Thursday 19 August 2010

People Present


Actions Recorded

  • JamieBennett to arrange a bug bot for #linaro-meeting

  • JamieBennett to arrange a vexpress/omap image discussion with lool and asac

  • JamieBennett to propose starting the meeting 30 mins earlier to accommodate all topics including adding a Landing Team discussion


Meeting opened by JamieBennett at 16:03

  • <JamieBennett> [TOPIC] Bugs and feature summary

Bugs and feature summary

  • <JamieBennett> Bug #615765

    <JamieBennett> also this bug


    <asac> o/

    <JamieBennett> slangasek: I presume the latter will be worked on by hrw next week?

    <JamieBennett> (hey asac)

    <slangasek> JamieBennett: 605042> bug log suggests it's been worked around in eglibc; I'm not sure why the java side is targeted to beta in that case, it doesn't really seem critical-path

    <JamieBennett> I saw the comment about eglibc, do I take it that is upstream and will make its way back or that we will carry the patch for 10.11?

    <slangasek> I expect Ubuntu is carrying that patch until openjdk is fixed

    <JamieBennett> slangasek: right

    <JamieBennett> Other than the two bugs above I see nothing really affecting us drastically ATM, anyone spot any other bugs we should be aware of?

    <slangasek> 615765> yes, I'll make sure hrw is looking at that next week

    <JamieBennett> slangasek: great

    <JamieBennett> OK, lets move on then if theres no more bug concerns

    <JamieBennett> [TOPIC] Progress Information

Progress Information

  • <JamieBennett> Foundations is letting the side down a little ;)

    <JamieBennett> but I'm sure slangasek can fill us in in just a moment

    <asac> alf is rocking the user platforms ;)

    <JamieBennett> [TOPIC] Team reports

Team reports

  • <JamieBennett> [PROGRESS_REPORT] Foundations

  • JamieBennett hands slangasek the mic

    <lool> slangasek: First time I see LP #615765, it sounds like it would be a binutils upstream bug and not a packaging bug, should I add a binutils-linaro task?

    <lool> We have no bugbot here

    <slangasek> a number of the outstanding items are about getting cross compiler support merged into the base Ubuntu packages; some of those are already resolved but I haven't confirmed which, the rest will fall out in short order when hrw returns

    <JamieBennett> :(

    <slangasek> lool: sure - is linaro doing binutils releases currently?

    <JamieBennett> [ACTION] JamieBennett to arrange a bug bot for #linaro-meeting

JamieBennett to arrange a bug bot for #linaro-meeting

  • <lool> slangasek: We didn't need to so far, but it will happen eventually

    <lool> I believe either binutils or gdb will be included in the next release

    <slangasek> arm-ael-alip-evalution had been blocked for a while so there's a backlog of WIs, but progress is being made on these now

    <slangasek> overall blueprint status:

    <slangasek> * arm-m-ael-alip-evaluation: v5 & v6 compilers working; running their way through the stack now

    <slangasek> * arm-m-cross-compilers: toolchain patches merged into the archive last week

    <slangasek> * arm-m-debugging-with-oprofile: waiting for upstream feedback on final patch

    <slangasek> * arm-m-kernel-version-alignment: working to fix vexpress image so it boots; imx51 being added to the prebuilt images

    <slangasek> * arm-m-memory-footprint: bootchart changes awaiting feedback from Ubuntu; getting beagleboard up for testing

    <slangasek> * arm-m-missing-security-features: setting CONFIG_COMPAT_BRK=n to see if exec ASLR already works

    <slangasek> * arm-m-testsuites-and-profilers:

    <slangasek> * arm-m-uboot-features-and-performance: refining package output to work on non-OMAP targets (imx51, vexpress)

    <slangasek> * arm-m-xdeb-cross-compilation-environment: xdeb 0.6 can now resolve/build only the specified packages (using --only-explicit) but needs support for grabbing Packages files for the correct arch (n progress)

    <slangasek> * foundations-m-multiarch-support: pending LSB discussion of pathnames; dpkg support uncertain

    <slangasek> as of the foundations meeting yesterday, there's good progress on all fronts... except for multiarch, perhaps

    <slangasek> mea culpa

    <asac> slangasek: do you need any help on the ael-alip spec ... e.g. does ppearse need other seeds than the full UI alip seed tgall is working on etc.

