Measuring and Trends in Linux Kernel Quality

Software quality is a hard topic, especially for large open source projects such as the Linux Kernel, it’s a tough problem. Is each new version really better? It’s often hard to look beyond that one bug that 1 million people hit, or to see beyond the 500 reports that may be caused by defective memory.

The kerneloops.org project tries to quantify and prioritize a serious class of kernel bugs, to find the common patterns and identify the bad-hardware flukes.

With the data from this project and a few other tools, this presentation shows some historic trends on where the Linux kernel was and is going, and has some reflections on what it means to measure the quality of a project this big.