软件实现的DSP内存保护技术

发布时间：2018-05-26 18:35

本文选题：软加固 + EDAC　；参考：《国防科学技术大学》2014年硕士论文

【摘要】：在许多高可靠性系统中,设计者会通过硬件方式对内存中的数据进行编码保护,其实现方式主要是增加额外的数据校验位以及编码和解码电路。然而,在某些空间电子设备上并不具备上述硬件条件,这时就需要通过软件实现的数据编码方法对内存中的数据进行保护。软件实现的数据编码保护方法省去了额外的硬件成本以及功耗开销,但由于其需要占用设备的计算资源,在使用时应控制好其带来的时间开销,以免造成过于严重的性能损失。在空间开销方面,对数据进行编码产生的校验位的冗余度远小于对数据进行复算的冗余度,因此数据编码保护方式十分适合对空间开销限制比较苛刻或者硬件条件受限的情况。内存中的代码和数据对程序正确运行有至关重要的作用,因此有必要在运行时对内存中的关键数据进行检查。采用编码校验的方法对数据进行检查,具有开销小、检错快等特点,适合在实时内核中使用。但由于编码方法的运行时间受限制于读取内存数据的速度,当数据量较大时,执行周期较长,在内核中完成编码检查对操作系统的实时性会造成严重影响。本课题提出了一种低成本高可靠性的DSP内存保护方案,即在较低时间开销和空间开销下对DSP上的内存数据进行检错和纠错。具体地,本课题采用了基于(39,32)最优奇权码的编码校验方案,并针对DSP的体系结构进行了线性汇编级的高效实现,然后提出了一种基于实时操作系统的代码区和堆栈区编码校验机制,即通过加入低优先级的校验任务,利用系统的空闲时间对代码区及堆栈区进行编码和校验。通过实验结果分析可知,该方案不会对原系统的实时性产生影响,且能够使系统的可靠性得到显著提升。
[Abstract]:In many high reliability systems, the designer encodes and protects the data in memory by hardware, mainly by adding additional data check bits and encoding and decoding circuits. However, the above hardware conditions are not available in some space electronic devices, and the data in memory need to be protected by the data encoding method implemented by software. The data coding and protection method realized by software saves the extra hardware cost and power cost, but because it needs to occupy the computing resources of the equipment, the time cost should be controlled well when it is used, so as to avoid too serious performance loss. In terms of space overhead, the redundancy of the check bits generated by encoding the data is much less than the redundancy of the complex data, so the data coding protection method is very suitable for the case where the space overhead is very strict or the hardware condition is limited. The code and data in memory is very important for the program to run correctly, so it is necessary to check the critical data in memory at run time. The method of coding check is used to check the data, which has the advantages of low overhead and fast error detection, and is suitable for use in the real time kernel. However, because the running time of the coding method is limited to the speed of reading the memory data, when the amount of data is large, the execution period is longer, and the completion of coding checking in the kernel will seriously affect the real-time performance of the operating system. In this paper, a low cost and high reliability DSP memory protection scheme is proposed, that is, the memory data on DSP is checked and corrected under low time and space overhead. In detail, the scheme of coding check based on the optimal odd weight code is adopted, and the efficient implementation of linear assembly level for the architecture of DSP is carried out. Then, a code and stack coding verification mechanism based on real-time operating system is proposed, which encodes and verifies the code area and stack area by adding low-priority checkout tasks and making use of the idle time of the system. The experimental results show that the proposed scheme has no effect on the real-time performance of the original system and can improve the reliability of the system significantly.
【学位授予单位】：国防科学技术大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP333

【相似文献】