    <JamieBennett> arm-ael-alip-evalution and arm-m-missing-security-features have stalled for a while, anything archive critical there?

    <asac> just wonder how ew can make something that is "deliverable" out of it

    <asac> i would think seeds would probably something archive relevant (but overlay is ok) ... if ppearse needs that at all

    <slangasek> asac: spec lists a minimal and a ship use case for ael-alip, in addition to the 'full' seed currently assigned to tom

    <slangasek> asac, JamieBennett: I don't believe any of the ael-alip is aimed at archive delivery; he's currently not even targeting armv7

    <asac> is minimal the "just initramfs with busy box" ?

    <slangasek> asac: I don't have the technical details on that - it may well be

    <asac> heh right.

    <asac> ok. lets try to make something out of it that is deliverable. even if not that exciting ;)

    <JamieBennett> slangasek: on a similar note, anything from the uboot spec hitting the archive?

    <slangasek> JamieBennett: I need to work with jcrigby to get a u-boot-linaro package into the archive; we should have it there if we can, yes, and request an FFe for it. Should have that request in to the Ubuntu release team this week.

    <JamieBennett> slangasek: great

    <lool> asac: Yes, that's about it

    <lool> slangasek: Do we need a FFE for an entirely new package?

    <slangasek> lool: yes

    <slangasek> lool: but the approval is usually trivial

    <asac> yeah. i dont see ubuntu release denying that

    <JamieBennett> slangasek: agreed

    <JamieBennett> slangasek: anything else foundations related?

    <slangasek> not unless there are other questions

    <JamieBennett> slangasek: the only other question would be multi-arch and the status for 10.11 delivery

    <slangasek> yes

    <JamieBennett> yes its going to happen or yes its a question ;)

    <slangasek> yes it's a question. :-) there are two at-risk elements here: 1) dpkg support, which is currently an external dependency; 2) pathname standardization, which is currently blocking on me getting the prototype done so that I can then try to push this up to the LF WG

    <asac> multiarch on 10.11 sounds unlikely ;)

    <asac> (just a feeling without knowing anything)

    <lool> slangasek: So I've asked jcrigby to add an imx51 flavor to linux-linaro; I suspect it also needs a FFE, even if we're not going to build any image out of that

    <slangasek> I can continue working on the library packages in parallel, but at least without 1), we don't get a working implementation in 10.11

    <slangasek> lool: right; I also have an action item from yesterday to ask the Ubuntu kernel team whether linux-fsl-imx51 should go away

    <lool> I thought it got removed

    <lool> Oh no

    <lool> My bad

    <asac> maybe not suitable here, but what does a flavour need to support before we enable it in linaro-linux?

    <slangasek> JamieBennett: by the end of next week I'll have a clearer idea if we're going to defer multiarch

    <asac> i was told that the ste board can boot to command line most likely (but no mmc support) with mainline

    <JamieBennett> Maybe we should capture the FFe requests on a wiki page?

    <slangasek> not just as bugs?

    <lool> asac: Happy to discuss with you on #linaro

    <slangasek> I think a wiki page would be make-work, unless you're planning to generate it automatically

    <asac> lool: right

    <JamieBennett> slangasek: I was thinking something else but I'll continue thinking on that one

    <JamieBennett> OK, lets move on

    <JamieBennett> [PROGRESS_REPORT] Infrastructure

    <JamieBennett> sbambrough: please take the stage

    <sbambrough> Ok, here's a summary of blueprints:

    <sbambrough> * LEP/DerivativeDistributions: Backend work under way. UI mockups look good. Not as much progress as I had hoped.

    <sbambrough> * arm-m-private-archive-hosting-infrastructure: Navigation work completed in vostok. Work to allow custom images. Basic preparation for implementing new views in vostok. Plugging away at work items.

    <sbambrough> * arm-m-derived-archive-rebuild: Slow progress on the archive deletion due to detailed DB work, and slow turnaround for checking changes when doing DB work.

    <sbambrough> * arm-m-image-building-tool: Blueprint's one remaining work item postponed. Blueprint now complete.

