Interface to the exit

The start/stop exit is invoked in task mode, problem state, and key 8 and the job-step task is APF-authorized. The active task runs with the same access authority as the job-step task. The exit must restore this state before returning to its caller.

Control is passed to the exit using the BAL instruction. The exit must return to its caller using the address and addressing mode passed to it in general register 14.

The exit is entered in AMODE 31 but must switch to AMODE 24 before performing any input or output operations, and then switch back to AMODE 31 before returning to the caller.

When the exit is entered, register 1 contains the address of the parameter list. Each address in this list is used to locate the parameter value. These parameters are passed to the exit:

 ACTION    DS   CL8    (Start/stop action)
 MCAUSERF  DS   A      (User field)

ACTION has the value START when the exit is called during Tivoli Workload Scheduler for z/OS start. MCAUSERF is zero for this initial call. Normally, this exit will perform exit initialization functions for the start call when you start Tivoli Workload Scheduler for z/OS. If the exit needs to allocate storage that is used while Tivoli Workload Scheduler for z/OS is active, you should update MCAUSERF to address this storage. ACTION has the value STOP when the exit is called during Tivoli Workload Scheduler for z/OS termination. Normally, this exit performs exit termination functions for the stop call when you stop Tivoli Workload Scheduler for z/OS. If MCAUSERF is updated by the start call, the same value is passed to the exit for the stop call.