Include: Nothing found for "^## meetings$"!

Include: Nothing found for "^## end-meetings$"!

Getting Started

We also maintain a set of how-to's related to cross-compiling


Out primary goal is to provide a common platform to use as the basis for Validation test heads and optimized hardware-specific builds. We work with a few teams to help make this happen:

Meeting Information

  • We have weekly meetings that anyone can attend. You can view the schedule and see logs of past meetings here.

  • We also meet face-to-face at Linaro Connect events.



  • Actions from last meeting - Steve chasing
  • Engineer progress reports - All team members
  • Work Item Progress and Team Issues discussion - Steve and JamieBennett questions to the team

  • Any other business (AOB)

Past Action items

Action Items

  • tgall_foo to seed a wiki page to discuss how to approach 1111 TR P1 (LEB as base for product builders)

Engineers Reports

Your Name <irc nick>


  • Short bullet points of work you've done that week which convey progress and highlights which can be used to report on how the team is progressing as a whole.


  • Your individual plans for the coming week(s).


  • Your individual plans for the coming week(s).

Peter Pearse <ppearse>


  • From 320 ALIP packages
    • 56 Need patching to cross build
    • 206 Build OOB
    • 10 Staged build for gobject introspection
    • 67 to do


  • Cross remaining alip packages other than those dependent on gobject introspection.
  • Attempt to cross gobject introspection.


  • Just realized xdeb imported some binary packages during my auto build, rather than crossing them.
    • Have now commented out all xdeb.cfg::native_import:

John Rigby <jcrigby>


  • Addressed all issues from comments on U8500 U-Boot patches.
  • Moved discussion on upstreaming Android Fastboot to u-boot list. Currently working on strawman design doc to facilitate discussion.
  • As usual, I have a packaged version of latest linaro linux kernel but it has had no testing yet.


  • Finish initial Fastboot on U-Boot design doc and send to U-Boot mailing list.
  • Normal kernel package releasing.
  • Investigate U8500 U-Boot mmc issues on some SD cards.
  • Do I need to find a new upload sponsor?


  • None.

Avik Sil <aviksil>


  • lttng-modules package has been revised and put into linaro-maintainers overlay PPA
  • Packaged new upstream release 0.19.9 of lttng-modules, but not yet put to my PPA
  • Participated in Linaro 11.05 beta 2 testing


  • Test latest lttng-modules package after it is added to linaro overlay
  • Verify oprofile works well on panda board after the fix


  • Uploading *source.changes file to my PPA using dput stalls





  • Not going to get everything in blueprint done

Tom Gall <tgall>


  • Off Friday + Monday as a result status is short
  • libturbo-jpeg - awaiting feedback from Mandeep Kumar
  • libav - awaiting feedback from Rony Nany
  • Helped Jamie update the manual testing information pages :

  • Chromimum - gflags, gmock, google-glog done
    • (gtest examined don't think we need)
    • flimflam looks to be already packaged. Vectormath might be as well. validating both.
    • Updating my chromium google recommended chroot.
  • Images dowloaded of beta 2. Getting them onto SD currently.
    • ltrace isn't working. Bug 771805


  • Chromium-wm packaging continues


  • None.

Riku Voipio <suihkulokki>


  • Getting introduced to Linaro tools and processes


  • Evaluate tracing tools from newcomper POV
  • Write a getting started doc to wiki based on experiences so far
  • Take over merging tasks from ppearse


  • Lack of testing HW.

IRC logs

Meeting opened by slangasek at 15:01

  • slangasek waves to ppearse

    <slangasek> [TOPIC] past action items

past action items

Engineer progress reports

  • <slangasek> you know the drill :)

    <slangasek> $ echo $(shuf -e hrw jcrigby dmart wookey slangasek ppearse aviksil tgall_foo kunal fturgis suihkulokki)

    <slangasek> hrw suihkulokki fturgis kunal aviksil wookey dmart ppearse jcrigby slangasek tgall_foo

    <slangasek> suihkulokki: hey there! how's this week been?

