NVIDIA Approach for Achieving ASIL B Qualified Linux: minimizing expectations from up...
summary
NVIDIA claims a path to ASIL B by keeping most kernel code unqualified and concentrating verification on a small set of mechanisms that can be merged or compiled off. It implements per-thread capabilities and memory pools by address ranges with per-thread per-core MMU maps and redundant PFN bitmap checks to confirm pool membership rather than proving the allocator correct. Patches live mostly out-of-tree as extra files plus light tags and hooks, stack hardening breaks ftrace today but rebases to 6.18 succeed and the design is deliberately non‑invasive.
AI-written summary · about
chapters (7)
- 0:00Introduction to ASIL B qualification
- 2:28Safety qualification challenges in Linux
- 5:15Thread capabilities and memory pools
- 9:11Resource partitioning for safety
- 13:10End-to-end memory verification
- 16:50Minimal kernel impact design
- 19:19Upstreaming benefits and future
related talks
41mIs Linux Kernel Development Good Enough to Make Your Life Depend on It?
21mDriving Innovation: Evolving the Role of Software in the RISC-V Ecosystem Beyond...
42mDecoding Safe(ty) Linux Architectural Approaches for Critical Systems
57mPanel Discussion: Super Long Term Kernel Maintenance
18mThe Significance of the RVA23 Profile in Advancing RISC-V Ecosystem
32mCommSec D2 - ProdSec: A Technical Approach
42mZephyr Functional Safety BOF
53mBridging Safety and Architecture: Zephyr’s Path To IEC 61508 Complian...