This month's meetings

WorkingGroups/Kernel/Meetings
<< <  2011 / 8 >  >>
Mon Tue Wed Thu Fri Sat Sun
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30 31        

Meeting info

  • Kernel WG meetings are conducted over IRC: #linaro-kernel on Freenode.

Attendance

Morning Attendees

  • Loïc Minier: x
  • Paul McKenney: x

  • Arnd Bergmann x
  • Anand Gadyar:
  • Per Forlin: x
  • Niklas Hernaeus: x
  • Nicolas Pitre: x
  • Dave Martin: x
  • Shawn Guo: x
  • Mounir Bsaibes: x
  • Grant Likely: x
  • Andy Green:
  • Venkatraman S: x

Evening Attendees

  • John Stultz: x
  • Jeremy Kerr:
  • Jason Hui:
  • Grant Likely:
  • Aneesh V:
  • Thomas Abraham
  • Sanjeev Premi:
  • Nishant Kamat:
  • Paul McKenney:

  • Mounir Bsaibes: x
  • Deepak Saxena: x

Agenda

  1. 11.11 Cycle Planning
  2. BUGS review
  3. Blueprints / Around the table

Action Items from Previous Meeting

To see the completed action items: WorkingGroups/Kernel/ActionItemsDone

Minutes From Today's Meetings

  • K7.10 DMA Engine on Arm - waiting for feedback from TSC. Deepak has the action. Will probably change prio to LOW.
  • Perf is no onger handling K7.2 ( Reduce overhead of data movement between kernel and user)he is now handling K7.11 (Separate data/control USB buffering, optimize USB buffer types)
  • Tixy estimate for K5.1 (Thumb2 kprobes work) is 20 days.
  • Dave still need to discuss with Deepak and others about how the various items from K5 and K8 will be assigned.
    • New items came up during blueprint discussions which weren't part of the initial plan, so we need to consider how to prioritize those (kvm investigations etc.)
  • Shawnguo has put a few work items on https://blueprints.launchpad.net/linux-linaro/+spec/linaro-kernel-o-ras-fix-ltp, Mounir to review

  • Nico: many of the things I have to do are on-going tasks not bound in time
  • Paulmck On maintainership: http://lwn.net/Articles/443510 appeared Thursday. Still thinking through what I should put on a blueprint.

    • Paul raised the issue of "how should maintainership activity be tracked?" -- especially given its reactive nature.
  • Niklas will handle K2.2 (Demonstration of Complete Device Tree) for Snowball, Niklas will replace Mian in the allocation.
  • The alarmtimers went upstream at the end of last week! so John can mark that as done for 11.05 and scratch it from the 11.11 blueprint
  • John raised 2 issues:
    1. A concern about the seemingly large split between what is available for testing in the hwpacks, and nico's tree.
    2. Handling last Minute fixes. We need more clear "freeze" period for nico's tree, to reduce churn in the linaro+android tree

Action Items From Today's Meetings

  • Deepak to check on K7.10 with the TSC
  • Mounir to update the estimate for Tixy as he does not have access to the spreadsheet.
  • Tixy to provide an estimate for his work on K7.1 (Improve performance of standard filesystems on flash via remapping layer)
  • gcl to sort out blueprints DT blueprints.
  • Dave and Deepak to close on K5 & K8 assignments

  • Dave and Deepak to look into handling new items that came up in the blueprint discussion
  • Mounir to review the Work Items for https://blueprints.launchpad.net/linux-linaro/+spec/linaro-kernel-o-ras-fix-ltp blueprint

  • Mounir/Deepak to decide how to allow for on going activities planning.
  • Mounir/Deepak have to create the milestones for the linux-linaro project for example 11.05, 11.06, 11.07 etc...
  • Mounir/Deepak to communicate the right way to assign Work Items to a milestone
  • Deepak to sync/talk to Nico and Jcrigby and TL's about the 2 issues raised by John: about not being too much later then a week or so out from the hwpack images to nico's tree and baout a freeze period for Nico's tree.