    <suihkulokki> Hi

    <suihkulokki> Continued familiarizing with l-m-c and linaro-hwpack-create with Qemu

    <suihkulokki> xdeb works also quite well, but didn't manage to get the linux-linaro kernel compiled with it..

    <suihkulokki> Lack of of working hw for linaro made it not possible to try out any of the tracing stuff

    <suihkulokki> I grabbed now a beagleboard B5, which has it's issues and iirc isn't supported?

    <suihkulokki> that's in the nutshell

    <slangasek> I'm not sure what the status of the B5 is - I know there were some late u-boot fixes to try to support it

    <suihkulokki> Ok, then I'll try it out

    <slangasek> suihkulokki: I think there's supposed to be a snowball en route to you - hasn't arrived yet?

    <rsalveti> slangasek: I tested the fix at my b5 and it worked fine

    <suihkulokki> I pinged asac earlier this today and he said it still wasn't sent :(

    <rsalveti> at least it's already included at ubuntu

    <slangasek> suihkulokki: ah, doh

    <slangasek> xdeb> I haven't tried using it to build a kernel... for kernel builds I think a straight debuild -aarmel invocation is more typical. But if you see something to fix in xdeb to make kernel builds work, great :)

    <rsalveti> suihkulokki: the problem with b5 is the lack of proper multi-read support, even with the fixed u-boot you could still see issues at the kernel

    <slangasek> kunal_:

    <kunal_> hello

    <wookey> jcrigby found that kernel packages need fixes to cross-build interenal tools

    <wookey> not sure if those patches made it in

    <wookey> once they are dpkg-buildpackage should work just like everything else

    <kunal_> this week integrated review comments in smartt application by MMWG

    <wookey> (not that it matters much as cross-building kernels is trivial anyway)

    <suihkulokki> well, crosscompiling kernel without packagin is trivial

    <wookey> exactly

    <kunal_> tried to test on panda board

    <suihkulokki> the challenge is to modify an existing linaro kernel to something lmc will accept to the ime

    <suihkulokki> image

    <slangasek> suihkulokki: if it's not working for you, definitely check with jcrigby for the most recent runes

    <wookey> (and cc me)

    <kunal_> regarding the kwin packages the blueprint status says the testing is done, need to check with jammy


    <slangasek> kunal_: what's the outcome of your testing on panda?

    <kunal_> on panda bug is already there for hw events while running perf

    <jcrigby> suihkulokki:

    <kunal_> comment added by sachin for the same

    <kunal_> so not able to run the application

    <kunal_> in dmesg log it can't reserve memory for pmu

    <slangasek> hmm, ok. do you have the bug number handy?

    <kunal_> no

    <aviksil> kunal_: you may want to try the latest linux-linaro with pmu/perf patch

    <kunal_> i am using hw pack and image from latest build

    <kunal_> can you share that patch

    <kunal_> i'll try with that

    <aviksil> kunal_: ok, i'll share it

    <kunal_> thanks

    <slangasek> ok, great

    <slangasek> kunal_: anything else?

    <kunal_> no, i'll mail the bug number

    <slangasek> thanks

    <slangasek> aviksil:

    <aviksil> ok

    <aviksil> lttng-modules package has been revised and put into linaro-maintainers overlay PPA

    <aviksil> Packaged new upstream release 0.19.9 of lttng-modules, but not yet put to my PPA

    <aviksil> having issue in uploading by dput

    <aviksil> not got any response from #launchpad

    <slangasek> hmm

    <slangasek> and the problem is consistently reproducible?

    <aviksil> slangasek: yes

    <aviksil> and finally, participated in Linaro 11.05 beta 2 testing

    <slangasek> have you tried running strace / lsof on the dput process to see what it's doing when it stalls?

    <aviksil> tested on panda board

    <aviksil> slangasek: i tried strace

    <aviksil> it stalls at recv(3,...)

    <suihkulokki> have a firewall that blocks ftp?

