Notify bit in SBP-2 ORBs, especially Management ORBs
Stephen_Finch at notes.ssi1.com
Stephen_Finch at notes.ssi1.com
Fri Oct 31 10:32:31 PST 1997
* From the T10 (formerly SCSI) Reflector (t10 at symbios.com), posted by:
* Stephen_Finch at notes.ssi1.com
*
According to SBP-2:
The notify bit, if set to one, FORCES the target to return a
status in all conditions.
If the notify bit is set to zero, the target MAY return a
status if the command completes without error. The target
MUST return a status if an error occurs.
These rules are true for all types of ORBs.
-------
First, I question the usefulness of the notify bit, but I have
been convinced in the past that there may be devices for which
this might make sense, especially sequential devices that
always execute device commands in order.
I do call into question the viability of this for Management
commands. Take the logout command. If the host doesn't set
the notify bit, the target could choose not to return a status.
Hmm... Did the "failed" status get lost, or was it not sent....
I don't think any driver writer will every code Management
commands without the notify bit set to one. In fact, I would
prefer to mandate it.
I would like to propose that SBP-2 require that the notify bit
be set to one for all management commands.
-------
I opened this can of worms, but how do the rest of you like the smell????
steve finch
PS: I know this is late for next weeks meetings, but...
*
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at symbios.com
More information about the T10
mailing list