20100928

September 28th, 2010, 13:00 UTC in #ubuntu-meeting.

Agenda

Action Items from last meeting

  • NCommander to clear out assigned bugs (c/o)

Special Items

  • None

Standing Items

Meeting Outcome

[13:59] <NCommander> #startmeeting
[13:59] <MootBot> Meeting started at 07:59. The chair is NCommander.
[13:59] <MootBot> Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE]
[13:59] <rsalveti> hey
[13:59]  * NCommander pokes GrueMaster dyfet ogra persia rsalveti and davidm
[14:00] <dyfet> that tickles
[14:00]  * ogra is here
[14:00] <mpoirier> o?
[14:00] <NCommander> [link] https://wiki.ubuntu.com/MobileTeam/Meeting/2010/20100928
[14:00] <MootBot> LINK received:  https://wiki.ubuntu.com/MobileTeam/Meeting/2010/20100928
[14:00] <mpoirier> o/
[14:00]  * GrueMaster grumbles over his ginormous coffee mug.
[14:01] <NCommander> [topic] Action Item Review
[14:01] <MootBot> New Topic:  Action Item Review
[14:01] <NCommander> [topic] NCommander to clear out assigned bugs
[14:01] <MootBot> New Topic:  NCommander to clear out assigned bugs
[14:01] <NCommander> c/o
[14:01] <ogra> well
[14:02] <ogra> we'Re hard frozen now for RC
[14:02] <NCommander> [topic] Standing Items
[14:02] <MootBot> New Topic:  Standing Items
[14:02] <ogra> so you should have some time
[14:02] <ogra> :P
[14:02] <NCommander> [link] http://people.canonical.com/~pitti/workitems/maverick/canonical-mobile.html
[14:02] <MootBot> LINK received:  http://people.canonical.com/~pitti/workitems/maverick/canonical-mobile.html
[14:02] <davidm> G'day
[14:02] <NCommander> morning davidm
[14:02] <ogra> persia still has 3 todos
[14:03] <ogra> making us touch the trendline !
[14:03] <ogra> time to postpone them i'd say
[14:03] <GrueMaster> Ouch.  It burns!
[14:03] <persia> I've been postponing spec work, since none of it directly affected the release.  if it's causing you guys issues, I'll focus on that this week.
[14:04] <ogra> juts makes us look bad in the chart
[14:04] <persia> can't have that :)
[14:04] <ogra> (the overall one)
[14:05] <NCommander> [topic] http://qa.ubuntu.com/reports/team-assigned/canonical-mobile-assigned-bug-tasks.html
[14:05] <MootBot> New Topic:  http://qa.ubuntu.com/reports/team-assigned/canonical-mobile-assigned-bug-tasks.html
[14:05] <ogra> the milestone release one lookd fine
[14:05] <ogra> *looks
[14:05] <ogra> NCommander, you could port erlang this week ;)
[14:06] <NCommander> :-P
[14:06] <NCommander> cna I move on?
[14:06] <ogra> sure
[14:06] <NCommander> [topic] Kernel Status (cooloney, mpoirier, lag)
[14:06] <MootBot> New Topic:  Kernel Status (cooloney, mpoirier, lag)
[14:06] <mpoirier> It's going to be a shorter one this week:
[14:06] <mpoirier> Marvel (mvl-dove):
[14:06] <mpoirier>  * Marvell new LSP 5.3.5 drop has been integrated in Lucid.  Maverick is coming soon.
[14:06] <mpoirier> Freescale (fsl-imx51):
[14:06] <mpoirier>  * nothing new this week.
[14:06] <mpoirier> OMAP3:
[14:06] <mpoirier>  * SGX driver has been integrated to maverick and proven to work properly.  Will be introduced in N.
[14:06] <mpoirier>  * ftrace graph tracing integration has been put on hold as it need a lot of work.
[14:06] <mpoirier> OMAP4:
[14:06] <mpoirier>  * B647890 - FIXED: Building error when turning off CONFIG_SMP
[14:06] <mpoirier>  * B633227 - ONGOING: Instabilities with highmem activated - we have 1G memory without highmem but need to disable SMP or L2 cache.
[14:07] <mpoirier>  * B637947 - ONGOING: A solution seems to be in line for panda with proper renaming of sound card.  Awaiting patch from TI.
[14:07] <mpoirier> ..
[14:07] <ogra> there is a dove metapackage upload missing
[14:07] <ogra> the new kernel most likely wont end up in the images
[14:07] <ogra> (bjf uploaded one yesterday)
[14:07] <persia> I thought I saw a request in -release to push that pre-RC
[14:07] <ogra> ah, ok
[14:08] <ogra> i didnt follow release that closely, the spam^Wqueuebot was so noisy
