An SBC-2 question regarding persistent reservation.

Tom Coughlan coughlan at missioncriticallinux.com
Mon Aug 21 14:21:50 PDT 2000


* From the T10 Reflector (t10 at t10.org), posted by:
* Tom Coughlan <coughlan at missioncriticallinux.com>
*
Ralph,

I don't think the separate requirements were the result of any specific
implementation concerns.  My guess is that it went more like this (as
expressed in an imagined ENDL-style dialog): 

"We need to specify that the logical unit shall check the reservation
before the command changes the media.  That is what that o.s.
requires."   

The appropriate words were penned.

"That looks good, but it doesn't say anything about commands that don't
change the media."

"Hmm, not much we can say about their execution steps. Lets keep it
simple and just require that they shall be checked before they start
executing." 

"Okay then, lets add a paragraph that says that."

At this point the group decided to take advantage of the astonishingly
quick resolution of the persistent reservation topic, and  moved on to
the next agenda item....

 :^)
 
Tom

Ralph Weber wrote:
> 
> * From the T10 Reflector (t10 at t10.org), posted by:
> * Ralph Weber <ralphoweber at compuserve.com>
> *
> Ah yes, this question:
> 
>  > A command that does not explicitly write the medium shall be checked
>  > for reservation conflicts before the command enters the current task
>  > state for the first time. Once the command has entered the current task
>  > state, it shall not be terminated with a Reservation Conflict due to a
>  > subsequent reservation. A command that explicitly writes the medium
>  > shall be checked for reservation conflicts before the device server
>  > modifies the medium or cache as a result of the command. Once the
>  > command has modified the medium, it shall not be terminated with a
>  > Reservation Conflict due to a subsequent reservation.
> 
> I've been wondering if the two requirements might relate to the possibility
> that write data may be sent to the target in the same interconnect
> tenancy as the CDB with the expectation that the target would buffer (cache)
> the data until the reservations have been checked.
> 
> Note, I'm not suggesting that I know the answer only curious about a
> possibility.
> 
> Thanks.
> 
> Ralph...
*
* 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