Linux Kernel Testing: Providing Better Results to the Upstream Community - Laura Nao, Collabora

Linux Kernel Testing: Providing Better Results to the Upstream Community - Laura Nao, Collabora

Automated testing can be an invaluable asset for the Linux Kernel developers and maintainers, aiding in both the development process and in the detection and tracking of regressions. However, its effectiveness depends on the quality of tests and the results provided. The path to make CIs more effective and valuable for the upstream community is long and winding. To build greater trust in CI systems, it's crucial that tests are crafted to be both reliable and maintainable over time. This includes use of stable ABI and standardized output formats. Engaging the community in the process is paramount, as we need to maintain such test for years to come. In this presentation, we will review the key challenges faced by the main test frameworks and present our contributions and ongoing efforts to improve upstream testing, from device probe regression tracking to kselftest standardization. Through some firsthand examples, we will explore how to address current shortcomings in upstream testing and improve the quality and dependability of the existing tests.