[14:08] <rsalveti> for bug 633227 we're still trying to find an easy way to reproduce it, to try to understand it better
[14:08] <ubottu> Launchpad bug 633227 in linux-ti-omap4 (Ubuntu Maverick) "instabilities with highmem activated" [High,New] https://launchpad.net/bugs/633227
[14:08] <GrueMaster> Hope they fix bug 634161 when they push it out.
[14:08] <ubottu> Launchpad bug 634161 in linux-mvl-dove (Ubuntu) "Unable to handle kernel paging request at virtual address 0ce003be - parport_pc driver load issue" [High,New] https://launchpad.net/bugs/634161
[14:08] <rsalveti> even when we remove highmem of the picture, we still face the issue
[14:08] <ogra> mpoirier, any plans to bring the IGEP2 stuff into an SRU post release ?
[14:09] <persia> GrueMaster, It's just a meta-upload.  Nothing substantive.
[14:09] <rsalveti> I was only able to make it work properly when running without highmem and no SMP, or no L2
[14:09] <mpoirier> ogra: the new paches or support the board ?
[14:09] <rsalveti> so probably a racing condition
[14:09] <ogra> mpoirier, yep
[14:09] <ogra> seems we have a lot progress on that platform
[14:09] <ogra> would be sad to lose the fixes
[14:09] <rsalveti> yup, would be good to push them forward
[14:09] <GrueMaster> persia: The bug only requires turning off the parport_pc driver in the config.
[14:09] <ogra> i understand they cant go into the release though
[14:10] <ogra> but probably in a post release SRU
[14:10] <rsalveti> post release would be fine
[14:10] <mpoirier> ogra: I think time started that - I'll look into it.
[14:10] <ogra> no hurry though
[14:10] <persia> GrueMaster, Right, but config is part of the source package for the kernel, not the -meta source package: unlikely the release team will approve such an upload, as it breaks the installer, etc.
[14:10] <NCommander> cna I move on?
[14:10] <rsalveti> and would be good if at least one of us could get one, to be able to check and test the kernel
[14:10] <ogra> post release is ... well after release :)
[14:10] <ogra> how is the progress on sound on OMAP4 ?
[14:11] <mpoirier> After conversation on alsa-soc yesterday it is clear the fix is not in the kernel.
[14:11]  * ogra didnt see much discussion in #ubuntu-arm recently
[14:11] <mpoirier> putting it there, even nicely, would be refused by upstream and therefore wasted time.
[14:11] <ogra> mind to update us on that ?
[14:12] <ogra> what exactly is required and what do we have
[14:12] <mpoirier> well, soc guys have purposely avoided putting a scheme forward for config in kernel.
[14:12] <persia> mpoirier, I read that differently: I read it as changing the kernel to ensure we had unique strings, and then creating userspace configs.
[14:13] <mpoirier> persia: I never denied that.
[14:13] <persia> Sorry, then I misunderstood your "the fix is not in the kernel".
[14:14] <mpoirier> persia: we won't apply the configs for the cards in the kernel.
[14:14] <mpoirier> the kernel names the card properly, which is pickup in userspace, where configs files are.
[14:14] <persia> berco has some config files that mostly work for panda: the same scheme could probably be extended to other boards.  There's an outstanding issue with saving the mixer state on reboot which needs investigation (if someone can replicate on beagle, I'm happy to investigate, but that requires someone with panda and beagle to confirm it's a parallel issue)
[14:15] <GrueMaster> Sounds like something for me to test.
[14:15] <ogra> so we'll use /usr/share/alsa* and the kernel gets a patch that /usr/share/alsa picks up stuff properly ?
[14:15] <mpoirier> we still have a problem in omap3 where cards for beagle C4, XM and gumstix are named the same.
[14:15] <mpoirier> ogra: yes.
[14:15]  * ogra is most concerned about OMAP4 atm :)
[14:16] <persia> ogra, Precisely.
[14:16] <ogra> OAMP3 is all nice to have
[14:16] <mpoirier> again, if we want to follow the same in omap3, a little bit of work is needed.
[14:16] <persia> mpoirier, Do we have differentiation for the differing wiremaps for the omap4 variants?
