Skip to content
This repository has been archived by the owner on Nov 1, 2023. It is now read-only.

2020.11.16: Working Group Meeting

Jump to bottom
Al Stone edited this page Nov 20, 2020 · 3 revisions

Profiles and Platform Spec Technical Subcommittee (P2TS) Agenda: 16 Nov 2020

Meeting Info: weekly on Mondays (09:00 PST, 16:00 UTC)

Passthrough:

  • Anti-trust statement
  • Open and collaborative

In Process:

  • Proposed scope/charter
  • Reusing the PSCI spec: moving this to far background; this is now completely dependent on ARM for further progress; lack of public governance is the primary concern
  • Update on hypervisor work: recommendations for reuse in platform spec?
    • There are likely parts that will be referred to as required
    • Software needs may be different from ISA considerations, so we do need to pay attention to both sides
  • SBI updates:
    • Anything that can be merged has been
    • Reset and PMU extensions pending approval
  • Updates from any other groups present? Nothing volunteered

Topics

  • Tech chairs meeting conflict (2nd and 4th Wednesday needed)
    • Vice chair will attend these (thanks!)
  • Volunteers:
    • to provide updates? AIC? CMO? Config? Others? (new)
    • to investigate subsections? (new)
      • Firmware? DT? ACPI? AIC? CMO? Config? IOMMU? RV64A? Boot protocol?
      • Serial console? Buses? SMBIOS? IPMI?
    • These are just suggested topics; the ask is for those that are interested to step up and start thinking about what we need to specify, if anything for these topics
    • Not surprisingly, everyone has far too much to do so there were no immediate volunteers
  • TSC Request (revisit)
  • SBI Reset extension (revisit)
    • discussion: https://lists.riscv.org/g/tech-unixplatformspec/message/430
    • Key point is that this is now approved and can be merged using the latest patches sent to the list. Hooray! Progress has been made :).
    • Topic veered off the road a bit, to discuss where we're going with the spec and the meaning of "done".
      • the first question: given that we are specifying the requirements for a platform -- that may or may not even exist yet -- what sort of PoC (Proof of Concept) is useful and/or required? The gist of this was that we will most definitely need a PoC to show the spec can actually be implemented; it may end up as simulator or emulator code but without a PoC, we will have to ask the TSC for exceptions.
      • second, we need to be careful about what we specify; on the one hand we need a spec that actually molds what people produce so that, for example, when someone says a board supports the Linux spec, it actually does so without a lot of hand waving and calisthenics. But we don't want the spec so loose that anything qualifies. At the other end of the spectrum, we can't have the spec so specific that only mega-corporations can afford the investment in IP (e.g., with PCIe). We know a priori that there will be a broad range of products developed for the architecture over time, so let's have the spec allow for both the low cost hobbyist and the high end data center servers -- and not all at once, but let's mature with the market. These seemed to be generally agreed sentiments and is indeed the overall plan.
  • SBI PMU extension (revisit)
  • Proposal: Hart Suspend Extension for IDLE (revisit)
    • discussion: https://lists.riscv.org/g/tech-unixplatformspec/message/274
    • This proposal is really part of the SBI reset extension proposal
    • It would be very good in the future if the authors of proposals could make themselves available to make an initial presentation of their proposal to this subcommittee, and allow for some discussion and Q&A. The chair would like to see this in addition to the mailing list.
    • In this case, we will continue any discussion on the mailing list. There were some suggestions made, to be mentioned there.

Open Floor

  • Nothing raised.
Clone this wiki locally