Wednesday 17 Nov 2010

People Present

  • slangasek
  • jcrigby
  • hrw
  • aviksil
  • npitre
  • wookey
  • ppearse
  • dmart
  • doko


  • JamieBennett: MeeGo Conference


  • Review action items from last meeting
  • Round table
  • weekly report reminder
  • volunteer needed to chair next week's meeting

Action Items from this Meeting

  • npitre to write up more formal documentation of what our kernel tree is (and isn't)
  • aviksil to talk to Mathieu about creating an lttng "for-linaro" branch that will merge cleanly
  • wookey to chair next week's foundations meeting

Action Items from Previous Meeting

  • slangasek, JamieBennett to discuss concerns raised at ELC about the new armhf port

  • jcrigby to register a linaro blueprint for this cycle's ux500 uboot mainlining work (DONE)
  • aviksil to put amitarora in touch with hrw on powerdebug packaging (DONE)

Status Reports

Peter Pearse



  • armel eclipse does not install - version differs from that in AMD64/i686, won't build on efika SmartBook

  • The final 5 packages of alip are all cross dependent.


Meeting Log

Meeting opened by slangasek at 16:10

  • <slangasek> waiting a tick for npitre to join us

    <slangasek> [TOPIC] Review action items fromlast meeting

Review action items from last meeting

  • <slangasek> * slangasek, JamieBennett to discuss concerns raised at ELC about the new armhf port

    <slangasek> * jcrigby to register a linaro blueprint for this cycle's ux500 uboot mainlining work

    <slangasek> * aviksil to put amitarora in touch with hrw on powerdebug packaging

    <slangasek> mine's still outstanding... JamieBennett is off at the MeeGo summit this week, maybe he'll bring back another set of concerns for us to talk about ;)

    <jcrigby> mine is done

    <slangasek> the blueprint registration is done:

    <hrw> powerdebug is done

    <aviksil> mine is also done :)

    <slangasek> thought so - great :)

    <hrw> I have one or two suggestions but will do that on irc

    <slangasek> ok

    <slangasek> [TOPIC] Round table

Round table

  • <hrw> hmm... linaro meeting on laptop + bob the builder at tv - jummy ;D

    <slangasek> which is like "blueprint status" but more general, because I haven't tried to collate the list of blueprints yet that we'll want to report on in these meetings... but with the drafting done at this point, you should all be hard at work on implementation :)

    <slangasek> $ echo $(shuf -e hrw npitre jcrigby dmart wookey slangasek ppearse JamieBennett aviksil)

    <slangasek> npitre wookey JamieBennett aviksil ppearse hrw jcrigby dmart slangasek

    <slangasek> npitre: you get to go first

    <npitre> got involved in a bunch of offline conversations about what the Linaro kernel tree is and is not

    <npitre> so I think we should document more formally what our kernel tree is

    <npitre> I suppose that falls on my shoulders

    <slangasek> npitre: I was just about to ask if you would take that on :)

    <slangasek> [ACTION] npitre to write up more formal documentation of what our kernel tree is (and isn't)

npitre to write up more formal documentation of what our kernel tree is (and isn't)

  • <npitre> other than that, been reviewing and commenting on a bunch of patches on lak

    <wookey> npitre I thought you had it just righ in recent thread (i.e linaro tree !=ubuntu tree)

    <slangasek> npitre: will that encompass questions like "why do we want to merge with 2.6.36 instead of trunk"?

    <wookey> I think some people are fogetting that linaro is just a dept of ubuntu

    <wookey> isn't !

    <npitre> right, but many other folks are wondering and they're not subs to linaro-dev

    <wookey> indeed - wider clarification is good

    <npitre> some folks, particularly from TI, are wondering if the Linaro kernel tree is good for them, etc.

    <slangasek> I look forward to the documentation :)

    <slangasek> anything else?

    <npitre> so I guess writing down in a wiki page what I've been saying for the last 2 weeks would be good (so I won't have to repeat it again to other people)

    <npitre> other than that, still trying to cohabitate with the workmen in the house renovating our kitchen

    <npitre> nothing else I guess

    <slangasek> ok

    <slangasek> wookey:

    <wookey> Written spec, but still working up some implemnetation details on xdeb sub-commands so I'm reasonably sure it'll all actually fit together.