[14:16] <mpoirier> having a common solution for all our omapX platform would be nice.
[14:16] <GrueMaster> We have the same issue in Dove as well.
[14:17] <mpoirier> persia: omap4 variants ?
[14:17] <persia> We have the same issue for all SoCs where the kernel driver is reporting the codec chip rather than a unique per-wiremap string.
[14:17] <persia> mpoirier, I've at least heard "panda" and "blaze" mentioned on #ubuntu-arm.  I presume there are more design wins pending.
[14:17] <ogra> persia, we dont even have HW to test atm
[14:17] <persia> ogra, That hasn't been a blocker before :)
[14:17] <ogra> no ES2.0 blazes around currently
[14:18] <ogra> i'm sure you can get a wiremap though
[14:18] <mpoirier> i thought blaze was an omap3... I am I correct ?
[14:18] <ogra> no
[14:18] <ogra> blaze is omap4
[14:18] <mpoirier> ok.
[14:18] <GrueMaster> persia: W/o hw, we don't know what future designs will entail.
[14:18] <mpoirier> Again, the problem is exactly the same as omap3.
[14:18] <mpoirier> the name of the card is reported with no information on the platform.
[14:18] <ogra> blaze likely has a lot more options than panda
[14:18] <persia> GrueMaster, Right, but if we design something that is sufficiently modular that it will be a trivial patch for new designs, that's far better than assuming one device is the only way an SoC will behave.
[14:19] <ogra> but thats userspace stuff then as i understand
=== jorge_ is now known as jcastro_
[14:19] <GrueMaster> persia: Best to let the kernel expose all and tweak it in user space on a case by case basis.
[14:19] <GrueMaster> reduces kernel bloat, faster to fix.
[14:20] <persia> ogra, kernelspace needs to report a unique string.  userspace can then use that unique string to set configuration through autodetection.
=== jcastro_ is now known as jcastro
[14:20] <persia> GrueMaster, Indeed.  Precisely.
[14:20] <ogra> persia, right
[14:20] <mpoirier> persia: that's what I'be been saying.
[14:20] <persia> Right.
[14:20] <mpoirier> I had a chat with Irg on that front.
[14:21] <mpoirier> let see what they come back with.
[14:21] <NCommander> I hate to be a pain, but we have other things ot discuss, can we movw past kernel  bugs?
[14:21] <ogra> other things ?
[14:22] <mpoirier> ..
[14:22] <ogra> more important than kernel bugs ?
[14:22] <ogra> cant be
[14:22] <NCommander> on the agenda
[14:22] <ogra> :)
[14:22] <ogra> move
[14:22] <NCommander> :-p
[14:22] <NCommander> [topic] QA Status (GrueMaster)
[14:22] <MootBot> New Topic:  QA Status (GrueMaster)
[14:23] <GrueMaster> Not much to report.  Still finding obscure bugs.  Main focus has been on Dove audio (since omap audio is currently covered).
[14:23] <GrueMaster> I have a couple of bugs that I need to file, but absolutely no idea what to file against.
[14:23] <GrueMaster> Will try consulting #ubuntu-bugs again after the meeting.
[14:23] <NCommander> GrueMaster: pokem after the meeting and I'll help
[14:24] <NCommander> [topic]  * ARM Porting/FTBFS status ([[MichaelCasadevall|NCommander]], [[DavidSugar|dyfet]])
[14:24] <MootBot> New Topic:   * ARM Porting/FTBFS status ([[MichaelCasadevall|NCommander]], [[DavidSugar|dyfet]])
[14:24] <NCommander> pp[s
[14:24] <NCommander> *opps
[14:24] <ogra> NCommander, tsk
[14:24] <GrueMaster> (and I guess I'm done).
[14:24] <NCommander> Nothing new to report
[14:24] <ogra> let tobin finish :)
[14:24] <NCommander> [topic] QA Status (GrueMaster)
[14:24] <MootBot> New Topic:  QA Status (GrueMaster)
[14:24] <ogra> heh
[14:24] <ogra> hooray for complete messup :P
[14:25] <GrueMaster> Still looking into a bug query that is more useful to us.
[14:25] <NCommander> ;-P
[14:25] <GrueMaster> Not much else to report other than the images have been having build issues.
[14:25] <ogra> move ?
[14:25] <GrueMaster> move
[14:26] <NCommander> [topic]  ARM Image Status
[14:26] <MootBot> New Topic:   ARM Image Status
[14:26] <ogra> NCommander, skipping FTBFS ?
