BoF: Bash the Kernel Maintainers
summary
A bones-up discussion documents concrete failure modes: patches vanish into a black hole for months, maintainers apply wildly different timelines and expectations, and polite pings often fix things while silence does not. Proposed remedies include documenting subsystem response windows, using patchwork to track delegated review, adding automated CI (build/static-analysis/boot tests) for early feedback, tagging drafts as RFC, and organizing short face-to-face triage meetings when reviewers can’t agree. Feedback is collected anonymously for kernel-maintenance summit followup.
AI-written summary · about
chapters (29)
- 0:00Intro
- 3:23Time issues
- 5:56No feedback
- 6:32Documentation
- 7:28Submitting patches
- 8:05Selling patches
- 9:28Getting a microphone
- 10:05No harder availability
- 10:30Losing access to patches
- 11:58Media patches
- 12:27Communication
- 13:22LKML
- 14:12Pink
- 15:51Sock
- 16:17Re resend
- 17:07Patches being forgotten
- 19:01Problems are hard
- 20:16Problems are depending on a problem
- 21:35The maintainer doesnt give a clear view
- 22:30Ive done that
- 23:12Middlegrade solution
- 24:07Static analyzer
- 25:10Dont lose face
- 26:37Make it easier
- 28:33Automate testing
- 29:05Mailing archives
- 32:32RFC
- 34:27Kalika
- 35:12Laurent Pinchart
related talks
47mProject Talk: Maintenance Update
14mThe Power of Backports : Business Impact and Automation Strategies
51mApproaches to Ultra-Long Software Maintenance
23mDeprecating Stuff
09mAspects of Dependable Linux Systems
40mThe Need to Drive Towards Frequent and Sustained Software Updates
30mA Series of Unfortunate Bugs
29mMake It Fixable: Preparing for Security Vulnerability Reports