Autosense support

[Ralph Weber -- VMS -- ZKO3-4/U14]

Charles Monia wrote:

>>> I would like to propose that all SCSI-3 protocol standards be required to
>>> define some mechanism for aupporting autosense.
>>>
>>> As autosense is currently specified, an acceptable form of autosense
>>> support is through emulation entirely within the initiator (as is now
>>> the case for CAM SIM).  Since this kind of implementation can apply to 
>>> any SCSI-3 protocol, I see no reason why protocol support for autosense
>>> support should be optional.
>>>
>>> Charles

Defining an autosense protocol on the parallel bus would be an interesting 
challenge.  So, the only thing I could imagine the SIP saying about autosense
is that, "host software can implement autosense in its service delivery 
subsystem layer."  To me, that renders moot any SAM requirement for a protocol 
definition of autosense.

Have I missed something subtile here???

Ralph::