RISC-V Linux发行版软件包构建问题实证研究

An empirical study on build issues of software packages for RISC-V Linux distributions

RISC-V是一种新兴的精简指令集架构,将软件包源码在RISC-V架构下重新构建,是Linux发行版向RISC-V架构迁移过程中的重要工作。为了帮助开发者更快地理解和解决这些问题,对521个Debian发行版软件包近2年在RISC-V架构下的构建失败,以及219个架构相关案例进行了实证研究,对构建失败的根本原因进行了深入分析,得到了7种架构相关问题的修复模式,对开发者解决构建问题提供指导,同时提出了一个基于机器学习的RISC-V Linux发行版软件包构建问题的分类框架,以辅助架构相关问题的解决,通过实验验证了本文方法的有效性。

RISC-V is an emerging reduced instruction set computing(RISC)architecture.One important task during the porting of Linux distributions to RISC-V architecture is rebuilding the source code of software packages for the RISC-V architecture.An empirical study was conducted on 521 Linux software packages that failed to build on the RISC-V architecture in the Debian distribution over the past two years,along with 219 architecture-specific cases,to help developers quickly understand and address these issues.Through a thorough analysis of the root causes of these build failures,seven fix patterns for architecture-specific issues were identified,providing guidance for developers in resolving build issues.Furthermore,a machine learning-based classification framework for RISC-V Linux distribution software package build issues was proposed to assist in resolving architecture-related issues.Experimental validation was conducted to verify the effectiveness of the approach.