IRC Logs

Morning meeting IRC log

<npitre> hello there
<tixy> hello
<gcl> hi npitre
<shawnguo> hello everyone
<mounir> Hi
<perfor> hi
<npitre> this is a holiday here, so I'm currently traveling
* lpieralisi (~lorpie01@fw-tnat.cambridge.arm.com) has joined #linaro-kernel
<npitre> I therefore may disapear into 3g blackholes from time to time
<nhe> hi
<svenkatr> hi
<paulmck> Hello!
<tixy> Who's running the meeting?
<paulmck> I was thinking that dsaxena was.
<paulmck> Me, I was belatedly writing up my status.  :-)
<shawnguo> it's not holiday in US, right :)
<shawnguo> even it is holiday, npitre is here
<mounir> https://wiki.linaro.org/WorkingGroups/Kernel/Meetings/2011-05-23
<mounir> dsaxena: r u there?
<npitre> shawnguo: I'm in Canada
* dmart (~dmart@fw-lnat.cambridge.arm.com) has joined #linaro-kernel
<dmart> hi all, sorry for the delay
<shawnguo> npitre: yes, it is holiday in Canda :)
<arnd> hi dmart, we're all still waiting for dsaxena to show up to run the meeting
<dmart> ah, ok
<tixy> dmart: hi, dave, were still wating for deepak to show up
<dmart> hi
<shawnguo> dmart: hi dave
<dmart> hi there
* gcl is celebrating same holiday as npitre 
* sumits (~sumit@linaro/sumits) has joined #linaro-kernel
<shawnguo> gcl: you get a lot of gpio codes to review :)
<gcl> shawnguo: indeed.  Don
<gcl> oops
<gcl> Don't expect much of it to land in 2.6.40 though
<gcl> but we'll see
<gcl> some of it is pretty low risk code moves
<npitre> gcl: would be nice though
<shawnguo> gcl: I do not think we will hit .40 either, as rc1 is soon coming in less 2 weeks
<shawnguo> gcl: actually one week or so left?
<gcl> npitre: yes.  Carrying over to 2.6.41 doesn't excite me either.  It's really a question of how much I can get sorted out in the next week
<gcl> even then I'm really pushing it
<mounir> Should we go around the table and discuss agenda item #1? 11.11 blueprints, drafting progress, estimate correct or not, etc...?
<npitre> gcl: you might at least push the pure moves to Linus, tha would give an idea of where things are going
<gcl> npitre: agree
<tixy> mounir: I think we should, doesn't look like Deepak is turning up
<mounir> any volunteer to start the discussion? 
<tixy> mounir: Do we know what blueprints we should be putting our work items against? I noticed a few people using the LDS meeting blueprintd
<perfor> mounir: Is the google doc 11.11 kernel planning the one I should check
<tixy> mounir: I would have thought it would be the TR blueprints linked from https://wiki.linaro.org/Cycles/1111/TechnicalTopics/Kernel
<dmart> mounir, perfor: the agenda wiki links to: https://spreadsheets.google.com/a/linaro.org/ccc?key=0AoZqvK7R1biJdHotS19UamEtVEdjMWd4VWpkZ29HcXc&hl=en&ndplr=1#gid=2
<mounir> tixy: it is ok to use the LDS meeting blueprint, as long as it linked to the TR blueprint (i.e. make the TR dependant on that blueprint)
<dmart> Is that the right thing to be looking at?
<perfor> mounir: thanks
<mounir> perfor: yes,  check the doc 11.11 kernel planning to check the estimate 
<mounir> tixy: you are correct the TR blueprints are at  https://wiki.linaro.org/Cycles/1111/TechnicalTopics/Kernel
<perfor> mounir: K7.10 waiting for feedback from TSC. Deepak has the action. Will probably change prio to LOW.
<mounir> perfor: ok
<perfor> mounir:  I have moved from k7.2 to k7.11
<perfor> mounir: I am no longer allocated for k7.2
* sumits (~sumit@linaro/sumits) has left #linaro-kernel ("Leaving")
<mounir> perfor: can you update the doc? Notice I moved the allocation, to a separate sheet called allocation, you can see your total allocation there
<perfor> mounir: As far as I can tell the estimates look good.
<perfor> mounir: sure
<perfor> mounir: doc is updated. That's all from me I think
<mounir> Tixy: are you ok with your estimate and blueprint?
<tixy> mounir: I don't have access to the planning doc. I estimated 20 days for thumb2 kprobes work.
<mounir> tixy: I will giveyou access after the meeting and update the estimate for you
