Single Point Fault Detection

Versal Adaptive SoC Technical Reference Manual (AM011)

Document ID
AM011
Release Date
2023-10-05
Revision
1.6 English

The detection of single point faults is supported with these features:

  • ECC protection for OCM, PPU RAM, PMC RAM, RPU L1 cache, and TCM memories
    • Address decode error detection
    • Separate RAMs for ECC syndrome and data
    • 4:1 or greater interleaving of memory cells protected by ECC
  • Hash validation of RCU ROM contents at every boot
  • Lockstep and redundancy covers Cortex-R5F processor
    • Lockstep with physical and temporal diversity
    • Redundant logic in critical control logic including the lockstep checkers
  • PPU and RCU controllers are implemented with redundancy
    • MicroBlazeâ„¢ triple modular redundancy (TMR) cores with physical diversity
    • Triple redundant flip-flops for critical control bits such as security state
  • XMPU and XPPU protect memory space
  • Windowed watchdog timers in LPD and FPD