Participants

  • Andrea Gallo
  • Anmar Oueija
  • Fathi Boudra
  • Ilias Biris
  • Leif Lindholm
  • Rony Nandy
  • Ryan Harkin
  • Steve Capper
  • Steve McIntyre

Actions From Last Meeting

  • ACTION: Al to start breaking down the ACPI for ARMv7 to have something usable and demonstrable by Connect @ Hong Kong. Al should start breaking down into tasks to look then at how to start working with the cards with Ilias. -- TODO
  • CARRIEDOVER: Al to look into getting the list of Fedora packages which are dependent on inline assembly and would need to be reworked for ARM servers. -- TODO
  • ACTION Anmar will take care of sending the invitation letter to Rony for the sprint as he needs to prepare his visa. -- DONE
  • ACTION Anmar to take over running the US LEG engineering meeting. -- DONE
  • ACTION Ilias to create a proposal related to the way we are going to use JIRA for planning and status updates. -- DONE (need a separate action for some more wiki documentation)
  • ACTION Ilias should move related wiki pages from the previous Internal wiki location to this new area in the wiki. -- DONE
  • CARRIEDOVER: Rony to send the LAVA scripts for UEFI testing on Origen to Dave Pigott, and review the corresponding card with Ilias so that it can be closed. -- TODO
  • ACTION: Steve McIntyre should document his effort with the inline assembly dependent packages under the LEG Engineering wiki pages. -- TODO

Minutes

  • Checking the action items from 15.11.2012.
    • In progress still are:
      • ACTION: Al to start breaking down the ACPI for ARMv7 to have something usable and demonstrable by Connect @ Hong Kong. Al should start breaking down into tasks to look then at how to start working with the cards with Ilias.
        • TODO - to be discussed w/ Al after Thanksgiving.
      • CARRIEDOVER: Al to look into getting the list of Fedora packages which are dependent on inline assembly and would need to be reworked for ARM servers.
        • TODO - to be discussed w/ Al after Thanksgiving.
      • CARRIEDOVER: Rony to send the LAVA scripts for UEFI testing on Origen to Dave Pigott, and review the corresponding card with Ilias so that it can be closed.
        • LAVA scripts are being rebased to the latest LAVA release and should land in LAVA by end of wk47.
      • ACTION: Steve McIntyre should document his effort with the inline assembly dependent packages under the LEG Engineering wiki pages.

        • INPROGRESS: Steve is using a spreadsheet to review the scan results, the wiki page contains scan results for the ARMv8, but that does not exactly match the scan results for Enterprise. Steve will at least update the wiki page with respect to items which are different for Enterprise - at least put placeholders where needed for packages which are different.
  • Working with cards and subtasks - Ilias.
    • How is work organised in LEG: LEG will not use Launchpad to create blueprints for the roadmap cards it targets. Instead LEG will use only JIRA to plan its work.
      • JRA Cards will exist to cover the roadmap items planned by LEG, and each card will be broken down into JIRA subtasks to describe chunks of work that would be handled by blueprints in Launchpad.
      • Subtasks are lighter cards - each subtask needs to be associated to roadmap “parent” card.
      • Subtasks also follow a lighter workflow where the control of approvals is with the team itself - in order to speed up decision making when it comes to engineering work (as opposed to the lengthier review-based process followed for the roadmap cards).
      • A rule of thumb about the effort: an epic’s effort lasts roughly between 6m and 1y, roadmap cards last from a month to a quarter, whilst subtasks should last about a month.
      • Each subtask can contain an even finer breakdown of tasks needed to complete the work described by the subtask.
      • It should be noted that the JIRA progress tracking mechanism supports progress tracking from the level of subtasks upwards. Progress in terms of the work items inside a subtask will be tracked by humans - Anmar and Ilias.
      • ACTION: Ilias should write up in the wiki about how the planning + roadmap process are implemented in LEG
  • Cards and subtasks in LEG - Updates are needed for the following cards:

    • CARD-173: Shows that it is a legacy moving forward.

      • GRUB on Uboot was ok as viable option before LEG formation, now that we have UEFI on Arndale and VE, we should remove the Uboot from the card and update it to Arndale board.
      • Should maintain in the card a reference to the effort in upstreaming the uboot work done - doing that will make the upstreaming of the UEFI easier.
      • Need to add v8 Fast model too.
      • Drop Linaro Android as a dependency and just refer to Ubuntu Server for Arndale and OpenEmbedded on v8 FM

      • Let's add the dependency on GPLv3 FSF approval by LEG-SC and TSC.
      • ACTION Andrea/Ilias to add these comments to the card and request a change to update the card (perhaps need to split it in order to close with respect to what is already done).
    • CARD-174:

      • Whilst the original focus for this card was the Samsung Origen board, we should also add Arndale as possibility.
      • ACTION: Anmar should add the Arndale as a platform for CARD-174

      • sdmux and SATA boot testing should be added as a dependency (check link to blueprint).
    • CARD-240:

      • ACTION: Anmar to take the point to discuss with Alexander on what is the division of the validation work as described in CARD-240.

    • CARD-239:

      • Too dense to be a subtask.
      • Promote from subtask to turn this into a standard roadmap card - part of the cards for the CARD-210 epic.

