Automotive Linux Wiki

The IVI Expert Groups meets every other Thursday at xx:00 UTC. Meeting minutes the upcoming scheduled meeting dates are listed below.

Join Zoom Meeting https://zoom.us/j/98420206742?pwd=WFFleHBYY0lEejdUR2haQnBTMGsyZz09

Meeting ID: 98420206742

Meeting Passcode: 121632

Find your local number: https://zoom.us/u/ameRude4a

Confluence page https://confluence.automotivelinux.org/display/IVIPR/IVI+Production+Readiness+Activity+Home

Attendees:Walt Bernard Jerry Mikhail Scott Kurokawa Kobayashi Hosokawa Kato Nakao Nishiguchi Nomoto Date Takeuchi

Agenda:

• IVI-EG Schedule Update
• Production Readiness Requirement Specification
• Virt-IO device support priority

• IVI-EG Schedule Update
  • 21:00, Thursday on the same day which SAT meeting is held.
  • Next IVI-EG on May13th
• Production Readiness Requirement Specification

Attendees:Walt, Jan-Simon, Bernard, Daniel, Kobyashi, Jerry, Kato, Date, Nakao, Scott, Takeuchi

The following is the material I explained. https://confluence.automotivelinux.org/pages/viewpage.action?pageId=41680904

Agenda:

• Explain PowerManagement feature in Basesystem(↑material)
• Feature Plan with Virt-EG
  • Process Overall discussion in SAT → each EG
  • At next SAT, Jerry-san explains the each needs the prioritization
  • At 22 SAT, collect the prioritized list and make prioritization.

Attendees:Walt, Jan-Simon, Scott, Kurokawa, Date, Nomoto, Hosokawa, Nishiguchi

Agenda:

• To share the status.
  • Three topics from basesystem and RBA.
• 2021 Feature Planning by IVI-EG.
  • Add tasks in this sheet IVI-EG by March 10th
    • https://docs.google.com/spreadsheets/d/1WsFPh_-2wAYE8NvKlBGywYixhhJz9EBlbUR02r0UWF8/edit#gid=428149407
  • Prioritize by March 18th
  • Be approved from SAT or other meeting
  • Set schedule
    • Nomoto(★) will send the email how to proceed this feature list action.

Agenda:

• To share abstraction of AMM presentation from Toyota
  • The content of AMM
    • Background(Toyota AGL activity in 2020)
    • What is contributing for Production Readiness?
    • Introduction of Demo
    • Awareness in Trial Operation
    • What is plan to target next?
    • Conclusion

• To discuss some items if anyone have

Attendees:

Agenda:

• Discussion Plan (minor update)
• Introduction
• PR software structure should be more composable and selectable
• remained Topics
• Security Module
• Package Management & deployment
• API
• HMI
• IPC

https://confluence.automotivelinux.org/pages/viewpage.action?pageId=36765867

Attendees: Walt, Jan-Simon, Nomoto-san, Tanikawa-san, Kurokawa-san, Hosokawa-san, Date-san, Scott, Tsubone-san, Oki-san

Agenda:

• Overall Update
  • discussion plan
  • contribution status and review comment
• Basesystem HAL follow-up
  • What is HAL? Why is HAL needed?
  • Signal handling overview
  • HAL Example 1: Positioning HAL
  • HAL Example 2: CAN HAL
  • CAN HAL API
  • Typical use cases
    • Send CAN data
  • Plan (Idea) for Production Readiness

Click this link to see the slide. https://confluence.automotivelinux.org/pages/viewpage.action?pageId=36765830

Attendees: Walt, Jan-Simon, Nomoto-san, Date-san, Oki-san, Scott, Aketa-san, Tanikawa-san, Nishiguchi-san,

• Remained BaseSystem feature
  • Lifecycle management
• HAL
• Review of basesystem contribution
• SPEC status

Attendees: Nomoto-san, Kurokawa-san Nishiguchi-san Scott, Aketa-san, Tanikawa-san, Walt, Jan-simon

From Toyota(advance comments)

• The files below are the same name files, but at the moment there are some differences between the two.
• ./service/system/task_manager/client/libtskmcfg/conf/agl_thread.h
• ./agl-basefiles/include/agl_thread.h ^^
  • question: use different names to not confuse ? e.g. agl_thread_tm.h ?
• So our ideas are to keep the two in basesystem.git at the current state and to refactor them from next year.

