Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

  1. Kenta Murakami (PAS):

    • Upload slides to the conference page and attach the state diagram to the Jira ticket for GRPC integration.

    • Provide additional details or clarifications on Unified HMI architecture as requested by Marius Vlad.

  2. Marius Vlad:

    • Continue prototyping GRPC integration and provide feedback on missing information or challenges.

  3. Walt Miner:

    • Finalize the schedule for upcoming events and ensure alignment on demo preparation timelines.

  4. Michele Paolino (VOSYS):

    • Push updates to the AGL repository regarding loopback RFC and security improvements.

  5. All Participants:

    • Prepare for the next AGL All Member Meeting in Berlin, with a focus on showcasing Unified HMI demos.

Feb 04, 2025

Attendees:

Agenda: 

  • Topic 1: Unified HMI GRPC I/F

    • Jira Legacy
      serverSystem Jira
      serverId28dd2583-f97f-3492-8dcd-520f3459580a
      keySPEC-5280

  • Topic 2: VirtIO-loopback RFC

    • Michele Paolinopresented closing activities for phase3

      View file
      namevirtio-loopback-recap-rfc-2025-01-20.pdf

Minutes

Key Discussion Points:

  1. Architecture and Development Updates:

    • Kenta Murakami presented a revised architecture slide, proposing that the "Unified HMI Window Manager" may not be necessary due to gRPC compatibility with Go language.

    • Marius Vlad emphasized the need for a test component (AGL IVI Window Manager) to validate API connections and simulate demo requirements. This component is temporary and intended for testing purposes only.

    • The team discussed the importance of a testbed to validate data exchange between ULA Node and the compositor before finalizing implementation in the shell.

  2. Demo and Benchmarking:

    • The current demo setup uses KVM to run Zephyr and containers for Linux-based applications.

    • A proposal was made to explore dual architecture:

      • One using Xen Hypervisor.

      • Another using KVM and non-hypervisor solutions.

    • The goal is to maintain a unified interface while supporting different backend implementations.

  3. Xen Hypervisor Considerations:

    • Xen Hypervisor is seen as a strong candidate for production environments, especially in automotive use cases.

    • Challenges with Xen include limited support for graphics output (e.g., Virtio GPU).

    • Collaboration with ARM, Linaro, and Renesas was suggested to leverage existing work on Xen Hypervisor.

  4. AMM (Automotive Grade Linux All Member Meeting) Preparation:

    • Michele Paolino's session is currently on the waiting list. Walt Miner will review the schedule and confirm availability.

    • Michele was asked to prepare slides summarizing achievements for potential inclusion in the keynote or STVG session.

    • Discussions at AMM will focus on the future direction of hypervisor and non-hypervisor solutions.

  5. Future Work and Action Items:

    • The team discussed the possibility of a dual architecture approach:

      • Hypervisor-based solution using Xen.

      • Non-hypervisor solution using containers and KVM.

    • The importance of balancing committee member interests and technical feasibility was highlighted.

    • The need to address the graphics output limitations of Xen Hypervisor was noted.

...

Action Items:

  • Michele Paolino:

    • Prepare slides summarizing achievements for AMM and share them with Jerry.

    • Include options for dual architecture (Xen and KVM) in the presentation.

  • Marius Vlad:

    • Update the architecture diagram to clarify the testing and implementation process.

    • Continue working on the AGL IBI Window Manager as a testbed for validating API connections.

  • Jerry (PAS_趙健淙):

    • Coordinate discussions on Xen Hypervisor and KVM at AMM, including potential collaboration with ARM, Linaro, and Renesas.

  • Walt Miner:

    • Confirm Michele's session status for AMM and finalize the schedule.

  • All Members:

    • Continue discussions on dual architecture (Xen and KVM) and finalize the approach for Phase 4.

    • Prepare for in-depth discussions at the AMM regarding the future direction of hypervisor and non-hypervisor solutions.