Thursday 30th September 2010


People Present

  • JamieBennett

  • sbambrough
  • asac
  • mattman
  • slangasek
  • james_w
  • bsaibes

Actions Recorded

Actions from last week

  • asac to rename chessy-overlay to linaro-overlay

Meeting opened by JamieBennett at 16:01

  • sbambrough has to go answer the door

    <JamieBennett> [TOPIC] Action items from last week

Action items from last week

  • <JamieBennett> * asac to rename chessy-overlay to linaro-overlay

    <JamieBennett> asac left at the right time it seems ;)

    <JamieBennett> --> still needs to be done.

    <JamieBennett> [TOPIC] Announcements


slangasek to read to ensure it is correct

  • <asac> i did the rename, but didnt upload. let me find it and upload

    <JamieBennett> asac: thanks !

  • sbambrough back

    <JamieBennett> Also Linaro testing day is today, please test images if you can.

    <JamieBennett> [LINK]

  • asac will test alip

    <JamieBennett> asac: thanks, its still building with the changes I just did

    <asac> right

    <asac> i didnt hear much about vexpress

    <asac> is that going well?

    <asac> (testing)

    <JamieBennett> asac: Not really, matt was the only one to test, arnd, paul and peter have boards but I haven't heard much

    <JamieBennett> I did give them a friendly reminder early this week though ;)

    <JamieBennett> lets hope we hear something

    <asac> kk

    <mattman> i heard

    <mattman> testing now

    <JamieBennett> mattman: thanks, board all recovered now ;)

    <mattman> JamieBennett: recovered

    <JamieBennett> OK, lets move on

    <JamieBennett> [TOPIC] Bugs summary

Bugs summary

  • <JamieBennett> Lots of bugs here need some triage love.

    <asac> do we have way to mark something release critical for linaro?

    <asac> or just high and critical for now?

    <JamieBennett> asac: no tags but mark high/critical and subscribe linaro-release if your really worried

    <JamieBennett> We seemed to have accumulated quite a number of linux-linaro bugs, some need verifying on the latest releases.

    <JamieBennett> last one - [LINK]

    <JamieBennett> opps

    <JamieBennett> [LINK]

  • <JamieBennett> Seems we are OK apart from a couple of kernel bugs

    <JamieBennett> I'll be going through the list again today/tomorrow and assigning ;)

    <asac> omap kernel not supporting omap4 is medium? that feels wishlist ;)

    <JamieBennett> lool: set it as medium, as we are not committed to shipping OMAP4 it should be low I think

    <slangasek> agreed

    <JamieBennett> or wishlist

    <JamieBennett> changed

    <JamieBennett> [TOPIC] Progress Information

Progress Information

  • <JamieBennett> [PROGRESS REPORT] Overall


  • <asac> hmm initramfs not working on vexpress is also a bit scary. but that seems to be just from flash?

    <asac> anyway ignore me ... lets move on

  • slangasek sneaks his weekly status update in before the link is posted to the channel :)

    <JamieBennett> asac: Yeah, I'll go through them and change the importance/assign

    <JamieBennett> asac: feel free to comment on the bugs too ;)

    <asac> i think i am bad at not uses the status page i am supposd to use

    <slangasek> oh blast

    <slangasek> the include has been changed

    <asac> [ACTION] asac to ensure to use proper userplatform status page

    <asac> hah

    <JamieBennett> heh

    <JamieBennett> yours looks fine

    <JamieBennett> asac: ^

    <asac> mine?

    <asac> the summary is the default "please put your summary here"

    <JamieBennett> its the rest of the teams that should be worried (burndown pages)

    <asac> haha

    <asac> yeah that one is good

    <asac> but the summaries suck (not exist)

    <JamieBennett> asac: your summary needs changing though ;)

    <slangasek> so shall I delete the old FoundationsStatusSummary page, instead of leaving it there as a trap for myself? :)

    <JamieBennett> slangasek: yes

  • sbambrough says infrastructure burndown chart on link is incorrect, will Jamie be fixing it

    <JamieBennett> doh, just realised that I posted the wrong links above, let me fix

    <JamieBennett> ARGH, there is some wiki weirdness going on today

    <JamieBennett> I give up

    <asac> endmeeting then?

    <JamieBennett> I keep getting a "Please use the interactive user interface to use action edit!