    <wookey> Looking at sbuild

    <wookey> ppearse has done v5 compiler debs for me to make config for - intend to try that RSN

    <wookey> and secret internal stuff using xdeb. Be interesting to see how that works out

    <wookey> So yes I'm into implementing now

    <wookey> Easy bits first :-)

    <wookey> which includes multiarch deps compatibility

  • slangasek nods

    <wookey> not sure if steve has comments on order of taks on blueprint

    <wookey> it could be done in various orders

    <slangasek> I will - and will adjust the whiteboard to match, this week

    <wookey> ack

    <slangasek> multiarch dep compat is one of the low-hanging fruit I'd like to see done early

    <wookey> fine

    <slangasek> so we can start whittling away at the existing blacklist

    <wookey> who's doing the actual package multiarchification?

    <wookey> 'everyone'?

    <slangasek> yes

    <slangasek> we can dole that out once xdeb can handle the switch

    <wookey> OK, I think that's all

    <slangasek> ok

    <slangasek> JamieBennett: here?

    <hrw> JamieBennett is 3h idle

    <slangasek> aviksil: your turn, then

    <aviksil> Merged LTTng tree into both linux-linaro-next and linux-linaro-2.6.36. Got same set of conflicts for both the merges which I manually resolved.

    <aviksil> Will have to ask Mathieu Desnoyers to include few missing commits so that it merges cleanly.

    <hrw> Mathieu is lttng maintainer?

    <npitre> aviksil: sometimes the conflict resolution has to stay with the merge commit

    <slangasek> hrw: yes

    <slangasek> npitre: how do you want that to work for linux-linaro-next, though? Last cycle, the stated intent of that tree was that everything that didn't merge cleanly directly into it would get pushed back

    <slangasek> and since the unclean merge is due to a few upstream commits currently missing from the lttng tree, it seems better if the conflict resolution can be handled in that tree?

    <npitre> slangasek: there are different levels of "clean"

    <aviksil1> got disconnected at the wrong moment! :(

    <npitre> slangasek: simple merge conflicts are fine as long as they are not due to fundamental code structure issues such as different implementations conflicting with each others, etc.

    <slangasek> aviksil: do you need me to repaste anything for you? What was the last line you saw?

    <aviksil1> slangasek: yes. my update & discussion portion

    <slangasek> npitre: do you think it's *wrong*, then, for aviksil1 to ask this conflict resolution to be pushed into the lttng tree?

    <aviksil1> slangasek: last line was: (09:59:53 IST) aviksil: Will have to ask Mathieu Desnoyers to include few missing commits so that it merges cleanly.

    <slangasek> aviksil1: ok, pasted in /query

    <aviksil1> slangasek: thanks

    <npitre> slangasek: if the conflicts are due to small changes that exist only in the linaro tree then yes it is wrong for the lttng tree to carry those fixes... unless Mathieu is willing to have a "for-linaro" branch

    <slangasek> npitre: right - in this case they're due to his branch being behind upstream

    <slangasek> so it seems obvious to me we would want that conflict resolution in his tree

    <npitre> in fact, I think that is probably a good idea if Mathieu has a for-linaro branch so he can ensure good conflict resolution himself and maybe sanity test the resul;t

    <slangasek> aviksil1: can you propose this for-linaro branch to Mathieu?

    <aviksil1> sure

    <slangasek> [ACTION] aviksil1 to talk to Mathieu about creating an lttng "for-linaro" branch that will merge cleanly

aviksil1 to talk to Mathieu about creating an lttng "for-linaro" branch that will merge cleanly

  • <slangasek> thanks

    <slangasek> anything else?

    <aviksil1> Wrote two blueprint specs - other-linaro-n-lttng and performance-linaro-n-boot.

    <aviksil1> thats it from my side

    <slangasek> ok

    <slangasek> ppearse:

    <ppearse> More xdebbing...