<tixy> mounir: we had this issue before, I don't have any sort of ID you can use to give me access to google docs.
<tixy> mounir: I also have some work for K7.1 I volunteer for at LDS, I must produce some estimates for that
<mounir> tixy: I can do what is needed for you, if I cannot give you access 
<mounir> tixy: ok - let me know, I will update the doc
<mounir> gcl: How about you ?  are you set with your blueprints and estimates, do you have a lot of specification to do for DT in 11.11 or that is already been done in 11.05
<gcl> mounir: There is a lot of work for 11.11, but I'm not able to look at sorting out blueprints until tomorrow.
<mounir> gcl: ok
<gcl> most of the core code work was done during 11.05 though
<mounir> gcl: K3 was sent back to OCTO for definition, are you handling that?
<dmart> I still need to discuss with dsaxena and others about who the various items from K5 and K8 will be assigned to
<mounir> dmart: yes, we need to close on those items, I think dsaxena has some idea to engage the new assignees on those items
<dmart> mounir: There were also some things that came up in the blueprint discussions which weren't part of the initial plan, so we need to consider how to prioritise those (kvm investigations etc.)
<gcl> mounir: yes, the k3 stuff is being handled by octo
<gcl> all of the boot architecture stuff is my primary deliverable for the 11.11 cycle
<mounir> dmart: is this something that has to go back to OCTO to decide on, or it can be decided on by the KWG?
<mounir> arnd: anything from your side?
<dmart> mounir: I suggest we discuss in the WG first and then go to the TSC if it looks necessary
<mounir> dmart: ok - I will record an action item to handle
<arnd> mounir: I was on vacation most of last week, no noteworthy news from me since LDS
<mounir> arnd: are you planning to look over your bleuprints, and start the drafting next week, if you can please review the estimate and make sure they rae reasonable
<mounir> arnd: I mean this week
<arnd> sure, will do
<mounir> arnd: thx
<mounir> Shawnguo: how about you? Are you set with what you will be handling?
<shawnguo> mounir: I put a few work items on https://blueprints.launchpad.net/linux-linaro/+spec/linaro-kernel-o-ras-fix-ltp
<shawnguo> mounir: you please give more input there
<mounir> shawguo: ok - will take an action to look those over
<mounir> npitre: would go next? There is a lot of focus on you this cycle for the monhtly releases, etc... 
<mounir> paulmck: any update from your side?
<npitre> mounir: many of the things I have to do are on-going tasks not bound in time
<mounir> npitre: yes - but I beleive will have to communicate what you decide on doing in details so everyone would know how to feed to your releases, correct?
<paulmck> mounir: On maintainership: http://lwn.net/Articles/443510 appeared Thursday.  Still thinking through what I should put on a blueprint.
* CosmicPenguin (~nobody@soa.codeaurora.org) has joined #linaro-kernel
<mounir> paulmck,  should that article be broadcasted at this stage or not yet, until it is more clear what is being done?
<mounir> nhe: Do you know what you will be working on this cycle?
<paulmck> mounir: The article is already published, as requested and agreed to in Budapest.
<nhe> mounir: K2.2 for Snowball.  I am to replace Mian in the allocation.
<paulmck> mounir: The issue in my mind is "how should maintainership activity be tracked?" -- especially given its reactive nature.
<mounir> paulmck, we should allow in our planning for on going activities - I will add a item to the actions to address this question
<paulmck> mounir: fair enough!
<mounir> few minutes left to the meeting, anyone would like to raise an issue that we need to handle?
<mounir> In a couple of week, Linaro will conduct a public presentations for cycle 11.11, Deepak will handle the presentation, but he may need to come to you for review before he present
<mounir> anything else before we close the meeting?
<dmart> nothing from me
<mounir> thank you all, the meeting has ended

