Memory Management And Optimization In WebitOS

While the embedded technology becoming perfect, embedded system has being applied in various fields. As the kernel of its software, the Embedded Operating System is the hotspot research in embedded field. Although embedded system resource is limit, memory management is one of the indispensable parts. In order to effectively utilize the memory space, choosing a suitable allocation and collection of memory algorithm is so important to improving embedded system performance.The main parts is as following. At first, the previous memory management in WebitOS is analyzed. After researching, this thesis finds that the primary allocation algorithm is best-fit algorithm, but the memory management doesn't include memory defragment. When a thread releases its memory space, the collection inserts the free node into the free list. But it doesn't pack up the memory space. In this way, after allocating and collecting for a long time, the memory space contains many fragments, which will affect the utilization of the memory space. In order to compensate the above faults, memory defragment function is added. The previous allocation algorithm may not fit for original system any more for its adding of collection algorithm. So, based on the characteristic of the system, another two allocation algorithms are put forward, which are first-fit algorithm and next-fit algorithm. Two application programs test.c and httpserver.c are compiled and implemented. Using this two application programs, the modified collection algorithm is compared with previous collection algorithm, the result confirms that memory defragment function is necessary. In addition, the three algorithms' validity and response time are tested. The result shows that the first-fit algorithm is better than the others. At last, the three allocation algorithms and the modified collection are tested separately, which demonstrated that the memory...