    <ppearse> Set up for bug tracking

    <ppearse> About to start upstreaming cross build fixes

    <ppearse> Read wookey's spec, made him an armv5 chain

    <hrw> ppearse: bug 623212 is also reported in debian

    <ubot2> Launchpad bug 623212 in tcl8.5 (Ubuntu Maverick) (and 2 other projects) "[cross] compiling should run build host tclsh, not newly built target binary (affects: 1) (heat: 14)" [Undecided,Won't fix]

    <hrw> ppearse: debian bug 597708 may interest you

    <ubot2> Debian bug 597708 in tcl8.4 "tcl8.4: Cross compilation does not detect proper CC" [Normal,Open]

    <ppearse> I'll cross ref

    <wookey> ppearse: aha - bugs to read - excellent

    <wookey> we should compare with OE fixes and ptxdist fixes

    <hrw> ppearse: bug 645373 too

    <ubot2> Launchpad bug 645373 in gmp (Ubuntu) "Add --target option for cross builds (affects: 1) (heat: 64)" [Undecided,New]

    <slangasek> ppearse: have you had a chance to talk to tgall_foo about the differences in the set of packages you're working on, vs. what's being used in the linaro-alip image currently?

    <ppearse> Nope - on my list

    <slangasek> ok

    <slangasek> anything else? questions for ppearse about the xdebbing progress?

    <hrw> ppearse: list of packages is still the one in google sheet?

    <ppearse> hrw:Yes

    <hrw> ok

    <slangasek> hrw: your turn

    <hrw> ok

    <hrw> edited specs and workitems.

    <hrw> workitems for 'integration of linaro toolchains in natty' may need changing

    <hrw> started work on other-n-other-compilers

    <hrw> slangasek: please review newer gcc-defaults-armel-cross but not upload it

    <slangasek> I'll be looking over those workitems this week, talking to robbiew about getting it signed off - since it's really for the Ubuntu foundations team to decide there

    <slangasek> hrw: yes, will do

    <hrw> this is part of bug 676454

    <ubot2> Launchpad bug 676454 in gcc-defaults-armel-cross (Ubuntu) (and 2 other projects) "switch to gcc-4.5 and symlinks for default versions (affects: 1) (heat: 6)" [Undecided,In progress]

    <hrw> plan is to get rid of u-a for gcc cross packages

    <hrw> I discussed it with doko today on #linaro and then created that bug to track work

    <wookey> isn't u-a useful?

    <hrw> wookey: I want cross compiler to be same as normal one

    <slangasek> "useful" - it's not the right tool for the job

    <hrw> so no u-a but gcc-defaults-armel-cross

  • dmart wonders what u-a means

    <slangasek> if you want to use a compiler version other than the default, set CC=

    <wookey> I thought normal compiler was selected with u-a?

    <hrw> update-alternative

    <slangasek> dmart: update-alternativse

  • dmart is enlightened

    <hrw> wookey: no, by gcc-defaults

    <doko> wookey: no way, this would be insane

    <wookey> OK, my bad.

    <slangasek> :)

    <hrw> other things: discussed with kubuntu team about qt/kde problems on armel

    <slangasek> wookey: u-a is wrong for compilers because it detrimentally impacts build predictability

    <wookey> i'll read the spec :-)

    <hrw> my pandaboard is hard to reach since it started building qt4-x11..

    <slangasek> "hard to reach"?

    <wookey> 'busy'

    <hrw> also played a bit with efikamx smart[top|book] support - genesi will discuss with freescale about renaming libraries

    <hrw> slangasek: 150% of cpu in use? no respond on network/serial/console?

    <slangasek> doh

    <hrw> slangasek: only heartbeat led and harddrive sounds say that it still works

    <hrw> when qt/kde will be solved I will spend a bit of time to help kubuntu guys to get it running on 512MB ram machines

    <slangasek> ok

    <slangasek> anything else?

    <hrw> slangasek: armel-cross-toolchain-base 1.53 will be sent for review once I get it built.

  • slangasek nods

    <hrw> thats all

    <slangasek> jcrigby:

    <jcrigby> I believe my specs/work item drafting is done

    <jcrigby> working on new kernel package based on linux-linaro-2.6.36 tree

    <slangasek> how's that shaping up?

    <jcrigby> using better git-fu this cycle for adding the Ubuntu sauce so it will be easier to keep track of

    <jcrigby> I compiled for our three flavours last night

    <slangasek> do you think you'll have a preliminary package for uploading to Ubuntu this week?