        • ACTION Ilias to promote CARD-239 into a standard roadmap card

      • ACTION: Anmar to move the ACPI points (8,9) from CARD-239 into a separate card - with the several steps in getting ACPI to work.

    • CARD-211:

      • Should mention not only ARM but also x86 assembly dependencies.
      • ACTION Andrea to add comments to the card CARD-211 regarding mentioning not only ARM but also x86 assembly dependencies.

    • CARD-234, CARD-235:

      • Should we merge them in one card? SteveM: what we have is reasonable now.
      • ACTION: Steve to add work items to CARD-234, CARD-235.

    • CARD-215 mentions CARD-214 as an epic but CARD-214 is not in the LEG structure but in the main Linaro roadmap. Since virtualisation is a cross target (common Linaro and LEG) - LEG will have specific work on eg using openstack as a validation for KVM but it is not clear whether we need to add the epic card in the LEG roadmap structure in JIRA. Perhaps it would be better to clarify CARD-214 and what is the plan with it, in order to clarify what is the relationship. If CARD-214 stays in the main Linaro roadmap it should be owned by TSC otherwise it is a LEG epic and should be followed up by the LEG-SC.

      • ACTION Ilias to clarify CARD-214 with Loic.

  • Engineering build: starting from the “template” information given in CARD-222: ENGBUILD: LinuxLinaro Ubuntu a spreadsheet is created to collect information about the LEG requirements for the engineering builds.

    • This spreadsheet should be reviewed by the team before Tuesday 27 November so that we can pass any initial drafted requirements to Platforms team.
    • Some feedback regarding LAVA is already in https://wiki.linaro.org/LEG/Engineering/LAVA. For example network boot is a key support - net boot onto a device and install onto a SATA drive would be the way to validate an installation on a SATA device.

    • For the kernel: on one hand we have the multi-platform support in the kernel now however Versatile Express Cortex A9 version does not use the LPAE (the TC2 on Versatile Express can do it) whereas Arndale wants to run with LPAE enabled. Therefore they cannot be used in the same kernel image - we would need separate kernel images.
      • We are not going to have initially a single kernel or single configuration to support all the boards needed by LEG anyway - we can use the hwpack concept and use 1 hwpack for each kernel configuration needed. We should of course work towards the single kernel image as part of a later target.
    • For Fedora we need to discuss with Al - to see what requirements are needed for the engineering build. Maybe current state for Fedora 18 is usable? Could we start using that one?
    • ACTION: Anmar to check related to Al (when is he going to be full-time, when can he start on ACPI and also ask on the right Fedora build to use).
      • There is one more assignee from Redhat but we do not have any details yet.
    • For the Ubuntu engineering build:
      • The current instructions on Arndale setup focus on Ubuntu ‘precise’ Server image with a LAMP stack - from next cycle this image will not be produced by Platforms. If in LEG we want to have LAMP stack we should add it in the spreadsheet as a requirement for the Ubuntu engineering LEG build.

      • From next cycle (12.12 Linaro release) we are moving to ‘quantal’ 12.10 - if there is any reason for LEG to stay on ‘precise’ this should be clarified now. If there is any 3rd party wanting to evaluate their sw stack against an LTS release then that has to be supported by Canonical. LEG is not in the business of product type quality. Eventually we may need to coordinate with Canonical and Redhat for the production of releases which can be used for validation of 3rd party applications but the actual distro should come from Canonical/Redhat.
  • Other business:
    • Such review meetings where we look at the backlog of cards and possibly suggest changes to them need to be planned better. The meeting overrun and status was not covered sufficiently.

Action From This Meeting

  • CARRIEDOVER: Al to look into getting the list of Fedora packages which are dependent on inline assembly and would need to be reworked for ARM servers.
  • CARRIEDOVER: Al to start breaking down the ACPI for ARMv7 to have something usable and demonstrable by Connect @ Hong Kong. Al should start breaking down into tasks to look then at how to start working with the cards with Ilias.
  • CARRIEDOVER: Rony to send the LAVA scripts for UEFI testing on Origen to Dave Pigott, and review the corresponding card with Ilias so that it can be closed.
  • CARRIEDOVER: Steve McIntyre should document his effort with the inline assembly dependent packages under the LEG Engineering wiki pages.

  • ACTION: Ilias should write up in the wiki about how the planning + roadmap process are implemented in LEG
  • ACTION Ilias to promote CARD-239 into a standard roadmap card

  • ACTION Andrea/Ilias to add the comments from Andrea to CARD-173 and request a change to update the card (perhaps need to split it in order to close with respect to what is already done).

  • ACTION Andrea to add comments to the card CARD-211 regarding mentioning not only ARM but also x86 assembly dependencies.
  • ACTION: Anmar should add the Arndale as a platform for CARD-174

  • ACTION: Anmar to discuss with Alexander on what is the division of the validation work as described in CARD-240.

  • ACTION: Anmar to move the ACPI points (8,9) from CARD-239 into a separate card - with the several steps in getting ACPI to work.
  • ACTION: Anmar to check related to Al (when is he going to be full-time, when can he start on ACPI and also ask on the right Fedora build to use).
  • ACTION Ilias to clarify CARD-214 with Loic.

  • ACTION: Steve to add work items to CARD-234, CARD-235.

LEG/Engineering/Meetings/2012-11-22 (last modified 2012-11-25 23:52:47)