摘要闪存的读写单位是页,擦除单位是块,一个块包含很多页,要写一个页,就必须先擦除该页所在的块。而闪存的瓶颈就在块擦除操作上,不仅耗时长,更重要的是擦除次数有限。因此闪存多采用异地更新策略,需要维护物理地址和逻辑地址的映射表关系,并且尽可有分散擦除操作到每个块上,以延长使用寿命。作者主要查看各种相关资料,然后对导师给的一个u盘固件代码中的flash管理模块进行剖析,并分析使用的负载均衡方法,简单与资料中的方法进行比较。63868
毕业论文关键词 磨损均衡 闪存 U盘固件代码
毕业设计说明书(论文)外文摘要
Title U disk wear leveling Analysis and Research
Abstract Flash unit is the page to read and write, erase unit is a block, a block contains many pages, write a page, you must first erase the block where the page is. the flash bottleneck is in the block erase operation, not only time-consuming is large, more importantly, the number of erase operations is limited.Therefore, remote update strategy is used,which need to maintain the physical address and logical address mapping table relationships. in order to prolong life,erase operations need to be scattered as possible to each block. Author mainly views many relevant papers, and analyzes a u disk firmware code ,especially the flash management module。Then analyzes the wear-leveling method in the firmware code,try to Compare it with methods which are discribed in relevant papers.
Keywords wear-leveling; flash; u disk firmware code
目 次
1 绪论 1
1.1 背景 1
2 算法综述 2
2.1 随机性磨损均衡算法 2
2.2 确定性磨损均衡算法 2
2.2.1 周期型算法 3
2.2.2 全局型算法 3
3 源代码分析 3
3.1 一些比较重要的表结构描述 4
3.1.1 量产表 4
3.1.2 索引表 4
3.1.3 Zone表 4
3.1.4 缓存块相关表 4
3.1.5 其他一些表或结构 5
2.2 物理读写操作接口的简介 5
2.3 内部基本表的读写管理 5
2.4 一般的坏块处理方法 6
2.5 缓存块的基本管理 7
2.5.1 SBS_GetNotUsed 7
2.5.2 SBS_SetUsed 7
2.5.3 SBS_ClearUsed 7
2.5.4 STO_AddNewUnit 7
2.5.5 STO_WhereIsBlock 8
2.5.6 STO_CalcPhyAddr 8
2.6 地址转换 8
2.6.1 NF_CalcAddrUnits 9
2.6.2 NFI_AddrCalcBP2RC 9
2.6.3 zone表地址映射 11
2.7 逻辑读写的接口