    <james_w> switching to/from sftp upload may help

    <wookey> are you uploading from other end of ssh connection?

    <aviksil> no, because other files are uploaded successfully

    <aviksil> problem occurs only in .changes file

    <aviksil> james_w: i'll try that

    <aviksil> that's all from me

    <slangasek> alright, thanks

    <slangasek> wookey:

    <wookey> ssh -t has fixed prblems with dupload for me in the past

    <wookey> maybe dput too?

    <wookey> it appears to upload files then stalls forever

    <wookey> days bank holiday so more drilling than hacking

    <wookey> 2

    <wookey> Changed cross-pkg-config support so wrapper provided by pkg-config

    <wookey> links provided by gcc-default-armel-cross package. So only armel currently supported, but no cross-pollution of normal systems without cross-tools.

    <wookey> Seems to work OK

    <wookey> patches at:



    <ubot2> Launchpad bug 771567 in pkg-config "cross-compiling wrapper support for pkg-config" [Undecided,New]

    <ubot2> Launchpad bug 605189 in gcc-defaults-armel-cross "Provide symlinks for cross-compilers" [Undecided,New]

    <wookey> steve, are you happy with that implementation?

    <slangasek> in principle, but I've not had a chance to look at the code yet :)

    <wookey> the gcc-related version number is a bit odd, and you don't get cross-pkg-config unless you install pkg0config-arm-linux-gnueabi package (continaing nothing but a link)

    <slangasek> can you submit the pkg-config patch to the Debian BTS and get Tollef's take on it?

    <wookey> OK, good idea

    <wookey> ested multistrap and pdebuild-cross - does in fact still work on natty, despite out-of-date config

    <wookey> Now back to working on xdeb to get 731079 fixed sharpish

    <wookey> Seem to have accidentally acquired a GSOC student interested in bootstrapping

    <wookey> that's it

    <slangasek> I still think it would be better to shove it into the 'gcc' binary package so it's available opportunistically, without having to specially install another package

    <slangasek> "accidentally", eh ;)

    <wookey> OK. the problem with that is adding pkg-config dependency to gcc package

    <wookey> is that acceptable?

    <slangasek> right, I think we should just ship a possibly dangling symlink

    <wookey> I implemneted that first

    <wookey> OK. I haven't investigated if anything bad happens there

    <slangasek> things that look for but don't really need pkg-config for building should DTRT with a dangling symlink

    <wookey> (e.g things testing for 'does triplet-pkg-config exist')

    <wookey> one hopes so, I just wondered...

    <slangasek> I could be wrong, but I guess it's a quick test, yes?

    <wookey> I'm happy with either arrangement

    <wookey> I can test with xdeb easily. I just wonder about some random package or autofoo doing odd things and going wrong

    <wookey> I guess it's a but if a dangling symlink counts, so we should fix that if found

    <wookey> OK. I'll send in a patch for that arangement too then

    <slangasek> autoconf should behave consistently, so testing that should answer the common case

    <wookey> are we sure it's better in 'gcc-cross' than in pkg-config?

    <slangasek> other things may test in funny ways... but we can burn that bridge when we come to it

    <slangasek> there's only one pkg-config but may be any number of cross compilers for different targets, so I thikn so

    <wookey> fine

    <slangasek> thanks :)

    <slangasek> ppearse:

    <ppearse> Just hacking away at the the alip packages as per my report....

    <ppearse> < 50 to go - should finish all except gobject introspectors next week

    <ppearse> Then spend UDS week on intenses contemplation of introspection ;-)

    <ppearse> s/ses/se/

    <slangasek> heh :)

    <ppearse> Questions?

    <wookey> nope - I asked mine in person earlier :-)

    <slangasek> the progress on the ALIP packages looks great - nice job!

    <suihkulokki> the alip image appears seriously bloated... on the first boot you get mlocatedb, apt-update-xapian, mandb-update and whatever spinning for quite a while before the UI becomes responsive

    <ppearse> BTW should I propose merges any more?