    <asac> ;)(

    <JamieBennett> " message when I try to save

    <slangasek> I got that once earllier

    <slangasek> does that mean it doesn't think you're logged in?

    <slangasek> LP has been logging me out all morning

    <slangasek> s/LP/Ubuntu SSO/

    <JamieBennett> slangasek: no it clearly says I'm logged in

    <JamieBennett> anyway

    <JamieBennett> basically your status pages moved to e.t.c. per an anmar request

    <asac> cool

    <asac> can you please remove the old ones?

    <JamieBennett> asac: will do

    <asac> or rename them so we see that they are gone?

    <asac> great ... /me happy that he never remembered where the old ones were

    <JamieBennett> [ACTION] JamieBennett to remove old team burndown pages from the wiki

JamieBennett to remove old team burndown pages from the wiki

  • <JamieBennett> OK, back on track

    <JamieBennett> [TOPIC] Team reports

Team reports

  • <JamieBennett> [PROGRESS REPORT] Foundations


  • <slangasek> arm-m-ael-alip-evaluation: gstreamer package now cross-builds, with ppa patches; work on providing flavored cross-compiler packages for other ARM optimization targets

    <slangasek> arm-m-cross-compilers: lucid backport cross-compilers available at

    <slangasek> arm-m-kernel-version-alignment: tracking Ubuntu kernel package changes for an immediate post-maverick SRU of linux-linaro

    <slangasek> arm-m-memory-footprint: working on memory usage statistics from the headless image

    <slangasek> arm-m-missing-security-features: most Linaro kernel patches now pushed upstream and accepted!

    <slangasek> arm-m-uboot-features-and-performance: device-tree testing from smoothwall, reported upstream; ux500 u-boot has been merged into the -next tree

    <slangasek> arm-m-xdeb-cross-compilation-environment: updates to multistrap and pdebuild-cross packages in Ubuntu maverick. xdeb fix for correct Packages file processing is due this week.

    <slangasek> EOF

    <JamieBennett> slangasek: great

    <JamieBennett> any questions?

    <JamieBennett> OK

    <asac> nope

    <JamieBennett> [PROGRESS REPORT] Infrastructure


  • <sbambrough> * LEP/DerivativeDistributions: Backend work for new distroseries initialization expected to be done next week. Expect package difference backend to start next week, along with code to do synchronization. UI should be mostly finished tomorrow, there remain some changes to do to completed pages based on feedback. Hoping to have a beta ready by December.

    <sbambrough> * arm-m-private-archive-hosting-infrastructure: All work items postponed. We will mark items as INPROGRESS as we start them. We have some UI mockups for archive creation and package upload workflow in circulation within the team. Wider circulation is expected after incorporating feedback from the team.

    <sbambrough> * arm-m-derived-archive-rebuild: No progress. Remaining work postponed.

    <sbambrough> * arm-m-image-building-tool: Blueprint complete.

    <sbambrough> * arm-m-image-building-console: Cycles from OEM scarce atm. One single trivial branch merged this week. Backend for hardware pack generation was to be merged, but hasn't been yet. Backlog of branches for merging remains. We currently only expect to high priority or trivial merges to happen before UDS. Still expect this to release in November, it just won't progress in a speedy fashion. Planning for deployment of Linaro instance of Offspring

    <sbambrough> . Test builds from the London data centre made available today. These require testing to ensure the infrastructure in London is working as expected. Lexington infrastructure still available as backup until after release on November.

    <sbambrough> * arm-m-automated-testing-framework: Continued work on integrating test suites and benchmarks. See for details. Goals mostly met here. Phoronix test suite not amenable to running from Abrek; it is being broken up and implemented in a piecemeal fashion right now. Some web page rendering benchmarks postponed due to technical and legal issues. Other work mostly consists of bug fixes and testing

    <sbambrough> submission of test results to the dashboard.

    <sbambrough> * arm-m-validation-dashboard: Linaro instance of version 0.1 deployed this week at Version 0.2 was released today. The public instance needs to be updated. Working on next release which will concentrate on authentication and more views of data. Also major push to populate the dashboard with real data by running abrek and submitting results to the dashboard. Expected to use this data to aid in view definition.

    <sbambrough> * hardware-packs: Not much work here. LexBuilder backend to build hardware packs needs to be merged. Source hardware pack work still needs to be completed.

