Stuck in 2009 - How I Survived
summary
Product ships on a five‑year‑old 2.6.28/2.6.29 kernel so teams spend months cherry‑picking fixes instead of upgrading; claim is that total backport work roughly equals porting to mainline. Tools used include git cherry-pick, nm, the backports project and a recompiled toolchain (uClibc/NPTL issues). Specific fixes: refactor ESOP codec into modular parts, massage WL127x SDIO driver (partial), and avoid crypto features added around 3.15. Outcome: project stays on the old kernel, with many fixes working but some patches failing to apply.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 4:20A fond look back
- 6:42The Clash
- 7:10How Much Effort
- 9:37Is It Vanilla
- 11:24Benefits
- 13:20External Peripherals
- 15:42Crypto
- 18:32ESOL codec
- 23:04Unexpected benefits
- 25:02Conclusions
- 27:18Questions
related talks
34mMaintaining a Linux Kernel for 13 Years? You Must be Kidding Me. We Need at Least 30?
18mBackporting Linux Mainline Drivers
39mThe End of the Paved Road: Maintaining Linux Kernel 4.4 Beyond LTS
40mLearnings from Upgrading an Outdated Yocto-Based System
01h 15mCurrent Developments in the Linux Kernel
14mThe Power of Backports : Business Impact and Automation Strategies
40mSurviving 19 Jan 2038 on 32 Bit Platforms: Lessons Learned and Common Problems
27mAn SRE guide to Linux Kernel upgrades