• Current HALs are stub and they don't have function.
  • As actual product, Basesystem has HALs and cannot be separated for supplier.
  • TOYOTA and related company are implementing only can_hal for R-Car aiming for June.
    • question: what makes the can_hal rcar-specific ? (only vcan - Kurokawa-san)
    • question: what is the abstraction provided by the can_hal (and other *_hal) .
      • e.g. over the BSP like meta-rcar-gen3
      • or e.g. why need to abstract the standard linux kernel socketcan

• Almost all of the other review comments are done but I don't have good answer for
  • https://git.automotivelinux.org/staging/basesystem/tree/agl-systemd/99-agl.rules
    • MODE 666 or 777 → security ??
      • MODE 666 → 660
      • MODE 777 → 770
    • Are these better ?
      • Yes. But: you might have to use group and group membership for the apps reading/writing from these then. E.g. for 'drm' in the agl-demo-platform with smack, there is a 'display' group and any app requiring display access need to be member of the 'display' group. Similar there is an 'audio' group.
      • example (with smack): https://git.automotivelinux.org/AGL/meta-agl-demo/tree/recipes-graphics/wayland/weston-init/zz-dri.rules.in
      • or app/service runs as 'root'

• xpath:
  • need in recipe: 'inherit perlnative' (Nomoto-san)
  • need to patch xpath recipe to use '#!/usr/bin/env perl' (Scott)

AI:

• check if documentation about HAL can be shared
• review agl-basefiles recipe. are these needed at all given:
  • headers are installed by the recipes that ship them
    • agl-basefiles/agldd: these headers originate in the evk.
    • They should be shipped evklib-headers.bb
    • meta-oem-production-readiness/meta-agl-basesystem/recipes-kernel/kernel-module-evklib/evklib-headers .
  • include/ cannot be split.
    • https://confluence.automotivelinux.org/pages/viewpage.action?pageId=32833688

Plan:

• any pending changes to basesystem.git go first,
• then merge recipes (if the image builds)

Next call: January 7th 2021

Attendees: Nomoto-san, Aketa-san, Tanikawa-san, Yamaguchi-san, Haraki-san, Kurokawa-san, Walt, Jan-Simon, Scott

• Agenda review
• review of the current state
  • slides with findings since Monday
  • https://docs.google.com/presentation/d/13MEi1ors6PGbsUQn1Ms8vxaZHsiamwMPzyndRc7QgXo/edit?usp=sharing
• next steps / tasks
• open mic
• next call
  • Monday, December 21st 2pm

Attendees: Nomoto-san,

• Agenda review
• Basesystem review

1. - Need quick intro from Toyota what is where (very brief does each folder contain)
   • agl-basefiles: used by all other components (e.g. Makefiles, some headers)
   • agl-systemd: systemd files used to start systemmanager
   • hal: SOC abstraction
   • module: kernel module
   • service: contains all function source code
   • stub: stub functions
2. - Questions:
   • agl-basefiles
     • agl_thread.h vs sm_launch_conf.h ???? they look basically same ??
     • agl_types_obsoluted.h → agl_types_deprecated.h
       • obsolete = no longer in use (aka should not be used at all)
       • deprecated = still in used but being faded out / reworked / replaced
     • ifdef KERNEL + license mentioned in header
   • agl-systemd
     • https://git.automotivelinux.org/staging/basesystem/tree/agl-systemd/99-agl.rules
       • MODE 666 or 777 → security ??
   • kernel-module-evklib:
     • https://git.automotivelinux.org/staging/basesystem/tree/module/kernel-module-evklib/evk_lib.c#n56
     • evk_lib.h → check
   • stub:
     • clock/localtime: why duplicate zoneinfo ? → there is tzdata package in yocto ?
3. - Q&A

• Yocto recipe review (gerrit)

1. - walkthrough on recipes in gerrit
   • 25647 ok, pending changes to basesystem (agldd)
   • 25648 ok, but pending changes to kernel module
   • Fails to compile: 25652
     • | ./src/resmgr_api_lib.c:19:10: fatal error: resmgr_api.h: No such file or directory
     • possibly generated? → need to fix build dependency !
2. - discussion on simplification and path forward
   • 'SDKTARGETSYSROOT=${STAGING_DIR_HOST}'“ not required ('–sysroot=${STAGING_DIR_HOST}' is already part of the $CC variable !
     • – influences basesystem-common.bbclass
     • – Makefile makros in basesystem.git
3. - work on recipes
   • 25647-25651 (JS & Scott) BSMAKE_FILE fix
   • 25652 Tanikawa-san (resmgr_api.h)
   • 25653-25654 → Nomoto-san (xpath and BSMAKE_FILE)
   • Nomoto-san: need to deal with basesystem.git issues mentioned above
     • agl-thread.h
     • agl_types_obsoluted.h
     • evk_lib.h
     • stub/clock/localtime ?
     • 'SDKTARGETSYSROOT=${STAGING_DIR_HOST}'” not required ('–sysroot=${STAGING_DIR_HOST}' is already part of the $CC variable !

* Next meeting on wednesday.