    <sbambrough> * x-project-blueprint-references: Picked up Launchpad spec Now on

    <sbambrough> * UDS summit schedule: This is done and live. One bugfix this week to correct problem displaying plenary sessions.

    <sbambrough> EOF

    <JamieBennett> sbambrough: wow, looks like you guys are busy atm

    <sbambrough> yup

    <asac> heh

    <JamieBennett> do we have an ETA on when the hwpack integration with lexbuilder will be done

    <sbambrough> its done

    <asac> if you lag whole cycle end gets always busy :-P

    <asac> (j.k.)

    <JamieBennett> sbambrough: deployed then

    <sbambrough> needs Cody to merge and deploy

    <sbambrough> was supposed to happen for Wed

    <asac> awesome

    <JamieBennett> sbambrough: OK so I suppose its *soon* then

    <JamieBennett> asac: looks like the hwpack moving isn't needed afterall

    <sbambrough> may be a month unless Cody hasn't done the deployment yet

    <JamieBennett> asac: unless you feel its still necessary?

    <asac> JamieBennett: ?

    <JamieBennett> sbambrough: ah, a month?

    <JamieBennett> asac: moving hwpacks to releases.l.o

    <asac> JamieBennett: i dont care as long as you copy weekly testing hwpacks to releases

    <sbambrough> they deploy monthly

    <asac> JamieBennett: also i am not sure what level of trust the current host has

    <sbambrough> we are building hwpacks via Hudson ATM

    <JamieBennett> sbambrough: right but they are not in an 'official' place

    <asac> e.g. are there viruses and trojans injected to our hwpacks ;)? ... no seriously, i think its fine, at best we would get those auto synched somewhere though (and signed and verified)

    <JamieBennett> sbambrough: we either need these moving to in the short term or LexBuilder deals with them

    <sbambrough> I can get it deployed sooner, i'll push it to steve

    <asac> sooner is always better.

    <JamieBennett> OK, sbambrough can you give me a heads up if we miss the roll-out window so we can sync hwpacks manually instead

    <asac> i leave it to jamie, steve etc. i am fine with everything. own infrastructure would be cleaner obviously

    <james_w> It's not a month

    <james_w> The rollout is delayed. Should be "soon"

    <sbambrough> good, then we will make it

    <asac> ok i think as long as we have a backup plan and we keep good updates on the status of rollout all should be fine

    <asac> thanks

    <JamieBennett> sbambrough, james_w: OK, keep me informed :)

    <JamieBennett> thanks sbambrough, james_w

    <JamieBennett> [PROGRESS REPORT] User Platforms

User Platforms

  • <asac> hey

    <asac> * Good news first: unity is working with mesa, omap4 and most likely imx51 (BSP)!

    <asac> * next: try meego omap3 drivers and see if they are better

    <asac> * next: let someone try on ste with mali drivers internally

    <asac> * next: speed things up further by adding TFP extension support to clutter (on its way)

    <asac> * heads are improving thanks to great work by tgall

    <JamieBennett> thanks tgall_foo !

    <asac> (and thanks everyone for helping him!)

    <asac> he is really doing a good job sorting all those small warts etc. that make a pleasent experience

    <asac> * ux500 bringup has X fbdev support now by default; next working on multimedia hw accell with open source bellagio and touch screen drivers

    <asac> * final spec status:

    <asac> * arm-m-graphics-stack-on-x (8/5/18) -> will be implemented within a week now that unity works

    <asac> * arm-m-ui-and-test-heads (48/2/52) -> same here ... would require spinning a new test head

    <slangasek> I hope to be more efficient at helping him with the head tweaking / size improvements now that I have live-helper hacked to cross-build images for me ;)

    <asac> * rest will be cleaned up by alf and me this week to have zero left; telephony will be kept open as opportunity for hacking at UDS

    <asac> slangasek: you have that? wow

    <asac> slangasek: did you also rebase the few patches we need to latest lh?

    <slangasek> asac: requires a one-liner change to debootstrap, and another to live-helper; currently nothing upstreamable

    <asac> i think most is upstreamed ... and the stuff that wasnt is not used anymore (like uboot etc.)

    <slangasek> asac: haven't rebased, no

    <asac> slangasek: does it use qemu-arm-static ?

    <asac> anyway. sounds good. would love to try ... lets take it offline

    <asac> ...continuing

    <asac> * > 90% of workforce is now on planning of next cycle topics etc.