[14:26] <NCommander> ailed to build due to archive skew due to libgtk+2.0
[14:26] <NCommander> ogra: already did it
[14:26] <ogra> no
[14:26] <ogra> you mumble jumbled around
[14:27]  * ogra has been heavily grilled for FTBFS stuff by the release team
[14:27] <NCommander> [topic]  * ARM Porting/FTBFS stats NCommander/dyfet
[14:27] <MootBot> New Topic:   * ARM Porting/FTBFS stats NCommander/dyfet
[14:27] <ogra> the plplot disaster mader it all up to the release manager level
[14:27] <dyfet> I had worked on plplot yesterday, but it seems uneeded
[14:27] <ogra> the bug was lying dead for two weeks ...
[14:28] <ogra> please, if you cant work on bugs, tell me
[14:28] <ogra> it was pretty bad
[14:28] <persia> Um, about images, today is the day RC images spin, so it's kinda important to get them sorted...
[14:28] <ogra> persia, we'll get to that
[14:28] <persia> OK :)
[14:28] <ogra> NCommander, that was all i had to say about FTBFS
[14:28] <ogra> doko worked the whole wekend apparently to solve the issue
[14:29] <ogra> after we didnt move a bit on that bug
[14:29] <ogra> NCommander, move :)
[14:29] <NCommander> [topic]  ARM Image Status
[14:29] <MootBot> New Topic:   ARM Image Status
[14:29] <doko> ahh, that was the weekend before. just had the idea to ask upstream
[14:29] <ogra> should be building fine again with the next build
[14:30] <ogra> doko, thanks anyway for saving our butts
[14:30] <ogra> :)
[14:31] <ogra> the omap3 image had an issue with switching to linaro u-boot since that wasnt seeded yet
[14:31] <ogra> thats why we ended up without partition table on it
[14:31] <GrueMaster> oops
[14:31] <persia> So vstehle's bug is closed now?
[14:31] <ogra> beyond that the archive is in bad state currently
[14:31] <ogra> persia, fix committed
[14:31] <ogra> i'll set it to released after a testbuild
[14:31] <persia> OK.
[14:32] <ogra> once the archive has setteled
[14:32] <persia> Be glad you aren't trying to make this work on powerpc :)
[14:32] <ogra> also we're in RC preparation ... no manual builds ... Riddell is the master of antimony now
[14:32] <persia> Or, hrm.  Seems things are mostly caught up again.
[14:32] <ogra> all builds should be channelled or at least nodded off by him
[14:32] <ogra> (image builds i mean)
[14:33]  * Riddell feels the power
[14:33] <ogra> :)
[14:33] <ogra> NCommander, anything from you for images ?
[14:33] <GrueMaster> how long will it take to churn through the gtk update?
[14:33] <ogra> dove builds all fine ?
[14:34] <ogra> GrueMaster, just watch the build queue :)
[14:34] <GrueMaster> yea, that will keep me awake.  :P
[14:34] <ogra> its exciting :)
[14:35]  * ogra tickles NCommander 
[14:35] <davidm> NCommander, anything from you on images?
[14:36]  * davidm thinks NCommander needs a better Internet connection
[14:36] <ogra> or stronger coffee
[14:36] <ogra> well, we still have nearly 30 min ...
[14:36] <ogra> *twiddle*
[14:37] <davidm> We what is next on the agenda?
[14:37] <NCommander> sory, laptop haes me
[14:37] <davidm> Ah he's back
[14:37] <davidm> NCommander, , anything from you for images ?
[14:37] <NCommander> Worked on Dove ubiquity iconand other things
[14:38] <NCommander> go a patch but didn'tmake it in tfrom rc freeze, will test it some more and submit it before EOD today
[14:38] <NCommander> I think that's it for me
[14:39] <NCommander> [topic] ABO
[14:39] <MootBot> New Topic:  ABO
[14:40] <NCommander> folks, its spec drafting time
[14:40] <NCommander> onl6 news I have
[14:40] <ogra> and activity report time too :)
[14:40] <ogra> only rsalveti has submitted his yet
[14:41] <ogra> we others suck !
[14:41] <NCommander> anyway, anything else, or I am going to close out the meeting
[14:41]  * GrueMaster is ready to post
[14:41] <NCommander> once
[14:41] <NCommander> twice
[14:41]  * ogra not but will do so today
[14:41] <NCommander> three timeis the charm
[14:41] <NCommander> and good night folks
[14:41] <NCommander> #endmeeting
[14:41] <MootBot> Meeting finished at 08:41.

