Status Detecting And Exception Analyzing Of Sofware On The Board

In the communication system &Equipments for mobile, A large, concurrent software of embedded computer with high reliability usually works for it. Because of the special requirement for working environment of the equipments, When there is a software failure, It usually informs the maintenance people to troubleshoot problems only by the way of distal alarm. Software running environment, however, is usually related to many factors, for example, human factors, environment factors, etc. So it is almost impossible To reproduce the fault time running environment, especially for the real-time systems, even more so. Therefore, it will bring a lot of help for troubleshooting by using a kind of technology, which can record information at the scene of the fault, and can analyze exception. In this paper, the technology of software to detect abnormal state and analyze exception on the board is studied, the main achievements are as follows:1. The current situation of the development of fault diagnosis technology at home and abroad has been Studied, the general method to solve the problem has been mastered. Researched the company laboratory failure and product field failure library, determined the requirement, the data needed to solve the problem has been selected. The data sources include: system scheduling, resource allocation, saving the system itself and reset type recognition, etc, and the data type has been classified, facilitating subsequent expansion2. The use of system memory has been studied, If the board has restarted not because of power down, the data in high memory will not be lost, Setting up the data storage structure in high memory, And the data frame type can be configured, It improves the scalability. for the situation of main controlled board & controlled plate exist at same time, while controlled plate has no local disc. To achieve the controlled plate data can be saved on main controlled board, and main controlled board data can be saved on net-work management center.3.The technology of Abnormal analysis has beed studied, providing warning signs on the situation of high loading of CPU and high temperature of the board, providing the self-recovery mechanism for the situation dead-loop, dead-lock, long term task suspension.For the record of reset, designed the different kinds of site information saving strategy under each reset situation, including commands triggered by high layer, exception and artificial typing. it can parse function call relationship while board resetting, according to the stack ®ister information of Power PC which collected before resetting. To fix the problem rapidly by different types of reset,By using the scheme, It can effectively monitor the state of the board, and rapidly fix problem, improve the testability and maintainability of the equipment.