    <asac> thats it i think

    <slangasek> yes... I hacked debootstrap to manually copy qemu-arm-static into the chroot <hangs head in shame>

    <asac> yeah.

    <asac> i misused ROOT command config at some point to try that

    <asac> the problem we face there is taht if you start installing a head with mono it will hang qemu when it runs mono

    <asac> to compile scripts (iirc)

    <JamieBennett> asac: answer = !mono

    <slangasek> yeah, I know

    <JamieBennett> ;)

    <asac> hehe

    <asac> good. anyway very good.

    <slangasek> if ALIP pulls in mono, we'll consider that a bug of a different sort ;)

    <asac> we should think about making that available

    <asac> also we should think about really trying the latest maverick lh rather than the low version thing we have. ..

    <asac> or pseudo upgrade our live-helper to have a version greater than what is in maverick

    <JamieBennett> asac: I smell a blueprint coming on ;)

    <asac> anyone volunteers that task?

    <slangasek> asac: yeah, I've only just hacked this last night and have spent the rest of the time waiting for a local mirror of maverick/armel to populate here so I can iterate... I'll generalize it soon

    <asac> slangasek: right. but there are other heads that might want mono ... we can avoid it in linaro luckily ;) /me loves linaro

    <asac> slangasek: thanks

    <asac> keep us updated

    <JamieBennett> right, thanks for the update asac

    <slangasek> I've thought about updating live-helper; are there any features we're missing, or is it just to have a sane version number in the ppa for upgrades?

    <asac> any questions on UP/Middleware?

    <asac> slangasek: lets talk offline. now that we have hwpacks we might get away just using what is in archive + 1 patch

    <slangasek> yep

    <asac> (to allow producing rootfs without kernel)

    <JamieBennett> [ACTION] asac and slangasek to discuss live-helper upgrade and report back results

asac and slangasek to discuss live-helper upgrade and report back results

  • <JamieBennett> OK, unless there are more questions ...

    <asac> thanks JamieBennett

    <JamieBennett> [PROGRESS REPORT] Landing Teams

Landing Teams

  • <JamieBennett> sbambrough: you have the floor

    <sbambrough> * Looking at slow MMC access on VExpress. Could be a hardware problem, doing testing to verify this is the case ATM. This causes slow boot with the current deployment strategy. Will add support for uboot/kernel in flash, and rootfs on USB media. This will become the default for Nov release.

    <sbambrough> * IGEP LCD display fix works on IGEP. Needs to be tested on Beagle to ensure no regressions.

    <sbambrough> * OMAP hardware pack produced. Working to get access to PPA with TI OMAP3 SGX drivers. Will be testing this out on the Beagle and IGEP and integrating them into the OMAP hardware pack.

    <sbambrough> * STE landing team a go. Needs to be staffed now.

    <sbambrough> EOF

    <JamieBennett> Great. I've seen a patch for the USB support for VE, I'll review and push today

    <sbambrough> will make booting a whole lot faster

    <JamieBennett> sbambrough: I think we are lacking testing for VE atm, mattman is doing a sterling job but we need the rest of the people with boards to test and report bugs too

    <sbambrough> how many do we have?

    <JamieBennett> sbambrough: 4

    <JamieBennett> ppearch, armd, paul and mattman

    <JamieBennett> ppearce

    <sbambrough> ok

    <JamieBennett> I'm pushing for more testing so we will wait and see I suppose

    <JamieBennett> thanks for the report sbambrough

    <slangasek> JamieBennett: has one been shipped to the DC also?

    <JamieBennett> slangasek: yes, 4 available for testing 1 in the DC AFAIK

    <slangasek> ok

    <JamieBennett> [PROGRESS REPORT] Kernel Consolidation

Kernel Consolidation

  • sbambrough has to step away for a minute

    <JamieBennett> sbambrough: OK

    <JamieBennett> lool: paulmck: available for a report?

    <JamieBennett> OK, moving on

  • sbambrough back

    <JamieBennett> [PROGRESS REPORT] Toolchain


  • <JamieBennett> bsaibes, all yours

    <bsaibes> Focusing on assigning and writing the blueprints for the new release. Usually, Michael send the status report for the current release

    <bsaibes> I will check with him for next week, to co-ordinate presenting the status to this meeting

    <JamieBennett> bsaibes: yes, I noticed the wiki is missing this weeks report

    <JamieBennett> bsaibes: thanks

    <bsaibes> I have 2 project management questions to ask before this meeting is over (if we have time)

    <JamieBennett> bsaibes: OK, do it in the AOB section

    <bsaibes> ok

    <JamieBennett> [PROGRESS REPORT] Power Management

Power Management

  • <JamieBennett> ST-E BSP was subjected to our current set of tools and missing interfaces in the BSP are being investigated; first impressions are quite good with powertop displaying the various low-power states correctly.

