| 摘要 |
An improved maintenance system and interface for use in a large-scale data processing system is provided. Within the data processing system, logic is partitioned into multiple logical groups. Each of the groups is associated with a bi-directional port, and each group may be intercoupled to other groups via the associated port. Within each of the groups, an internal condition detection circuit selectively monitors selected logic circuits to detect the presence of certain predetermined conditions or events. When an internal condition is detected, a notification is provided to the bi-directional port and broadcast over the interface to all other logic groups simultaneously. Each logic group further includes an external condition detection circuit interconnected to the respective bi-directional port whereby condition indicators are received from other logic groups as an external condition indicator. Within each logic group, the respective internal indicator and a received external condition indicator are provided to a control unit within the group. The control unit generates control signals which are provided to selected ones of the monitored circuits within the group to perform a predetermined function such as stopping the clock. The configuration allows actions to be taken selectively throughout the system substantially simultaneously and with minimal delay following the condition necessitating the action. Because only a single bi-directional port is needed to interconnect the various logic groups, pin usage is significantly reduced when each group is implemented within a different respective physical device. |
