Open Issues from January SAM Working Group - Jeff Stai's Response

Charles Monia, SHR3-2/W4, 237-6757, monia@starch.enet.dec.com 11-Feb-1994 1745 monia at starch.enet.dec.com
Fri Feb 11 14:44:45 PST 1994


The following is extracted from the original note:

>>Item 2: After a contingent allegiance condition occurs, how should a
>>	logical unit respond to a CLEAR ACA task management function
>>	while an ACA command is pending?
>>
>>
>>Alternative a)The CLEAR ACA task management function shall cause the
>>		ACA command to be aborted.
>>
>>Alternative b)The CLEAR ACA task management function shall not affect
>>		the pending ACA command. However, other pending commands
>>		may then complete.
>>

>
> In response to the above, Jeff Stai wrote:
>

"Regarding item 2: Option b) seems more deterministic. If the initiator
wants the ACA command aborted, it should do it explicitly"

> The following observation is from past correspondence  on
> this issue. I regret that this didn't come to mind when I sent
> the original note.
> 	
> Anyhow, with respect to item 2, option B - what happens if another contingent 
> allegiance condition occurs while the ACA command is still pending?
> Especially if the new condition was caused by a command from another
> initiator. As I interpret the rules for ACA tasks, both commands would be
> allowed to complete.
>
> I thought the intent of the original rule was to insure that only one ACA task
> would be executed while an ACA condition was in effect. In any event,
> I feel SAM should specify what happens in this case.
>
> Comments? Suggestions?
>
> Charles






More information about the T10 mailing list