A reconfigurable fault-tolerant startup method for multi-core operating systems

Abstract

A reconfigurable fault-tolerant startup method for a multi-core operating system, which includes setting a synchronization signal as an initial state and setting a spin lock as unoccupied; each processor core applies for a spin lock after completing initialization; the processor core that has applied for the spin lock Enter the critical area of ​​mutually exclusive access between cores, and the rest of the processor cores are blocked on the spin lock; the first processor core that applies for the spin lock sets its processor core ID as the main core ID number, and sets the synchronization Signal, and then release the spin lock; the processor core that subsequently applied for the spin lock knows that it is a slave core through the read value, and the processor core corresponding to the master core ID is the master core, and releases the spin lock; the master core After completing the initialization of the memory, bus device and operating system kernel, the slave core is started by setting the inter-core synchronization signal, so that the master core and the slave core start task scheduling at the same time. The invention can still normally start other processor cores to run when the main core fails, thereby improving the reliability of the multi-core computer.

Description

technical field [0001] The invention relates to a reconfigurable fault-tolerant startup method for a multi-core operating system, in particular to a reconfigurable fault-tolerant startup method for a multi-core operating system based on shared memory. Background technique [0002] The start-up process of the embedded multi-core processor is mainly completed by the main processor core (hereinafter referred to as the main core). After the processor is powered on and reset, the master core first makes the slave cores enter the sleep or waiting mode of operation, the master core completes the initialization of the memory, bus devices and operating system kernel, and then wakes up the rest of the slave cores, and finally enters the normal multi-core operating mode. Schedule tasks to run in parallel. [0003] In traditional embedded multi-core operating systems, the selection of the main core is usually pre-specified manually. When the main core fails to start, it renders the en...

AI Technical Summary

Problems solved by technology

[0007] The technical problem solved by the present invention is: to overcome the deficiencies in the prior art, to provide a multi-core operating system reconfigurable fault-tolerant startup method, which solves the problem that the master core fails to start the slave core in the traditional startup method, and can It can still start other processor cores to run normally, which improves the reliability of multi-core computers

Images