gtpm3m0m | Multi-Processor Interconnect Facility Reference |
MPIF performs blocking of data elements received from separate system services in order to maximize I/O efficiency. Prior to invoking the I/O routine (during MPIFC execution or during a write I/O interrupt), MPIF assembles numerous data elements into a single staging buffer. These data elements are removed from a PDT entry's output queue. The I/O routine initiates execution of a channel program to write the contents of the staging buffer to the other processor.
Upon receipt of a read interrupt, the I/O routine places the staging buffer on the ready list or the input list for post interrupt processing. The post interrupt routine recognizes that blocking has been performed. The individual data elements are de-blocked and the appropriate user exits will be scheduled.