Seagate Ultra 320 ユーザーズマニュアル
Parallel SCSI Interface Product Manual, Rev. A
143
The handling of tasks created by initiators other than the faulted initiator depends on the value in the TST field
in the Control mode page (see ANSI SCSI Primary Commands-4, T10/1731-D).
in the Control mode page (see ANSI SCSI Primary Commands-4, T10/1731-D).
If TST=000b, tasks created by other initiators while the ACA or CA condition is in effect shall not be entered
into the faulted task set (except for a PERSISTENT RESERVE command with a Preempt and Clear action as
described in Section 7.2.1, note [7], Table 62). Tasks rejected from the task set due to the presence of an ACA
or CA condition shall be completed with a status of ACA Active (if NACA=1 in the new command’s CDB Control
byte, see Section 7.2.1, note [7], Table 62) or Busy (if NACA=0).
into the faulted task set (except for a PERSISTENT RESERVE command with a Preempt and Clear action as
described in Section 7.2.1, note [7], Table 62). Tasks rejected from the task set due to the presence of an ACA
or CA condition shall be completed with a status of ACA Active (if NACA=1 in the new command’s CDB Control
byte, see Section 7.2.1, note [7], Table 62) or Busy (if NACA=0).
If TST=001b, tasks created by one initiator shall not be rejected based on an ACA or CA condition in effect for
another initiator. Only ACA or CA condition for the sending initiator (as well as other task set management con-
siderations described in SCSI Architecture Model-4, T10/1157D, clause 7, “Task Set Management.”) shall
affect acceptance into the task set or rejection for a task from that initiator.
another initiator. Only ACA or CA condition for the sending initiator (as well as other task set management con-
siderations described in SCSI Architecture Model-4, T10/1157D, clause 7, “Task Set Management.”) shall
affect acceptance into the task set or rejection for a task from that initiator.
7.6.1.2
Clearing an Auto Contingent Allegiance condition
If the NACA bit is set to zero in the Control byte of the faulting command, then the SCSI-2 rules for clearing
contingent allegiance shall apply. In addition, the logical unit shall clear the associated contingent allegiance
condition upon sending sense data by means of the autosense mechanism described in Section 7.6.4.2.
contingent allegiance shall apply. In addition, the logical unit shall clear the associated contingent allegiance
condition upon sending sense data by means of the autosense mechanism described in Section 7.6.4.2.
While the SCSI rules for clearing the ACA condition are in effect, a logical unit that supports the Clear ACA task
management function shall ignore all Clear ACA requests and shall return a service response of Function
Complete (see SAM-4).
management function shall ignore all Clear ACA requests and shall return a service response of Function
Complete (see SAM-4).
If the logical unit accepts a value of one for the NACA bit and this bit was set to one in the Control byte of the
faulting command, then the SCSI-2 rules for clearing a contingent allegiance condition shall not apply. In this
case, the ACA condition shall only be cleared:
faulting command, then the SCSI-2 rules for clearing a contingent allegiance condition shall not apply. In this
case, the ACA condition shall only be cleared:
a. As the result of a power on or a logical unit reset (see ANSI SAM-4 document T10/1683-D);
b. Through a Clear ACA task management function issued by the faulting initiator as described in ANSI
b. Through a Clear ACA task management function issued by the faulting initiator as described in ANSI
SAM-4 document T10/1683-D;
c. Through a Preempt and Clear action of a PERSISTENT RESERVE OUT command that clears the tasks
of the faulting initiator (see the ANSI SCSI Primary Commands-4);
d. A command with the ACA attribute terminates with a CHECK CONDITION status. The state of all tasks
in the task set when an auto contingent allegiance condition is cleared shall be modified as described in
ANSI document T10/1157D, clause 7, “Task Set Management.”
ANSI document T10/1157D, clause 7, “Task Set Management.”
7.6.2
Overlapped commands
An overlapped command occurs when an application client reuses a Task Address (see Glossary) in a new
command before a previous task to which that address was assigned completes its task lifetime as described
in SCSI Architecture Model-4 (SAM-4), “Task and Command Lifetimes.”
command before a previous task to which that address was assigned completes its task lifetime as described
in SCSI Architecture Model-4 (SAM-4), “Task and Command Lifetimes.”
Each system that implements a SCSI protocol standard shall specify whether or not a logical unit is required to
detect overlapped commands. A logical unit that detects an overlapped command shall abort all tasks for the
initiator in the task set and shall return CHECK CONDITION status for that command. If the overlapped com-
mand condition was caused by an untagged task or a tagged task with a tag value exceeding FFh, then the
sense key shall be set to Aborted Command and the additional sense code shall be set to Overlapped Com-
mands Attempted. Otherwise, an additional sense code of Tagged Overlapped Tasks shall be returned with the
additional sense code qualifier byte set to the value of the duplicate tag.
detect overlapped commands. A logical unit that detects an overlapped command shall abort all tasks for the
initiator in the task set and shall return CHECK CONDITION status for that command. If the overlapped com-
mand condition was caused by an untagged task or a tagged task with a tag value exceeding FFh, then the
sense key shall be set to Aborted Command and the additional sense code shall be set to Overlapped Com-
mands Attempted. Otherwise, an additional sense code of Tagged Overlapped Tasks shall be returned with the
additional sense code qualifier byte set to the value of the duplicate tag.
Note.
An overlapped command may be indicative of a serious error and, if not detected, could result in cor-
rupted data. This is considered a catastrophic failure on the part of the initiator. Therefore, vendor-spe-
cific error recovery procedures may be required to guarantee the data integrity on the medium. The
target logical unit may return additional sense data to aid in this error recovery procedure (e.g.,
sequential-access devices may return the residue of blocks remaining to be written or read at the time
the second command was received).
rupted data. This is considered a catastrophic failure on the part of the initiator. Therefore, vendor-spe-
cific error recovery procedures may be required to guarantee the data integrity on the medium. The
target logical unit may return additional sense data to aid in this error recovery procedure (e.g.,
sequential-access devices may return the residue of blocks remaining to be written or read at the time
the second command was received).