Reliability, Availability, and Serviceability (RAS) on ARM64 status
summary
Status update on ARM64 RAS and firmware integration: ARMv8 RAS extensions plus SDEI, APEI and a secure-partition prototype provide a firmware-first error flow that collects registers, creates a CPI table and drops error records into nonsecure memory for kernel pickup. A prototype super-library and inj table (fault-injection) run on QEMU; CPI protocol and kernel boot driver are upstream. Work in progress: move to SPSA QEMU, add BRT persistent crash storage, implement common fault-injection registers and external interrupt (TFE) routing.
AI-written summary · about
chapters (10)
- 0:00Introduction to RAS on ARM64
- 2:19RAS architecture overview
- 3:26RAS extension components
- 7:12RAS 1.1 new features
- 9:47RAS extension 1.1.0 proposals
- 13:05Security partition handling
- 16:41BRT table for crash recovery
- 19:14Error injection modeling
- 20:20QEMU prototype progress
- 22:26Next steps and future work