    <slangasek> ppearse: it may be better to hand off your patches to suihkulokki instead and let him drive any further merges

    <ppearse> OK will do

    <tgall_foo> suihkulokki, indeed

    <slangasek> suihkulokki: hmm, I have to say that I don't think we've looked at boot performance of ALIP... the focus has been on install footprint (which is obviously not great either yet, by embedded standards :)

    <tgall_foo> suihkulokki, getting alip down to something more reasonable will hopefully be an acceptible 11.11 item

    <wookey> runing on 1st boot is probably correct, but those things probably shouldn't be there at all..

    <aviksil> kunal_: you may try this patch:;a=commit;h=457520e3fbce80812c6901c226ec242fdb906c63 it's already available in linux-linaro-2.6.38

    <suihkulokki> agree with wookey

    <tgall_foo> slangasek, nano on a panda for instance is ~5-6 seconds

    <kunal_> aviksil: thanks

    <wookey> I think we all agree that there is plenty of opportunity for shinking minimal Debian/Ubuntu

    <tgall_foo> wookey, indeed

    <aviksil> kunal_: np

    <wookey> everyoine else it trying to make it fatter :-)

    <wookey> is

    <slangasek> ;-)

    <slangasek> jcrigby:

    <suihkulokki> ppearse: when you have time send the patches you have and give some explanation how and where they should be merged

    <jcrigby> short week, same usual stuff

    <ppearse> suihkulokki: Ack

    <jcrigby> U8500 u-boot patches have had no more comments so I guess they are good for 2011.06

    <jcrigby> move the discussion about upstreaming fastboot to u-boot mailing list and got some feedback from Wolfgang

    <slangasek> U8500 - great :)

    <jcrigby> after some back and forth I have signed up to a design doc to discuss how it should plug into u-boot

    <slangasek> nice job on diverting the fastboot thread onto a productive track... you're clearly a mailing list judo master

    <jcrigby> the doc is just a strawman based on the existing implementations (out of tree) but will serve to continue discussion

    <jcrigby> have packaged latest nico kernel but not tested

    <jcrigby> thats about it

    <slangasek> thanks

    <slangasek> me up

    <slangasek> this week has been mostly tying up loose ends for me; flushing the merge review queue, getting packages reviewed and sponsored into ppas where needed, etc

    <slangasek> closing out those last pesky work items and encouraging you all to do the same :)

    <slangasek> also working on getting our list together of topics for LDS

    <wookey> slangasek: if I have a work item blocked by another part which is 'postponed' then should it be mrked 'blocked', or 'postponed'?

    <slangasek> anything you think we should be discussing at LDS, please send it to me so I can make sure it gets on the list

    <slangasek> wookey: 'postponed' IMHO, 'blocked' is when you're blocked by something external

    <wookey> OK. I thought that might be the case

    <wookey> BTW Am I still supposed to be doing developer docs or can I considered that handed off to suihkulokki?

    <suihkulokki> hmm.. is everything going to end up handed over to me? :)

    <wookey> until you start to bleat, probably :-)

  • ppearse hides

    <slangasek> wookey: you have the better overview of the current state of play; whatever docs suihkulokki writes for his own edification as he goes should be integrated, but I want you to still drive this

    <wookey> I recall you being responsible for 'everything', or somesuch wording :-)

    <wookey> right-ho

    <tgall_foo> slangasek, do you have thoughts on the TR for starting point for product builders will play out? Is there a wiki page with thoughts about how to approach that?

    <slangasek> tgall_foo: not really; that's definitely something we should gather input on at LDS

    <slangasek> tgall_foo: if you want to seed a wiki page with your thoughts, feel free :)

    <slangasek> any other questions for me?

    <wookey> TR?

