This month's meetings

WorkingGroups/Kernel/Meetings
<< <  2011 / 4 >  >>
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  

Meeting info

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

Attendance

  • Arnd Bergmann: x
  • Dave Martin: x
  • Deepak Saxena: x
  • Grant Likely:
  • Jason Hui:
  • John Stultz: x
  • Linus Walleij: x
  • Manjunath Kondaiah:
  • Mounir Bsaibes: x
  • Nicolas Pitre:
  • Niklas Hernaeus: x
  • Rajendara Nayak:
  • Thomas Abraham:
  • Shawn Guo: x
  • Venkatraman Sathiyamoorthy:

Agenda

  1. Announcement / Upcoming Release Dates
  2. Pending Action items
  3. Around the table: Work Items updates, what's next
  4. BUGS review

Action Items from Previous Meeting

Minutes

https://wiki.linaro.org/WorkingGroups/Kernel/Thumb2Guide

  • Deepak will finish setting up the Sessions for Orlando Connect and will send email to session owners to work on creating an outline
  • we're going to be changing how we work in the KWG a little bit starting at Linaro connect, what we'll be doing in the future, is setting up small mini-teams to go tackle a specififc problem across all SOCs at the same time for example, we'll work on getting pinmux supported on all member SOCs
  • clk framework will be done by the PMWG
  • There will be a session about how to fir upsreaming with Linaro process. Mounir will send a link soliciting feedback.
  • arnd:
    • I have spent a lot of time merging stuff and verifying the results
    • the randconfig work is starting to pay off for me, I could take the for-next branch and merge it with the randconfig branch and see everything that broke :)

    • hopefully we got the ball rolling on the CMA patches again, a few people sent ACKs back to that, but it might be too late for 3.2
  • Shawn
    • back from holiday, I'm preparing the pull-request for imx5-dt and imx6q series to Arnd. And I also start looking into pinctrl subsystem for migrating imx iomuxc to it.
  • dmart
    • was working on the amba driver module autoloading issue, have acks an about half of the patches
    • also completed finished the Thumb-2 kernel porting advice. What's the best way to advertise this to the landing teams?
      • will include that in this week's kernel summary if it didn't make it on last weeks
  • jstultz_vm
    • focusing on the scheduler freezer idea right now to try to promote the wakelock alternative idea i had..
    • also having some discussions on the pm list about what userland approaches we could also try.
    • need to re-ping folks on the config merge script.. last submission didn't get any feedback..
    • need to figure out what the story will be for 11.10 android kernel, since nico is thinking 3.1 for the linaro tree.. so i might have to pull the unofficial android fwd port that andy green did
  • linusw
    • pin control core + pinmux stabilization for the merge window
    • then I'm researching pin control generics for biasing, driving etc.
    • found out what was blocking NHE:s work on DT

    • random reviews of random stuff and thinking out loud as usual
  • nhe
    • I am on dt for snowball. Used some three days for some adm that have been stacked up. Unless a Linaro <-> STE presentation goes as Linaro work, then 1.5 days

    • The bp is updated and the items should be trackable and nice.

