Error History (13-029r5) bug in the craw

Kevin D Butt kdbutt at us.ibm.com
Tue Sep 2 17:36:49 PDT 2014


Formatted message: <a href="http://www.t10.org/cgi-bin/ac.pl?t=r&f=r1409021_f.htm">HTML-formatted message</a>

Fred and all,
I have some comments regarding
SPC-5: Device Internal Status Dump
(by: Frederick Knight)
T10/13-029r5   Uploaded: 2014/08/28   259221 bytes
http://www.t10.org/cgi-bin/ac.pl?t=d&f=13-029r5.pdf
This proposal is addressing the first instance of a difference in what is 
created/retrieved from an error history.  The selection between a 
vendor-specific error history and a device internal status is just one of 
the possibilities where delineation may be needed.  I have a device with a 
similar need to delineate between which error history is being requested 
to be generated.  If I have to generate all my supported error histories 
(i.e., all the Buffer ID's) at the same time, then the time constraints I 
need to meet to generate the error history cannot be met.  If this 
proposal could take a more general approach to a solution to allow a way 
to indicate what is being requested that would be appreciated.
This proposal creates a mode specific field, but that is limited to a 
general type of error history - and has few available options for growth 
and does not permit selecting which BUFFER ID(s) are to be captured.  I 
have been thinking about what could be done to make this a more complete 
generalized solution.  I have an idea that at first causes me concern, but 
I think it would work and not conflict with things there already.
I think one value of the new MODE SPECIFIC field can be taken and have it 
indicate that for BUFFER IDs 00h, 01h, 02h, and 03h, when this MODE 
SPECIFIC value is indicated, the BUFFER OFFSET field (part of it) is used 
to indicate which BUFFER ID is to be captured.	This would provide much 
more granular control over what is requested to be generated.  If desired, 
we could even allow a range of buffer id's to be selected by providing a 
starting and ending BUFFER ID.	Today, for these four buffer id's, the 
BUFFER OFFSET field is required to be zero, so using the BUFFER OFFSET 
field should not conflict with anything. 
Thanks,
Kevin D. Butt
SCSI Architect, Tape Firmware, T10 Standards
Data Protection & Retention
MS 6TYA, 9000 S. Rita Rd., Tucson, AZ 85744
Tel: 520-799-5280
Fax: 520-799-2723 (T/L:321)
Email address: kdbutt at us.ibm.com



More information about the T10 mailing list