Quantum 10K II - Manual (Page 80)

SCSI Description
Quantum Atlas 10K II Ultra160/m SCSI Hard Disk Drives 5–13
5.5 Extended Contingent Allegiance Condition
Under SCSI-2 processing, the Extended Contingent Allegiance condition extends
Contingent Allegiance handling so the initiator has explicit control over the point at
which suspended (queued) commands are resumed. It allows the initiator to send a
REQUEST SENSE command and then send the non-tagged commands to access the
drive’s medium.
This action is provided to fix damage left by the failed command before permitting
the drive to continue with processing the queued commands.
The Extended Contingent Allegiance condition is enabled and disabled by the EECA
bit in the MODE SELECT Control Mode page. The Extended Allegiance Condition
differs from the Contingent Allegiance in that, under the Extended Contingent
Allegiance condition:
•
The drive sends an INITIATE RECOVERY message to the allegiance initiator
prior to returning the failed command’s COMMAND COMPLETE message. If
the initiator rejects the INITIATE RECOVERY message, the drive drops back to
contingent allegiance level of operation for the current error.
• An ABORT message does not terminate the extended contingent allegiance
condition.
• Non-tagged drive access commands from the “allegianced” initiator are accepted
by the drive and take priority over all queued commands. QUEUE FULL status
is returned if the host sends a tagged command.
• The drive delays resuming or aborting the queued commands until a RELEASE
RECOVERY message is received from the initiator.
The extended contingent allegiance condition remains in effect until cleared by one
of the following:
•
Hard reset
• BUS DEVICE RESET message from any initiator
• RELEASE RECOVERY message from the “allegianced” initiator