REQUEST SENSE LONG bit (01-199r3)

Elliott, Robert Robert.Elliott at COMPAQ.com
Wed Feb 6 21:40:40 PST 2002


* From the T10 Reflector (t10 at t10.org), posted by:
* "Elliott, Robert" <Robert.Elliott at COMPAQ.com>
*
I'm a bit worried about "should" because it lets the logical unit return
either format when LONG=1.  It could use short format for some commands
and long for others.  

I'd rather see it say "if the LONG bit is supported by the target and
the LONG bit is set to one, the long sense data format shall be
returned."  

The top of PDF page 110 (spc3r03) has an example of such wording for the
TUR (test unit ready) bit.  Slight difference: that wording says if the
bit is not supported but is set to 1, a CHECK CONDITION is required.
For the LONG bit, we want that handled per the logical unit's normal
reserved field checking behavior.

There are several ways to determine if LONG is supported:
1. Try it and look at the response code.  If you get back long format,
it is supported.  (With shall, that's conclusive. With should, it's
not.)

2. Use the INQUIRY CMDDT feature to see if that bit appears in the CDB
bitmask.

3. Use the new REPORT SUPPORTED OPCODES command to see if that bit
appears in the CDB bitmask.

4. If the control mode page LONG_SENSE bit is changeable (meaning that
the long format is available for autosense and AER), the LONG bit is
probably supported in REQUEST SENSE CDB.  The REQUEST SENSE description
could mandate that they match.

---
Rob Elliott, Compaq Server Storage
Robert.Elliott at compaq.com

> -----Original Message-----
> From: Ralph Weber [mailto:ralphoweber at compuserve.com]
> Sent: Tuesday, February 05, 2002 8:57 AM
> To: T10, Reflector
> Subject: REQUEST SENSE LONG bit (01-199r3)
> 
> 
> * From the T10 Reflector (t10 at t10.org), posted by:
> * Ralph Weber <ralphoweber at compuserve.com>
> *
> 01-199r3 contains the following statement:
> 
>   "If LONG is set to one [in the REQUEST SENSE CDB],
>   the long sense data format shall be returned."
> 
> This statement is not backwards compatible. If a
> REQUEST SENSE CDB is sent to an SPC-2 device server
> (to whom the LONG bit is reserved and need not be
> tested), that device server will return the non-long
> (current format) sense data, no if's and's or but's.
> 
> The above cite sentence is being "edited" to change
> 'shall' to 'should'.
> 
> The bottom line is that an application client that
> is willing to accept long sense data is obligated
> to check the response code in any sense data it
> receives in order to verify the sense data format.
> 
> FYI
> 
> Ralph...
> 
> 
> 
> 
> *
> * 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