Seagate Ultra 160, Ultra 320 manual Queued tasks formerly queued I/O processes, Target hard reset

7.6.7Target hard reset

a SCSI target port hard reset is a SCSI target port response to a SCSI target port Reset task management request (see SCSI Architecture Model-4, T10/1157D), or a reset event within the service delivery subsystem. The definition of target reset events is protocol and interconnect specific. Each SCSI product standard shall specify the response to a SCSI target port reset event including the conditions under which a SCSI target port hard reset shall be executed.

To execute a hard reset, a SCSI target port shall initiate a logical unit reset for all attached logical units as described in Section 7.6.8.

7.6.8Logical unit reset

A logical unit reset is a response to a Logical Unit Reset task management request (see SCSI Architecture Model-4, T10/1157D), or some other logical unit reset event, such as a SCSI target port hard reset (see Sec- tion 7.6.7). The definition of such events may be device-specific or dependent on the protocol and intercon- nect. Each appropriate SCSI standard shall specify the conditions under which a logical unit reset shall be executed.

To execute a logical unit reset the logical unit shall:

a.Abort all tasks in its task set;

b.Clear an auto contingent allegiance condition, if one is present;

c.Release all SCSI device reservations;

d.Return the device’s operating mode to the appropriate initial conditions, similar to those conditions that would be found following device power-on. The mode select conditions (see SCSI Primary Commands- 4 (SPC-4), T10/1731-D) shall be restored to their saved values if saved values have been established. Mode select conditions for which no saved values have been established shall be returned to their default values;

e.Set a Unit Attention condition (see Section 7.6.6); and

f.Initiate a logical unit reset for all nested logical units.

In addition to the above, the logical unit shall execute any additional functions required by the applicable stan- dards.

7.7Queued tasks (formerly “queued I/O processes”)

Queuing of tasks allows a drive to accept multiple commands for execution at a later time.

There are two methods for implementation of queuing: tagged and untagged. Tagged task queuing allows the drive to accept multiple commands from each initiator. Untagged task queuing allows the drive to accept one command from each initiator. Drives that have SCSI-2 implementation support tagged queuing while in SCSI-1 or SCSI-2 mode. They can use untagged task queuing mode if the initiator does not send task queue tag mes- sages.

Initiators may add or delete commands to the queue for the drive within the limitations specified in this specifi- cation. When adding a command, the initiator may specify fixed order of execution, allow the drive to define the order of execution, or specify that the command is to be executed next. See glossary in Section 1.2.1 for termi- nology definitions when reading the following explanations.

7.7.1Untagged task queuing

Untagged task queuing allows the drive to accept a command from a SCSI initiator port while a command from another initiator is being executed. Only one command for each I T L nexus may be accepted at a time.

A new task may be initiated any time the BUS FREE phase exists even if another task from a different initiator is being executed. If the disconnect privilege is not granted, the drive returns Busy status to the new task.

Parallel SCSI Interface Product Manual, Rev. A 147