Actions

  • Mounir to mention the thumb2 guide on the weekly summary

IRC Log

* arnd (arnd@nat/ibm/x-izbljwvbptcnplnm) has joined #linaro-kernel
* nhe (~nh@57.79.167.151) has joined #linaro-kernel
* dsaxena (~dsaxena@c-24-21-171-119.hsd1.or.comcast.net) has joined #linaro-kernel
<dmart> hi all
<dsaxena> good morning/afternoon/evening
<mounir> hi
<shawnguo> hello everyone
<jstultz_vm> dsaxena:  hey
<dsaxena> greetings shawnguo , jstultz_vm , dmart , mounir
<dsaxena> anyone else around?
<nhe> Hi...
<dsaxena> hi nhe
<linusw_> I'm here
<dsaxena> greetings linusw_
<dsaxena> ok, this might be a short meeting. let's start with a couple of quick announcements
<dsaxena> 1. I've started creating session blueprints for Linaro connect. We have a total of about 15 slots, I've created 8 BPs so far
<dsaxena> i'll make a few more today and send email to session owners to work on creating an outline
* 36DAAG1V2 (~a0131687@122.179.109.232) has joined #linaro-kernel
<dsaxena> we also have about 15 hours of hacking time available and we should start thinking about how we want to split that.
<dsaxena> which leads to the second announcement.
<dsaxena> we're going to be changing how we work in the KWG a little bit starting at Linaro connect
<dsaxena> currently we have everyone working on their own, targeting somewhat completely different deliverables
<dsaxena> this is making it hard for Linaro as an organization to say "We will have feature  X up and running on all SOCs by this timeframe"
<dsaxena> what we'll be doing in the future, is setting up small mini-teams to go tackle a specififc problem across all SOCs at the same time
<dsaxena> for example, we'll work on getting pinmux supported on all member SOCs
<linusw_> cool
<dsaxena> with linusw_ as the lead for the mini-team.
<dsaxena> this wouldn't be everyone's 100% effort. we'll figure out the details at LC
<dmart> that generally sounds like a good idea -- this is sort of what happens anyway, but without the team structure
<shawnguo> yeah, just like the mini-team for device tree led by Grant
<dsaxena> yep
<dsaxena> the hope is also that by having us work in parallel, we increase chances of new infrastructure we work on to get pushed up stream as there is more than one user for it
<linusw_> so like with DeviceTree I get sleepless nights if Torvalds doesn't pull in my code ;-)
<dsaxena> and also to build more of a sense of a team working in unison
<dsaxena> linusw_, :)
<dsaxena> other items will include transitioning regulator drivers to DT, inter-CPU IPC drivers (once we have a generic solution), and whatever other cleanup/consolidation efforts we dig into
<arnd> what about DT support for clocks? is that something we expect to have a patch for before LC, or should we work on that too?
<dsaxena> arnd, we'll have to work on that too. gcl mentioned some ideas around clock support @ LPC, we need someone to take lead.
<linusw_> arnd: do we have a way forward on generic clk device/subsystem now then?
<dsaxena> linusw_, mike t. and amit in the pmwg are taking lead of that . we're sharing rooms with them @ connect, so there will likely be a joint hacking session on that
<linusw_> OK sounds like a plan, is tglx coming?
<dsaxena> hmm, unsure, I will ping him
<dsaxena> ok, one last thing on my end is that Mounir, Michael H., myself, and some other tech leads have put together a little document talking about the upstream kernel process
<dsaxena> and different approaches for it to map to Linaro's process
<dsaxena> we'll send out a link and would really like input
<dsaxena> we have a session at LC to talk about this, the hope is to have some agreements across the organization before we get there so it is not just us talking about how the process doesn't work :)
* manjugk (cb653d07@gateway/web/freenode/ip.203.101.61.7) has joined #linaro-kernel
<mounir> dsaxena, the hope is to get an agreement at connect how to fit upstreaming with Linaro process as best as we can
<dsaxena> mounir, ack. i'm looking forward to having that discussion
<dsaxena> ok, let's do a go around, let us know what you're focusing on and if there's any action you need from anyone else to help you
