Contigent Allegiance question

Binford, Charles cbinford at
Wed Aug 16 06:29:00 PDT 1995

From: Charles Binford
  To: SCSI Reflector

'>' marks John Lohmeyer's comments
'>>' marks Steve Gaskill's comments

>From: Lohmeyer, John
>To: sgaskill
>Cc: SCSI Reflector
>Subject: RE: Contigent Allegiance question
>Date: Tuesday, August 15, 1995 7:22PM
>Hello Steve.  I've imbedded my response to your CA questions, below -- John
>Steve Gaskill wrote:
> ----------

.... stuff deleted ....

>>2. Is a SCSI-2 target that supports tagged command queuing and CA (but not
>>permitted to enqueue commands from other initiators?  Or _must_ it respond
>>BUSY status until the CA condition is cleared?
>SCSI-2 requires BUSY status to be returned to other initiators while a CA 
>ECA condition exists.  The word "shall" is legalese for mandatory.  You
>found the pertinent sentence:

Wait a minute.  We can't go directly to the 'shall' and skip the big 'IF'.
The words in question for CA conditions are pasted below from rev 10L (not 
the final version, but close enough.)

"While the contingent allegiance condition exists, if the target is unable
to maintain separate sense data, the target shall respond to any other
requests for access to the logical unit or target routine from another
initiator with a BUSY status."

The "shall return BUSY" is qualified by whether or not the target keeps 
separate sense data for each initiator.
Note, the words in the next section dealing with ECA do NOT contain the IF.

>>in section 7.7, I find "While the extended contingent allegiance condition
>>exists, the target shall respond to any other request for accessto the
>>unit from another initiator with BUSY status."
>>Does the lack of the word "shall" in the first quote mean that the target
>>not have to respond with BUSY status, and can accept more commands
>>to be placed into the frozen queue)?
>No, quite the opposite.  "Shall" means the behavior is mandatory, not

Section 7.7 is for ECA.  The rules are different for CA.

.... rest of message deleted ....

Charles.Binford at

More information about the T10 mailing list