    <sbambrough> * arm-m-image-building-console: Working on getting several branches landed. All security reviews complete, no surprises. New name for release to be chosen by beta time. Started talking about documentation. Postponed anything not necessary for open sourcing.

    <sbambrough> * arm-m-automated-testing-framework: License approved. Added support for subcommands. Looking at test suites and benchmarks with eye to correctness on ARM and inclusion in the test framework. Don't have details about which test suites and benchmarks handy.

    <sbambrough> * arm-m-validation-dashboard: License approved. Working on getting changes landed. Will start new work items when this happens.

    <sbambrough> * hardware-packs: New spec. Spec done, implementation languishing as I've been swamped this week. Will start looking at it tomorrow.

    <lool> hardware packs > I guess that should only be considered next cycle?

    <sbambrough> no, this cycle

    <JamieBennett> lool: we hope to see it this cycle if possible but I wouldn't count on it for image building totally, we have a back-up plan :)

    <lool> so you want us to have images using hw packs this cycle hmm

    <sbambrough> I don't have good status on LEP/Derivative distributions, james_w is there anything you want to add

    <lool> I already have used cases for hw packs, for instance support for igep versus a generic omap3 image, but it didn't seem ready for use now

    <lool> *use cases

    <JamieBennett> lool: I think we go with separate images until hardware packs are available

    <lool> Right

    <JamieBennett> that way we don't lose out

    <sbambrough> no other choice really

    <JamieBennett> sbambrough: You still have work items targeted at Alpha-3, please get your team to update their blueprints

    <sbambrough> any questions on these items?

    <james_w> on LEP/DerivativeDistributions, the user testing should be finished early next week, and then frontend implementation will start. There has already been some work on the backend.

    <sbambrough> JamieBennett: just move them to beta or final?

    <JamieBennett> sbambrough: where ever you think they will land ;)

    <JamieBennett> sbambrough: other than that do you have anything that could need a FFe?

    <sbambrough> i hadn't been moving them so I had visibility on what was late

    <sbambrough> but i can if you want

    <JamieBennett> sbambrough: Yes they need moving, for next cycle I hope to have POSTPONED-TO-MILESTONE implemented so you can keep the work item in two or more milestones

    <sbambrough> that'd be sweet

    <JamieBennett> sbambrough: I doubt it but do you have any FFe potential items?

    <sbambrough> FFe?

    <JamieBennett> Feature Freeze Exceptions

    <JamieBennett> anything archive landing

    <sbambrough> none that I can think of

    <JamieBennett> sbambrough: OK

    <JamieBennett> thanks sbambrough

    <JamieBennett> [PROGRESS_REPORT] User Platforms

    <sbambrough> Paul had been trying to get expect-lite into universe, but its not critical

    <asac> hi

    <JamieBennett> asac: hi :)

    <asac> lots of stuff going on. lots of stuff not really tracked here because its out of cycle (like all the STE landing etc.)


    <asac> thats our status report for this week ... i failed to add a summary for this meeting

    <asac> so here we go:

    <asac> we need FFe for linaro-meta

    <asac> i got FFe for geoclue-skyhook and that was uploaded ...

    <asac> gps application is still sitting here on my disk waiting for finalization ... so that will need FFe

    <asac> arm-m-ui-and-test-heads -> this is quite good state. all the rest is out of archive, besides cairo-perf, but i dont see this in the archive this cycle

    <JamieBennett> ppa?

    <asac> if it happens then through a sync from debian

    <asac> yeah

    <asac> for test-and-ui-heads we are still thinking to make a seed that ships a good test case collection ... but not sure. most likely we will postpone that item

    <asac> arm-m-qt-on-embedded -> this was uploaded, rejected by archive admins because of code duplication

    <asac> then reuploaded ... and now riddell and me lobby to get this in. i think it will go in; we filed a bug to track the massive code duplication; stay tuned

    <asac> arm-m-graphics-stack-on-x -> clutter work is finally done. seb128 accepted our mammoth patch and pulls it into the archive. debian also will take it

    <JamieBennett> \o/

    <asac> rest is bug fixing in the broader sense ... like: atm unity doesnt work at all on gles etc. but we are working with cross-vendor and the new graphics specialist i have in team to fix things

    <asac> so i am not too unpessimistic that we will get this

    <slangasek> asac: I have alf's forwarded information from upstream about why this needs to be a parallel build; I'm reviewing and will help shepherd it in (or tell you if I think it's still a problem)

    <JamieBennett> anything telephony I should be aware of?

