Minutes 14 Nov 2024

Host: Paul Albertella

Participants: Daniel Krippner, Pete Brink, Igor Stoppa, Gabriele Paoloni

Agenda

• Publish previous workshop presentations
• Review and publish 'Contributions' already in OSEP repo
• Review status of PRs

Discussion

Add author and contributors to the published document

• Can we automate this (e.g. with a plugin)

include other metadata?

What policy should we have on when material is published?

• Currently publishes on merge to main
• Could publish on tags / releases only

Add the names of the approvers

• Could we automate this?

Should we ask Min to link to this from the main Elisa site?

Review criteria: Completeness, correctness and comprehensibility

• Peer review by someone with the appropriate domain knowledge
• May need both safety and Linux domain knowledge
• Should we include reviewer bios to

Gab: Some subjective statements

• Cover this by verifiability claims
  • Where a claim is felt to be subjective by a reviewer, then we need to substantiate it
    • or qualify it, to make it clear in what specific context the claims apply
• Review against available technical documentation and the safety standards

Examples:

• Concept of self-interference may need to be reframed
• Some statements are very absolute, when there could be room for other approaches

Igor: Need to consider what the implications of putting an ‘official’ badge on a document from a liability perspective

Gab: If we focus on describing the risks that are involved in using Linux, and identifying possible mitigations, but avoid making specific statements about feasibility or suitability

• e.,g. enumerate the limitations of cgroups (and all kernel code) rather than making a value judgement about whether it is a ‘good’ or ‘dangerous’ solution
• describe the challenges objectively, and the limitations of any mitigations, and leave the decision about what is ‘safe’ to the system designer

Next time:

• Igor: Each feature in the kernel carries a certain amount of risk, based on its complexity, potential set of interactions (and therefore interference)
• Can we make a statement about how we might consider this in our approach?