Evening meeting IRC log

<dsaxena> mounir: greetings
<jstultz_vm> dsaxena: mounir:  hey!
<mounir> Hi dsaxena
<mounir> jstultz_vm, hello
<mounir> https://wiki.linaro.org/WorkingGroups/Kernel/Meetings/2011-05-23
<dsaxena> who else is around?
<dsaxena> jstultz_vm, mounir: might just be the 3 of us
<dsaxena> jstultz_vm and I met last week and I believe you're on track with filling out BP for the android upstreaming?
<jstultz_vm> dsaxena:  yea, i took a first swing through must of them.
<jstultz_vm> dsaxena:  did you get a chance to skim the work items? does it seem like "enough" for now?
<dsaxena> jstultz_vm: no, had internet issues most of today. :(   What's the correct URL? https://blueprints.launchpad.net/linaro/+spec/tr-kernel-android-upstreaming2 which is linked from the 11.11 session planning page is empty
* jk- (~jk@ppp121-45-241-16.lns20.per2.internode.on.net) has joined #linaro-kernel
<dsaxena> ah, but there's the spec @ https://wiki.linaro.org/WorkingGroups/Kernel/Specs/androidUpstreaming-11.11
<jstultz_vm> https://blueprints.launchpad.net/linux-linaro/+spec/linaro-kernel-o-android-upstreaming
<jstultz_vm> https://blueprints.launchpad.net/linux-linaro/+spec/linaro-kernel-o-android-commontree
<jstultz_vm> https://blueprints.launchpad.net/linux-linaro/+spec/linaro-kernel-o-android-commontree
<jstultz_vm> dsaxena:  those are the three i wrote up
<jstultz_vm> dsaxena:  one change there... the alarmtimers went upstream at the end of last week! so i can mark that as done for 11.05 and scratch it from the 11.11 blueprint
<dsaxena> :)
<dsaxena> jstultz_vm: looks good to me as a start point :)
<jstultz_vm> dsaxena:  ok, great.
<dsaxena> jstultz_vm: I think TSC wants monthly milestones, do you think you can break those high level tasks up in that way (just email me the info)?
<jstultz_vm> dsaxena:  ok..  should i put ====== between items? or is there a better way to do this in-blueprint?
<dsaxena> jstultz_vm: that works for me. mounir, any comments?
<jstultz_vm> i just hate the idea of duplicating and syncing between your list and what's in the blueprint.,
<dsaxena> jstultz_vm: agree
<jstultz_vm> dsaxena:  ok.. i'll split things out by month and then we can adjust them as we go along
<mounir> in the blueprint you can target a qork item to a milestone, is that what you are asking?
<mounir> s/qork item / work item/
<dsaxena> mounir: yes
<mounir> one sec let me find the workitemhowto url it has an example 
<jstultz_vm> https://blueprints.launchpad.net/linux-linaro/+spec/linaro-kernel-o-arm-config-managment
<jstultz_vm> oh.. i'll wait to see mounir's howto..
<mounir> https://wiki.linaro.org/Process/WorkItemsHowto
<mounir> go to the subcycle tracking section
<jstultz_vm> mounir:  what are the sub-cycle names?
<mounir> jstultz_vm, in the example lucid-beta-1 is the milestone (sub-cycle)
<mounir> jstultz_vm, dsaxena we have to create the milestones for the linux-linaro project for example 11.05, 11.06, 11.07 etc...
<jstultz_vm> mounir:  right, but what are our sub-cycles called?
<jstultz_vm> ok.. so just 11.06
<jstultz_vm> sounds great.
<jstultz_vm> i'll get those broken up then.
<jstultz_vm> dsaxena:  so i've got two things i wanted to bring up, if there's nothing else on the blueprint review..
<mounir> yes - dsaxena I saw a note from the TL's  meeting about this. do you remember how we should do it?
<dsaxena> mounir: not sure, I didn't make it to meeting this morning, will follow up
<mounir> jstultz_vm, we should take this as an action for me and dsaxena to handle and communicate the right way to do it
<mounir> dsaxena, do you agree?
<dsaxena> mounir: agreed. we'll send an email to whole team about it.
<mounir> dsaxena, we need to actually do it on the linux-linaro project as well before we send the note
<dsaxena> mounir: understood.
<jstultz_vm> mounir:  and its ok if work-items shift around between the sub-milestones? right?
<jstultz_vm> mounir:  we don't need to postpone them in one sub-milestone and create a new one in the second, right?
<jstultz_vm> mounir:  for instance, i have broken things up somewhat, but some items will be multi-month efforts, so i'll probably fill in details as i go..
<mounir> jstultz_vm,  yes, we don't want to go overboard with this. Specially it is the first time we do it. We want to do what is reasonable
<jstultz_vm> mounir:  cool..
<jstultz_vm> mounir:  ok.. i just broke things up in my 3.. send me mail if they need any tweaks..
