基于AER的PCI Express总线故障注入工具研究
发布时间：2018-07-23 17:40
【摘要】：随着信息化的进程，计算机系统越来越深入地融入了国民经济的各个领域，人们对计算机系统的依赖程度与日俱增。尤其是军事、金融、电信、航空航天等国民经济关键领域，不仅需要计算机系统拥有强大的计算能力，更需要其具有极高的可靠性。 PCI Express是由英特尔提出的最新的总线和接口标准，它代表了第三代I/O接口标准。它的主要优势是数据传输速率高，而且还有相当大的发展空间，另外带宽选择灵活性高，适应性强。PCI Express总线在现代计算机，特别是高端容错计算机的I/O系统中被广泛采用。高端容错计算机系统提供的网络服务、存储服务等都依赖于PCI Express总线，PCI Express总线的可靠性是高端容错计算机系统提供可靠服务的非常重要的一个环节。 为了评测计算机的PCI Express容错机制的可靠性，本文研究了基于硬件的PCI Express总线故障注入工具的工作原理，设计并实现了PCI Express总线测试套件。使用该测试套件对某国产高端容错计算机进行了测试，为该计算机系统提供了多项改进建议。 另外，本文通过研究PCI Express总线的故障检测和故障处理机制，设计并实现了基于Linux内核AER驱动的PCI Express总线故障注入工具，并在惠普Proliant ML150G6服务器上进行了测试，验证了故障注入工具的有效性，并根据测试结果对该服务器的PCI Express总线软件层的容错机制进行了分析。该工具的实现，弥补了基于硬件的PCI Express总线故障注入工具仅适于评测PCI Express硬件层容错机制的不足，两者相互配合，能够全面的评测计算机系统软硬件对PCI Express总线故障的容错能力。
[Abstract]:With the development of information technology, computer system is more and more integrated into all fields of national economy, and people rely on computer system day by day. Especially in the key areas of the national economy, such as military, finance, telecommunications, aviation and aerospace, not only must computer systems have strong computing power, The most reliable. PCI Express is the latest bus and interface standard proposed by Intel, which represents the third generation I / O interface standard. Its main advantage is that the data transmission rate is high, and there is still considerable room for development. In addition, the bandwidth selection flexibility is high, and the adaptability of .PCI Express bus is widely used in modern computers, especially in the I / O systems of high-end fault-tolerant computers. The network service and storage service provided by high-end fault-tolerant computer system depend on the reliability of PCI Express bus. It is a very important link for high-end fault-tolerant computer system to provide reliable service. In order to evaluate the reliability of PCI Express fault-tolerant mechanism of computer, this paper studies the working principle of PCI Express bus fault injection tool based on hardware, and designs and implements PCI Express bus test suite. The test suite is used to test a high-end fault-tolerant computer made in China, which provides several suggestions for improvement of the computer system. In addition, this paper designs and implements the PCI Express bus fault injection tool based on Linux kernel AER driver by studying the fault detection and fault handling mechanism of PCI Express bus, and tests it on HP Proliant ML150G6 server. The validity of the fault injection tool is verified and the fault-tolerant mechanism of the PCI Express bus software layer of the server is analyzed according to the test results. The implementation of this tool makes up for the deficiency that the hardware-based PCI Express bus fault injection tool is only suitable for evaluating the fault-tolerant mechanism of the PCI Express hardware layer, and the two tools cooperate with each other. It can comprehensively evaluate the fault tolerance of computer system software and hardware to PCI Express bus fault.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP302.8
本文编号：2140148
[Abstract]:With the development of information technology, computer system is more and more integrated into all fields of national economy, and people rely on computer system day by day. Especially in the key areas of the national economy, such as military, finance, telecommunications, aviation and aerospace, not only must computer systems have strong computing power, The most reliable. PCI Express is the latest bus and interface standard proposed by Intel, which represents the third generation I / O interface standard. Its main advantage is that the data transmission rate is high, and there is still considerable room for development. In addition, the bandwidth selection flexibility is high, and the adaptability of .PCI Express bus is widely used in modern computers, especially in the I / O systems of high-end fault-tolerant computers. The network service and storage service provided by high-end fault-tolerant computer system depend on the reliability of PCI Express bus. It is a very important link for high-end fault-tolerant computer system to provide reliable service. In order to evaluate the reliability of PCI Express fault-tolerant mechanism of computer, this paper studies the working principle of PCI Express bus fault injection tool based on hardware, and designs and implements PCI Express bus test suite. The test suite is used to test a high-end fault-tolerant computer made in China, which provides several suggestions for improvement of the computer system. In addition, this paper designs and implements the PCI Express bus fault injection tool based on Linux kernel AER driver by studying the fault detection and fault handling mechanism of PCI Express bus, and tests it on HP Proliant ML150G6 server. The validity of the fault injection tool is verified and the fault-tolerant mechanism of the PCI Express bus software layer of the server is analyzed according to the test results. The implementation of this tool makes up for the deficiency that the hardware-based PCI Express bus fault injection tool is only suitable for evaluating the fault-tolerant mechanism of the PCI Express hardware layer, and the two tools cooperate with each other. It can comprehensively evaluate the fault tolerance of computer system software and hardware to PCI Express bus fault.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP302.8
【参考文献】
相关期刊论文 前10条
1 刘福华;;容错技术与自动变速控制[J];重庆科技学院学报(自然科学版);2007年02期
2 李洪超;;计算机系统的容错技术方法[J];单片机与嵌入式系统应用;2010年11期
3 徐和平;何强;;联锁控制系统中的容错技术[J];川化;2005年02期
4 胡谋;;计算机容错技术及其展望[J];电子技术;1992年08期
5 李明;李霞;周轶男;;基于MPC8548嵌入式系统PCI-Express设备驱动开发[J];电子技术;2011年01期
6 张伟栋;刘志敏;;基于PowerPC8640的PCI-E总线接口设计[J];航空计算技术;2011年03期
7 陆阳;王强;张本宏;诸葛战斌;;计算机系统容错技术研究[J];计算机工程;2010年13期
8 宁新建;;航空火控计算机的容错技术分析[J];计算机与网络;2010年17期
9 王莲;徐萍;刘斌;;BIT验证中1553B总线故障注入设备的设计与实现[J];计算机测量与控制;2011年12期
10 刘逢清;;容错计算机系统的可靠性建模和分布式仿真[J];南京邮电大学学报(自然科学版);2008年05期
，本文编号：2140148
本文链接：https://www.wllwen.com/kejilunwen/jisuanjikexuelunwen/2140148.html