    <asac> on another front there is an unexpected opportunity we are currently trying. our new efl stack

    <asac> has gl engine

    <asac> which we switch to gles now ... so we can use that to try our drivers. according to raster its all tested on beagle so we might get a fully accelerated UI ;)

    <asac> of course not OOB unless we get drivers

    <asac> arm-m-telephony-stack -> still blocked on hardware. we are calling vendors atm. the simulater went nowhere

    <asac> so i think its definitly at risk to send a real SMS

    <JamieBennett> :(

    <asac> or do a phone call

    <JamieBennett> who are we talking to for hardware?

    <asac> telit is the thing we are looking at. thats a great thing, but quite expensive. we will probably buy one, but it might be too late to give a "ites working" for 10.11

    <asac> JamieBennett: we are trying to get in touch with telit if they would want to support us

  • JamieBennett notes that down

    <asac> also i am talking to STE because their board has a modem.

    <JamieBennett> OK

    <asac> but that only has proprietary drivers ... telit is all completely open and kind of perfect

    <asac> the rest is mostly covered above or not interesting for the release meeting

    <JamieBennett> OK asac, thanks

    <asac> questions?

    <JamieBennett> asac: mine were inline, anyone else?

    <JamieBennett> moving on

    <JamieBennett> [PROGRESS_REPORT] Kernel Consolidation

    <JamieBennett> lool: care to say a couple of words?

    <lool> Yes

    <lool> * /!\ no release

    <lool> * no new -next tree yet

    <lool> working on daily build infrastructure and release engineering; needs to syn

    <lool> c with nico

    <lool> Grmpf

    <lool> I made good progress on a buildbot based system this week, so I expect we will have daily builds soon, but probably on my home server (urgh)

    <lool> I'd like to use the Linaro ISO tracker to test our kernel releases if it's ready soon

    <JamieBennett> sbambrough: any comments on the ISO tracker?

    <JamieBennett> status

    <sbambrough> RT ticket filed, IS came back and asked us to justify new instance, Paul replied, no response from IS yet

    <lool> Otherwise, we'll test by hand

    <JamieBennett> sbambrough: does it need more pushing or has IS been responsive so far?

    <sbambrough> been responsive, just not fast enough for what we'd like

    <lool> I need to note there are broken expectations between what was scheduled to be delivered this cycle related to kernel consolidation which will be delivered in the kernel consolidation wg as a devleopment tree, and what will be delivered in Linaro archive proper

    <JamieBennett> OK, thanks for the update sbambrough

    <JamieBennett> lool: can we have a quick doc explaining the differences somewhere?

    <lool> so I expect we will continue landing new stuff, but it wont be integrated in images anymore

    <JamieBennett> to make it clear what is delivered wrt kernel?

    <lool> Doc for whom?

    <JamieBennett> to explain what will and won't be delivered, email me if its only me interested

    <lool> I think I'm being asked for something vaguely similar by the TSC

    <JamieBennett> lool: OK so we should see something on the wiki?

    <lool> I'm not sure, I didn't start yet; it could be wiki indeed

    <JamieBennett> OK

    <JamieBennett> anything else kernel wise? I see vexpress breakage atm

    <lool> So we only have 5 minutes left; for toolchain, there isn't much to report, focus is on runtime opts and completing 4.5 and hard-float stuff

    <JamieBennett> not sure its that a concern of the WG though

    <lool> JamieBennett: vexpress breakage?

