Fault Detection And Safety Design In High Security OS Driver Development

In order to carry out driver development on a high-safety operating system,ensure the functional safety of the driver,meet the requirements of IEC 61508 and EN 50128,and avoid hiding deeper defects,it is necessary to study in time to avoid some difficult observations in the process of operating system driver development Defect technology.In this thesis,by studying the mining and cloning detection of the version history of the code,using software source and evolutionary lineage analysis technology,building a big data environment,and proposing solutions to discover and avoid hidden defects in operating system-driven development.In order to cover the different implementations of operating systems in an all-round way,this thesis studies the history of 28 open source operating systems and dependent projects,involving a total of19.553 M code blobs and a total of 6470 M lines of code.This thesis also uses crawler technology to crawl and build open source forums and open defect libraries for error reporting,build a defect information library,and combine it with code history to help discover hidden defects.This thesis aims to make full use of code information with all-round time and space information.Build a multi-granularity operating system-driven evolutionary pedigree based on large-scale code resources.The constructed operating system driver evolution pedigree will be used to automatically and semi-automatically find similar and related codes of any given operating system driver code,discover its evolution history,and assist code review and analysts in discovering the target operating system driver system.The defect library is updated for a fixed period of time to ensure that newly discovered defects can be found in the defect library.The code library is scanned according to the latest defect library,and hidden defects in the code library are excavated for reference and processing by developers.