    <jcrigby> yes, it will be the new per-flavour source packages

    <slangasek> ok, great :)

    <jcrigby> if its ok I will just publish to ppa for first round?

    <slangasek> I really want to get those into the archive so we have them there for alpha-1

    <jcrigby> then embarassment on failure is reduced

    <jcrigby> ok, fine

    <jcrigby> found problem with emmc reading in ux500 u-boot so we can now support it in same u-boot pkg as our other flavors

    <hrw> jcrigby: so linux-linaro-{omap3,ux500,omap4} etc?

    <jcrigby> hrw, yes

    <jcrigby> for parallel builds

    <hrw> ')

    <jcrigby> and slangasek has volunteered to upload in place of tim

    <slangasek> jcrigby: anything else?

    <jcrigby> yes, I did a wiki page on how to make your own packaged kernel for landing team people

    <hrw> url?

    <slangasek> gotten any feedback on that page?

    <jcrigby> no feedback yet

    <jcrigby> did it for lag

    <slangasek> I guess they'll be trying it soon enough :)

    <hrw> as I would like to package nicely kernel for efikasb


    <hrw> as "fakeroot make deb-pkg" is not so nice

    <hrw> [link]

  • <jcrigby> that is it

    <slangasek> thanks

    <slangasek> dmart:

    <dmart> hi

  • hrw -> out

    <dmart> I've written up my three blueprints, but in all cases it's not decided yet who will undertake the workitems

    <dmart> I'm waiting for feedback on those

    <dmart> For now, I'm treating kernel tasks as the priority, since the sooner we can merge the better.

    <dmart> I've been working on getting Thumb-2 kernel builds working -- I've posted a few patches for that and still need to test it on omap, but it looks OK

    <dmart> Next I will look at SMP-UP (and see if we can fix the runtime patching frameworks to work for Thumb-2)

    <dmart> I create a tree in -- see linaro-dev for details of that.

    <dmart> I think that's all.

    <dmart> Oh, I'm currently taking a look at freescale's newly-published 2.6.35 BSP branch

    <dmart> That's really all.

    <slangasek> ok :)

    <slangasek> my turn

    <slangasek> briefly: spec review/approvals, and the public plan review call for Foundations yesterday

    <slangasek> if you weren't around for the call itself, there's an audio feed on the wiki page you can download and listen to me droning on :)

    <slangasek> that's it

    <slangasek> so, just a couple more agenda items

    <slangasek> [TOPIC] weekly report reminder

weekly report reminder

  • <slangasek> I haven't been getting very many reports the past couple of weeks

    <wookey> ooh. I did one and sent it to arm but forgot you :-)

    <wookey> I'll fix that :-)

    <slangasek> I do still want to get these, so I know where things stand - and so we have something I can stick in weekly reports that's more parseable than a meeting log

    <npitre> slangasek: yeah... I've been in catch-up mode since I was back from UDS, and that includes those reports

    <slangasek> npitre: ack

  • wookey excited because he hardly ever manages to write wookly report

    <slangasek> just a gentle reminder, nothing more :)

    <aviksil> slangasek: I'm not aware of this weekly report at all!

    <slangasek> aviksil: just a short email, Friday or Monday, summarizing what you've done the previous week

    <aviksil> slangasek: ok, i'll do :)

    <wookey> (unless you work at arm in which case you have to write lots of tedious guff)

    <slangasek> [TOPIC] volunteer needed to chair next week's meeting

volunteer needed to chair next week's meeting

  • <slangasek> I'm off on holiday for a week, starting next Tuesday

    <ppearse> wookey: cur & paste is your friend

    <slangasek> Thanksgiving and all, here in the US

    <ppearse> s/r/t/

    <slangasek> but we should still have this meeting - can I get a volunteer to chair it?

    <wookey> erm. I guess I could do that.

    <slangasek> [ACTION] wookey to chair next week's foundations meeting

wookey to chair next week's foundations meeting

  • <slangasek> wookey: thanks :)

    <wookey> (except the random bit - that's too clever

    <slangasek> [TOPIC] AOB


  • <slangasek> anything else that I've missed on the agenda?

    <slangasek> #endmeeting

Meeting closed at 17:18

Platform/DevPlatform/Meetings/2010-11-17 (last modified 2011-03-11 11:34:02)