Kfeature: Rendering the Kconfig System into Feature Models
Kaan Berk Yaman, Jan Willem Wittler, Christopher Gerking · 2024
Kconfig is a prominent configuration system that helps manage variability of software such as the Linux kernel. To make such software accessible to feature-oriented analysis, it is desirable to transform Kconfig into equivalent feature models. In this paper, we address the problem that existing transformations do either not cover an essential part of the Kconfig language constructs, or fail to produce human-readable feature models. We address both limitations by drawing up practical guidelines for the transformation, and enable feature-oriented analysis in terms of Kfeature as a tool interface of the FeatureIDE framework. To evaluate the semantic equivalence, we apply Kfeature to synthesized variability definitions.