Parallelism Safety-Critical Guidelines for C++
summary
A working group produces concrete, checkable rules for parallelism safety: a public spreadsheet and phase‑one analysis yield 24 rules ready with MISRA classifications (about eight mandatory rules and mixed tool vs human decidability). Rules target specific features: cooperative cancellation J threads, coroutine capture hazards, atomic_ref use, semaphores, latches/barriers, RCU/TS2 reclamation and std::async launch policy. Proposals aim to land in MISRA NEXT and C++ Core Guidelines; many rules are written to be automatable by existing static analysis tools.
AI-written summary · about
chapters (7)
- 0:00Parallelism safety guidelines intro
- 10:50Deadlocks in C++ core rules
- 19:53Lifetime violations and safety
- 33:24MISRA vs C++ Core Guidelines
- 42:20Future C++ parallelism features
- 57:16MISRA vs C++ Core Guidelines differences
- 1:00:06Summary and next steps
related talks
58mEliminating Hidden Dangers to Your Applications Patterns for Reliable Systems
28mAddressing the parallelism gap in replicated server systems
45m(...)Distributed Systems
39mFormal Modeling and Analysis of Distributed Systems
01h 03mThe Power of the Paradigm
42mFive Common Mistakes With Distributed Systems
25mConsistent Updates for Scalable Microservices
01h 00mDIY Async Message Pump: Lessons from the trenches