    <tgall_foo> slangasek, yeah ... something that's unformed like that might not translate into a very productive UDS session

    <slangasek> wookey: technical requirement:

    <slangasek> tgall_foo: right - if we can have the conversations with vendors before UDS and use the UDS session instead for formulating a plan of action that would be more productive, but I don't think there's time to actually gather that input before UDS

    <slangasek> tgall_foo: your turn

    <tgall_foo> slangasek, yup, it is what it is ... I'll take a crack at a wiki page ... feel free to make that into an action item :-)

    <tgall_foo> Off Friday + Monday as a result status is short

    <tgall_foo> libturbo-jpeg - awaiting feedback from Mandeep Kumar

    <tgall_foo> libav - awaiting feedback from Rony Nany

    <tgall_foo> Helped Jamie update the manual testing information pages :

    <tgall_foo> Chromimum - gflags, gmock, google-glog done

    <tgall_foo> (gtest examined don't think we need)

    <tgall_foo> flimflam looks to be already packaged. Vectormath might be as well. validating both.

    <tgall_foo> Updating my chromium google recommended chroot.

    <tgall_foo> Images dowloaded of beta 2. Getting them onto SD currently.

    <tgall_foo> ltrace isn't working. bug 771805.

    <ubot2> Launchpad bug 771805 in ltrace "[armel] ltrace hangs" [Undecided,Confirmed]

    <tgall_foo> and that's it

    <slangasek> [ACTION] tgall_foo to seed a wiki page to discuss how to approach 1111 TR P1 (LEB as base for product builders)

tgall_foo to seed a wiki page to discuss how to approach 1111 TR P1 (LEB as base for product builders)

  • <slangasek> tgall_foo: could you try building ltrace from aviksil's branch to confirm whether this is fixed upstream?

    <tgall_foo> slangasek, ... aviksil pointer ?

    <tgall_foo> slangasek, sure

    <aviksil> tgall_foo:

    <tgall_foo> aviksil, gratcie

    <slangasek> gflags, gmock, google-glog> what does that leave yet to do?

    <tgall_foo> slangasek, well presuming there aren't undiscovered deps ... about 7 packages

    <slangasek> ok

    <tgall_foo> libchrome, libhcromeos, libcros, flimflam, chromeos-asses, update_engine and vectormath

    <slangasek> publishing these to a ppa?

    <tgall_foo> slangasek, handling them off to jahz to publish in the graphical team's PPA

    <slangasek> alright

    <tgall_foo> but otherwise yes .. they are going into my own PPAs too :-)

    <slangasek> :)

    <slangasek> cool, thanks

    <slangasek> [TOPIC] AOB


  • <slangasek> what else...

    <slangasek> reminder: weekly reports to be added to the meeting wiki page so they can be summarized for management (

    <wookey> UDS subjects we should just mail you about. Is there a list to check repeats against? (i.e. mine might already be listed)

    <slangasek> not yet, I'll get that into the wiki today and send out the link

    <suihkulokki> If we have TRs to suggest for the next cycle, how do we proceed?

    <tgall_foo> seems like in the future we should devote time during a team meeting about future TRs

    <slangasek> and this is my last weekly meeting with you guys... next week I won't be with Linaro anymore :(

    <tgall_foo> and that should occur say a month before the next UDS

    <tgall_foo> slangasek, we'll miss ya!

    <jcrigby> so is asac our fearless leader until another is designated?

    <slangasek> so I want to thank you guys for being so great to work with, and wish you all well for the coming cycle

    <ppearse> send back T-shirt

    <wookey> :-)

    <asac> hi guys

    <asac> i will do my best to keep you utilized

    <slangasek> jcrigby: correct - they're still working on making the time-to-designation as short as possible, in the meantime asac is the go-to guy :)

    <asac> now that slangasek is gone holiday season is over :-P ... j.k.

    <tgall_foo> heh ... now asac is going to crack the whip

    <slangasek> suihkulokki: TRs to suggest> the TRs are agreed between the Linaro TSC and the engineering team; if you think something should be there that isn't, please talk to asac

    <suihkulokki> ok, thanks

    <slangasek> anything else?

    <slangasek> #endmeeting

Platform/DevPlatform/Meetings/2011-04-27 (last modified 2011-04-27 16:49:01)