I believe that I have found an inadvertent change between SPC-2 and SPC-3 that is still in SPC-4.  This came about in the addition of the descriptor format of the sense data.  The change is related to the SKSV bit when the sense key is ILLEGAL REQUEST.

SPC-3 states:
A sense-key specific valid (SKSV) bit set to one indicates the SENSE KEY SPECIFIC field contains valid information as
defined in this standard. An SKSV bit set to zero indicates that the SENSE KEY SPECIFIC field is not as defined by this
standard.

SPC-3 states in 4.5.2.4.2 Field pointer sense key specific data that
If the sense key is ILLEGAL REQUEST, then the SENSE KEY SPECIFIC field shall be [as shown in the table].

This brought in a SHALL that requires SKSV bit to be set to one and the field pointer to be filled in.  In SPC-2 this was optional.

The SPC-2 text reads:
7.20.3 Sense-key specific
A sense-key specific valid (SKSV) bit of one indicates the SENSE-KEY SPECIFIC field contains valid information as
defined in this standard. The SKSV bit and SENSE-KEY SPECIFIC field are optional. The definition of this field is determined
by the value of the SENSE KEY field. This field is reserved for sense keys not described below. An SKSV value
of zero indicates that this field is not as defined by this standard.

If the sense key is ILLEGAL REQUEST and the SKSV bit is set to one, then the SENSE-KEY SPECIFIC field shall be as
defined as shown in table 103. The FIELD POINTER field indicates which parameters in the CDB or the data parameters
are in error.

It seems this came about because the descriptor format, by definition, has the SKSV bit set to one.  I think it was not a conscience decision to require this for the fixed format.

Do you agree this was unintentional and what should be done about it now?

Thanks,

Kevin D. Butt
SCSI & Fibre Channel Architect, Tape Firmware
MS 6TYA, 9000 S. Rita Rd., Tucson, AZ 85744
Tel: 520-799-2869 / 520-799-5280
Fax: 520-799-2723 (T/L:321)
Email address: kdbutt@us.ibm.com
http://www-03.ibm.com/servers/storage/