<jstultz_vm> dsaxena:  so... the two issues i wanted to cover..
<dsaxena> jstultz_vm: go ahead
<jstultz_vm> dsaxena:  the first, is that i'm concerned about the seemingly large split between what is available for testing in the hwpacks, and nico's tree.
<dsaxena> in terms of patches added by LTs for board support?
<jstultz_vm> if i grab the natty current release from the linaro.org site.. i get the 1002 revision from john rigby's tree.. which is seemingly 3 months off from nicos
<dsaxena> that's not good. I think we're going to work on fixing that with the monthly kernel snapshots
<jstultz_vm> i just was having some issues w/ the android kernel, and wanted to see if the issue was present with nico's tree.. and it seemed to be, so wanting to narrow down if its a config issue, i grabbed the official tree
<jstultz_vm> and it worked fine, but its really old.. i know jcrigby has a wip (work in progress) tree that is more current, but it doesn't seem to be in the hwpacks
<jstultz_vm> so yea.. it would be good if we were not too much later then a week or so out from the hwpack images to nico's tree.
<dsaxena> yeah...I'll sync up with jcrigby about this
<jstultz_vm> second, talked a bit w/ alexander today..
<jstultz_vm> i had updated the linaro+android tree on friday, and asac was concerned that its awful close to a release to be adding churn.
<jstultz_vm> the problem is, that nico's tree has had quite a bit of last minute fixes landing..
<jstultz_vm> so to keep close sync w/ nico's tree i need to do the update.
<jstultz_vm> the down side is apparently there's a regression from a few weeks ago, and it may just be a config issue, but its a unwanted hiccup
<jstultz_vm> so asac was hoping there was a more clear "slush" -> "freeze" period for nico's tree, that i can then sync to to reduce churn in the linaro+android tree
<dsaxena> that's also part of the 11.11 plan :)
<jstultz_vm> because they're looking at a week out for a release and beagle hdmi is busted :(
<jstultz_vm> (but that is mostly my fault...)
<jstultz_vm> ok.. concerns vented.   thanks for listening :)
<dsaxena> i think some of the changes in nico's tree came from landing team? sounds like we need better agreement by all parties on when changes need to stop going into the kernel
<jstultz_vm> yea.. clearly there will be last minute changes, but its been a ton of churn..
<jstultz_vm> the hdmi on panda is looking good, but that changed how hdmi on beagle works, and the android tree needs a little extra time to sort those changes out
<dsaxena> so i was imagining one week freeze before release, but it might almost need more time than that?
<dsaxena> or we can freeze nico's tree for one week, release that, and then one week later release the android and LEB kernels?
<jstultz_vm> that can probably work. really it would be great to have something akin to the kernel tree..
<jstultz_vm> queue changes and have a 1week merge window, then 2 weeks slush, 1 week freeze.
<jstultz_vm> or something like that.
<jstultz_vm> that's probably wishful thinking :) but yea..  hopefully we'll be more in slush then merging..
<dsaxena> my concern is that a lot of work we're doing may not land quite within those windows. we may need to have parallel lines of development and I worry about overhead to manage all that
* dsaxena takes ACTION to convene a discussion of this with LT leads, Nico, jstultz_vm, jrcrigby
<jstultz_vm> dsaxena:  yea.. the one month release cycle will be tough..
<jstultz_vm> dsaxena:  we may need to sync w/ asac on that as well.. although i sense he was wanting the kernel team to just bake and deliver working stuff.
<jstultz_vm> and if its a regular schedule, they'll base around that
<dsaxena> jstultz_vm: ok, will sync up with him too
<jstultz_vm> dsaxena:  thanks!
<dsaxena> jstultz_vm: yvw.
<dsaxena> ok, anything else? mounir?
* dsaxena will poke other engineers to make sure they all know about the weekly meetings...
<mounir> no I am fine 
<mounir> dsaxena, yes we want to make sure all engineers attend the meetings and/or send weekly activity reports 
<dsaxena> sounds like another ACTION: remind engineers about importance to send in status reports and attend weekly meetings
<dsaxena> ok, I think we're done for now. tnx mounir, jstultz_vm
<jstultz_vm> dsaxena:  thanks!
<mounir> thx dsaxena - good night

WorkingGroups/KernelArchived/Meetings/2011-05-23 (last modified 2013-01-14 19:33:22)