<dsaxena> i'm going to be finishing up creating LC sessions and will need help on filling out the outlines
<dsaxena> also working on getting more requirements into our requirements system and working on an overall kernel consolidation roadmap
<dsaxena> arnd, want to start?
<linusw_> dsaxena: up in the log I noted something like more consolidation == more subsystems needing to be created
<arnd> sure
<arnd> I have spent a lot of time merging stuff and verifying the results
<arnd> the randconfig work is starting to pay off for me, I could take the for-next branch and merge it with the randconfig branch and see everything that broke :)
<dsaxena> arnd, have you subscribed to the kernel ci build's RSS feed for arm-soc?
<arnd> I've manage to get a few of the randconfig branches into maintainer trees, but there is much more work, and I found another ~30 problems checking linux-next with it
<arnd> yes
<arnd> but it seems that always breaks with git errors in the last weeks
<arnd> I've also helped out a few people doing patch reviews last week, but nothing spectacular there
<arnd> hopefully we got the ball rolling on the CMA patches again, a few people sent ACKs back to that, but it might be too late for 3.2
<arnd> I think that's all for me this time
<dsaxena> arnd, tnx. i'll look into the git issues
<dsaxena> with the builds
<dsaxena> dmart, you're it
<dsaxena> dmart, still around?
<dsaxena> shawnguo, want to give a quick update on what you're working on?
<shawnguo> dsaxena: sure
<shawnguo> back from holiday, I'm preparing the pull-request for imx5-dt and imx6q series to Arnd.  And I also start looking into pinctrl subsystem for migrating imx iomuxc to it.
<linusw_> shawnguo: much appreciated!
<dmart> dsaxena: yep, sorry -- I'll wait for shawnguo
<shawnguo> linusw_: I hope Linus T will pull your bits for v3.2 ;)
<shawnguo> dmart: that's it for me
<dsaxena> shawnguo, tnx!
<dsaxena> dmart, ping
<dmart> dsaxena: mostly last week, I was working on the amba driver module autoloading issue
<dmart> I have acks an about half of the patches
<dmart> This would have solved the aaci brokenness in the 11.09 image for vexpress (though really it was a config regression, it exposed the problem)
<dmart> I also completed finished the Thumb-2 kernel porting advice.  What's the best way to advertise this to the landing teams?
<linusw_> dmart: what about a Googledoc and share it "view" with everyone?
<shawnguo> dmart: great document, thanks a lot
<dmart> We've been building Thumb-2 for all the non-LT kernels for a while now, but except for omap our tree lacks the scariest code
<dsaxena> dmart, put it on the linaro wiki and post to linaro-dev and cc: techleads@linaro.org
<dmart> linusw_: Currently it's on the wiki -- I posted to linaro-dev, but that may tend to be buried in the noise
<dmart> I'll try the techleads alias
<linusw_> dmart: maybe post it to linux-arm-kernel@infradead.org? :-)
<dsaxena> mounir, we should make sure to include that in this week's kernel summary if it didn't make it on last weeks
<dsaxena> linusw_, dmart +1 on the l-a-k posting
<dmart> linusw_: Hmmm, that could be a good idea
<dmart> It's only draft quality -- if anyone has any comments, suggestions or fixes
<dmart> then please let me know
<mounir> dsaxena, ok
<dmart> There will also be mistakes, though hopefully not too many :P
<dsaxena> dmart, will do!
<linusw_> dmart: don't worry Russell will let you know all mistakes :-)
<dmart> Last time the topic came up, Russell decided that Thumb-2 was "too weird" for him
<linusw_> dmart: I can forward a copy to Rabin Vincent @STE he sort of like Thumb2 stuff I think
<dmart> There's nothing secret, so feel free to forward links
<dsaxena> jstultz_vm, how goes android hacking?
<dsaxena> dmart, tnx for the work on the doc, i think it would also make a good presentation for conferences...
<jstultz_vm> dsaxena:  ok.. focusing on the scheduler freezer idea right now to try to promote the wakelock alternative idea i had..
<jstultz_vm> also having some discussions on the pm list about what userland approaches we could also try.
<jstultz_vm> (although i'm skeptical)
<jstultz_vm> dsaxena:  need to re-ping folks on the config merge script.. last submission didn't get any feedback..