    <JamieBennett> imx51 frequency scaling driver is getting reviews and looks like it might make it to the next kernel version.

    <JamieBennett> Ongoing experimentation with powerdebug UI to find the best way to visualise clock trees

    <JamieBennett> Work is still ongoing (under review) to clean up the idle path in OMAP.

    <JamieBennett> Blueprints for 11.05 cycle have been created and assigned, assignees have been asked to fill out work items by 14th Oct.

    <JamieBennett> From

    <JamieBennett> [LINK]

  • <JamieBennett> Any comments on Power Management?

    <asac> yay for clock tree visualization. i have great hopes that this helps us in next cycle middleware validation suites topics

    <JamieBennett> asac: yeah, sounds interesting

    <JamieBennett> [TOPIC] AOB


  • <asac> bsaibes: go ahead ;)

    <bsaibes> 1) How and who have remote access to the Datacenters

    <JamieBennett> bsaibes: ssh and a bunch of us ;)

    <JamieBennett> bsaibes: depends on what you mean by remote access, what are you trying to do?

    <bsaibes> Jamie: can for example developers from IBM access new boards placed in the Datacenters to do their work, assuming they don't need to have a board

    <JamieBennett> bsaibes: this is something that IS needs to enable and we need to have a policy around who can access them. Lets take that one off-line.

    <JamieBennett> bsaibes: email me :)

    <bsaibes> Jamie: ok

    <JamieBennett> bsaibes: you had two questions?

    <slangasek> not all of the machines deployed in the data center on behalf of Linaro are intended as development machines; but for those that are, I think it's important for us to get all Linaro assignees access to them

    <slangasek> bsaibes is not the only one to bring this question up :)

    <bsaibes> 2) Can we assign sizing to Work Items in blueprints in order to do loadbalancing of assignments?

    <JamieBennett> slangasek: OK, we can get some answers after the meeting

    <asac> rick said once that work items should not be sized ... if something takes longer than a couple of days they should be split

    <JamieBennett> bsaibes: sizing, you mean effort involved?

    <asac> weight i guess

    <bsaibes> Jamie: yes

    <JamieBennett> asac: is correct, if its a task that cannot be broken down into 2 day pieces just write it as:

    <JamieBennett> do foo 1: TODO

    <JamieBennett> do foo 2: TODO

    <JamieBennett> do foo 3: TODO

    <JamieBennett> ...

    <JamieBennett> equating to the task length as guesstimated

    <asac> right. imo weighting would induce another level of complexity in planning and burn down interpretation

    <bsaibes> Jamie, the idea is to know whether a particular developer has more on his/her plate than can handle

    <JamieBennett> bsaibes: for a rough guess, it we do this right, you can use the number of work items

    <JamieBennett> but I can chat to you about that after the meeting

    <bsaibes> Jamie, after the meeting is fine

    <JamieBennett> OK, anyone else for AOB

    <JamieBennett> going once ...

    <JamieBennett> ... going twice

    <JamieBennett> #endmeeting

Meeting closed at 16:59


IRC Paste Text

[TOPIC] Action items from last week 
* asac to rename chessy-overlay to linaro-overlay 
--> still needs to be done.
[ACTION] asac to rename chessy-overlay to linaro-overlay
[TOPIC] Announcements
For people wondering what will happen to the Linaro release after Ubuntu ships Maverick,
should explain things.
Also Linaro testing day is today, please test images if you can.
[TOPIC] Bugs summary 
Bug #652306 - when network down, device disrupts network
Lots of bugs here need some triage love.
[TOPIC] Progress Information 
[TOPIC] Team reports 
[PROGRESS REPORT] Foundations 
[PROGRESS REPORT] Infrastructure 
[PROGRESS REPORT] User Platforms 
[PROGRESS REPORT] Landing Teams 
[PROGRESS REPORT] Kernel Consolidation 

Cycles/WeeklyReleaseMeeting/2010-10-07 (last modified 2011-04-15 19:51:54)