Memory Maganement Design And Realize Of Embedded Linux System

With the rapid development of multimedia technology and the Internet, and the enhancement of intelligence for electronic products, embedded systems are more extensive and in-depth applications. Networks everywhere and computing everywhere is bringing human being into a brand-new information society. The safety and reliability of embedded system are crucial in key field such as spaceflight, industry control and medicine. In order to improve the safety and reliability of data storage in embedded system, research on memory management for embedded system is particularly significant.In this paper we carry out a deep and detailed analysis upon memory management mechanisms of embedded linux system, and on the base of it proposing an improvement algorithm of buddy system, and taking some corresponding tests upon the improved performance. Specifically in the two aspects of physic memory maganement and virture memory maganement analyzing embedded linux memory management strategy.There are four proberms in the current buddy system: address-countinous block may not merge, external fragementation, inernal fragementation and efficiency issue.At last proposed based on need and merge-delay buddy system and do some test.This paper consists of five chapters. The first chapter summarizes embedded system briefly, including the presently developing status and researching status of embedded linux system. And then interprets level-classified structure of memory and memory management function. The second chapter elaborates linux memory management mechanism and strategy. Linux memory management is analyzed in the phase of physical memory management and virtual memory management. The third chapter analyzing the physic memory maganement algorithm - buddy system.The process involves the principles of the buddy algorithm, the process of allocation and free and bitmap operations.On this basis, point out the deficience of the current buddy system and propose improvement programme. The forth chapter realize this improved buddy algorithm, consequently tests the performance of the algorithm. The fifth chapter sums up the whole paper.