    <JamieBennett> lool: OK, maybe we can sync up on that outside the meeting

    <lool> I actually have some notes for landing teams as well

    <lool> I wanted to raise that we don't have a good place to discuss integration of WG and landing teams output into release

  • JamieBennett nods

    <lool> I think this weekly meeting where we review the state of the release is where it should be

    <lool> For landing teams, this is what I had to note:

    <lool> * /!\ no vexpress images

    <lool> * omap3 images probably lacking tons of stuff

    <lool> Matt and Torez are working with Jamie and asac to catchup

    <lool> I'm worried about these two

    <JamieBennett> vexpress will be set up image wise by EOW next week

    <JamieBennett> I'm on that

    <JamieBennett> omap is asac territory I think

    <lool> So I'm not exactly sure how well the OMAP3 image will work, what it will have in terms of hardware support and features; the lack of a vexpress image is being fixed

    <lool> Right

    <lool> Well I would feel bad to blame asac on the omap image

    <asac> its fine to blame me

    <asac> but please first report issues before blaming me to not fix it

    <asac> i hever heard of anything missing. wht is missing?

    <JamieBennett> OK, I'll sync with asac any you to try to fix this for the images

    <lool> He certainly put it together, but I suspect it's on landing team's shoulder to deliver good support for more OMAP3 hardware, unless we consider thisimage was so good in 10.05 that it transitioned to platform already

    <JamieBennett> [ACTION] JamieBennett to arrange a vexpress/omap image discussion

JamieBennett to arrange a vexpress/omap image discussion

  • <JamieBennett> running out of time

    <asac> lool: i would love to get one ack that beagle and XM is booting fine before flipping our images to that

    <lool> asac: I know that it needs work to support more platforms, it's not necessarily only missing packages, it might be missing patches or tweaks, we miss installation instructions for other boards, we miss some firmware packages and such

    <asac> new meta

    <lool> asac: I'm glad to hear it works fine on XM for you

    <asac> lool: right. the meta package is right just for 1-2 days now

    <JamieBennett> I think this needs more discussion outside of here though

    <lool> I think we failed defining which boards were supported

    <asac> lool: no ... i would love to hear that before switching to linaro-image ... but i will just do it after meeting

    <JamieBennett> [PROGRESS_REPORT] Power Management

    <lool> JamieBennett: I realize we're running out of time, I think there is a backlog of omap3 topics to cover and can cover wiht asac offline

    <asac> yeah

    <lool> amitk isn't here

    <asac> i am dialing in call now

    <lool> I also think we have too little time to cover WGs properly

    <JamieBennett> amitk: can't ne here, he sent a summary

    <asac> and open line if i am first

    <JamieBennett> * Changes to powertop sent to upstream, no comments yet

    <JamieBennett> * Want to land that patch into Ubuntu for 10.11 (was hoping for an

    <JamieBennett> ack from upstream before doing that, but it might be too late by then)

    <JamieBennett> * slangasek was to provide amitarora some help with packaging (if we

    <JamieBennett> are to do this in a PPA)

    <JamieBennett> powerdebug, new tool

    <JamieBennett> * will only be a PoC/alpha software while we figure out the feature set

    <JamieBennett> * Plans to host in a PPA only

    <lool> either because we don't go fast enough on the other reports, or because we don't do it right, or because we need longer meetings

    <JamieBennett> * cpuidle latency measurement infrastructure

    <JamieBennett> * Early patch for OMAP available, not targeted for 10.11 currently

    <JamieBennett> [TOPIC] AOB


  • <JamieBennett> anything?

    <slangasek> hmm - I thought the help I was to provide was around the packaging of the new tool, not merely adding patches to the existing powertop, which I don't think takes two people :)

    <slangasek> I'll check with amit on this

    <JamieBennett> slangasek: OK, thanks

    <JamieBennett> #endmeeting

Meeting closed at 17:01

IRC Paste Text

[TOPIC] Bugs and feature summary
[PROGRESS REPORT] Ubuntu milestoned bugs
[PROGRESS REPORT] Ubuntu release-targeted bugs
[PROGRESS REPORT] Ubuntu known regressions
[PROGRESS REPORT] Ubuntu milestoned features
[PROGRESS REPORT] ARM and Linaro bugs
Bug #605042
Bug #615765
[TOPIC] Progress Information
[TOPIC] Team reports
[PROGRESS REPORT] Infrastructure
[PROGRESS REPORT] User Platforms
[PROGRESS REPORT] Kernel Consolidation
[PROGRESS REPORT] Power Management

Cycles/WeeklyReleaseMeeting/2010-08-19 (last modified 2011-04-15 19:38:02)