* CosmicPenguin (~nobody@soa.codeaurora.org) has joined #linaro-kernel
<jstultz_vm> dsaxena:  basically that's all right now..  need to figure out what the story will be for 11.10 android kernel, since nico is thinking 3.1 for the linaro tree.. so i might have to pull the unofficial android fwd port that andy green did
<jstultz_vm> (or i guess has in his tree, was done internally by TI)
<linusw_> jstultz_vm: moving ahead of Android official baseline will anyway be necessary for things like Cortex A15 support...
<dsaxena> is it just an android fwd port or does it have other bits in it too? I always worry a bit about vendor-internal Android trees
* asac_ has quit (Ping timeout: 260 seconds)
<jstultz_vm> linusw:  yea.. with android.git.kernel.org still down, its a bad spot.
<jstultz_vm> dsaxena:  i think its just the forward port..
<dsaxena> ok, good to hear.
<jstultz_vm> dsaxena:  anyway, i'm a little worried about stability and bugs, so 11.10 might be a rougher release.
<jstultz_vm> hopefully 3.1 is solid enough already..
<dsaxena> npitre, jstultz_vm can we do an -rc pre-release to play with this week?
<dsaxena> dmart, sounds like we should pull in your amba patches into 11.10?
<jstultz_vm> dsaxena:  yea, i was planning on having something ready this week for testing..
<dsaxena> jstultz_vm, ok, great
<dsaxena> tnx!
<dsaxena> linusw_, what's on your plate this week?
<linusw_> dsaxena: pin control core + pinmux stabilization for the merge window
<linusw_> dsaxena: then I'm researching pin control generics for biasing, driving etc.
<linusw_> dsaxena: going to look up dept. of applied electronics @ Lund University and ASIC dept. @ STE for that stuff
<linusw_> dsaxena: found out what was blocking NHE:s work on DT
<linusw_> dsaxena: random reviews of random stuff and thinking out loud as usual
<dsaxena> linusw_, ping shawnguo, rajendra, thomas abraham, to get input on other SOCs
<dsaxena> linusw_, what was the DT issue?
<linusw_> dsaxena: compat node didn't match compiled-in value, then the error message was on earlyprint and earlyprint didn't work
<linusw_> dsaxena: I'm not very good at coordination, how do I know the key person for each member platform?
<dsaxena> so past it and we can boot with DT now?
<linusw_> dsaxena: I sorta need a list to go over or so
<dmart> dsaxena, npitre: I have no change requests on that series from anyone, and the changes are pretty straightforward -- so I think these could go into 11.10
<dsaxena> linusw_, ask me for now. :) we'll sort it out @ connect
<linusw_> dsaxena: the HREF500 boots with DT I think nhe still has some problem with the Snowball so I guess I fix that next...
<dsaxena> linusw_, and the go-to person might change per project too as people rotate into differnt things
<dsaxena> linusw_, ok
<linusw_> dsaxena: OK I guess they will come to me too...
<linusw_> dsaxena: and if nothing happens I'll force-convert their platforms myself haha
<dmart> dsaxena, npitre: For now though, it's not very interesting -- we do (or should) build most amba drivers into the kernel anyway.  And once the relevant drivers are fully DT-ised, I believe autoloading should be taken care of the OF way anyhow, at least for fdt-based kernels
<linusw_> dsaxena: that's all for me.
* 36DAAG1V2 has quit (Ping timeout: 245 seconds)
<dsaxena> linusw_, tnx
<dsaxena> i think that was everyone who's here today?
<nhe> Me also...
<dsaxena> nhe, hi
<dsaxena> nhe, sounds like some good progress on DT!
<nhe> Short then, I am on dt for snowball.  Used some three days for some adm that have been stacked up. Unless a Linaro <-> STE presentation goes as Linaro work, then 1.5 days
<nhe> mounir: The bp is updated and the items should be trackable and nice.
<mounir> nhe: thank you for handling
<nhe> ok, next, if there is someone left...
<dsaxena> nhe, i think that presentation counts, it's important to get intenral teams up and running on DT
<dsaxena> i think that's it?
<nhe> dsaxena: not dt, Linaro, slightly different...
<dsaxena> nhe, ah ok
<dsaxena> ok, we're a bit over 1 hr, so thank you all for your time
<linusw_> thx

WorkingGroups/KernelArchived/Meetings/2011-10-10 (last modified 2013-01-14 19:41:27)