Deterministic Assurance Technology Of Embedded Multi-Core Operating System

In the process of the processor's development from single core to multi-core,on the one hand,the performance of device processing is improved,but on the other hand,the use of multi-core also makes the system compete on shared resources such as buses and memory,and tasks are executed concurrently on multiple cores.A series of problems such as the uncertainty of the operating system have a great impact on the reliability,safety,and real-time performance of the equipment.The deterministic guarantee technology of the operating system poses a huge challenge on multi-core.At present,the software design still stays in the use of single-core.For example,in the aviation field,designers often turn off other processor cores when running safety-critical tasks.Retaining a nuclear operation safety-critical mission to prevent interference from other cores,the design and use of multiple cores is still in the exploratory research stage.In safety-oriented operating systems,both the ARINC 653 standard and the D0-178 C standard require the operating system to have the concept of partitions.Partitions include both temporal and spatial partitions to ensure mutual isolation between applications.This article introduces in detail the influencing factors of multi-core interference,and proposes a variety of multi-core deterministic guarantee strategies and mechanisms,including multi-core interference mitigation,multi-core real-time operating system architecture,multi-core real-time scheduling strategy,multi-core deterministic interrupt,and multi-core deterministic monitoring And other mechanisms.Finally,on the domestic Ruihua operating system,the single-core multi-partition two-level scheduling mechanism that meets the ARINC653 standard has been expanded into a multi-core and multi-partition two-level scheduling mechanism,which innovatively supports multiple applications of AMP,SMP,and BMP.Hybrid scheduling method of processing architecture.In the multi-core and multi-partition two-level scheduling mechanism,the first level of scheduling is based on the static scheduling table among partitions,and the second level of scheduling is based on priority preemption.The partition occupies one or more time windows in the main time frame of the periodic rotation defined by the system.The partition can only be scheduled within the allocated time window,which ensures the certainty of the running time of the partition.The accuracy and predictability of the system is guaranteed by the first-level inter-zone scheduling,and the real-time performance of the second-level intra-zone scheduling guarantees the determinism of the multi-core operating system,and the correct implementation of the two-level scheduling mechanism has been proved through experiments.