send a message with
> * 'info t10' (no quotes) in the message body to majordomo at t10.org
>
>
>
*
* 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