SIG: Reliability, availability, serviceability (RAS)

In order to promote development of RISC-V in server domain, we need a complete specification to guide implementation of RAS in the design of SoC, firmware and OS.


Tasks in scope include:


RAS terminology interpretation:
Interpretation of RAS concept and terminology (e.g. diagnosability, recoverability, types of error).


RAS framework design:

A framework covers the full path of error handling:

  • Error recording: Standard error record formats (e.g. register banks, APEI­)
  • Error reporting: Error event reporting methods (e.g. exceptions, NMI, local/global interrupts)
  • Error recovery: strategies adopted to handle the error (e.g. neglect/warning/recover/isolation/halt)

RAS feature support:

Engage specific RAS features into the framework:

  • E2E Data protection
  • error isolation
  • data poisoning containment;
  • advanced error reporting for PCIe

Group Information

  • 88 Members
  • 32 Topics, Last Post:
  • Started on
  • Feed

Group Email Addresses

Group Settings

  • This is a subgroup of main.
  • All members can post to the group.
  • Posts to this group do not require approval from the moderators.
  • Messages are set to reply to group and sender.
  • Subscriptions to this group do not require approval from the moderators.
  • Archive is visible to anyone.
  • Wiki is visible to anyone.
  • Members cannot edit their messages.
  • Members can set their subscriptions to no email.

Top Hashtags [See All]

Log In If You Are Already A Member

Message History