Action Items

  • None

Minutes

  • New LSP merged for Dove
  • New patches for OMAP and OMAP4
  • Audio still failing on ARM

Weekly Reports

Michael Casadevall (NCommander)

This Week

  • Added checks to casper to remove oem-config if installed on a live image on ARM to allow ubiquity's icon to properly show up
  • Began RC/final release Dove image testing
  • Did some work on packaging acceleration components for Dove

Ricardo Salveti (rsalveti)

This Week

Future

  • Work with Oliver to push the PowerVR SGX packages
  • Continue debugging memory and display issues on Panda
  • Help TI getting a proper package for PowerVR SGX Omap 4
  • Test PowerVR SGX packages for Omap 4

Tobin Davis (GrueMaster)

This Week

  • Retested alsa on Dove beta and compared results. Audio works fine at alsa level (speaker-test), but nothing from applications like Totem or Rhythmbox. Also discovered that changes made to expose speaker/headphone switch capability in Lucid no longer works. Further examination reveals a different audio codec on the board from X0. sigh.
  • Setup panda ES2 6L in place of defective beagleXM to test HDMI switchbox & 1440x900 monitor. Works great. Now I can test more omap4 stuff in parallel. Will need to add more cabling to make this semi-permanent for this cycle.

  • Promptly received new BeagleXM. Recabled and now it is online as well. Testing has resumed.
  • Couple of obscure bugs found.
    • Filed one regarding creating encrypted home directory causes /SWAP.swp to fill the drive. Seems user-setup uses dd to zero swap space without a size limitation. Bug 646421.
    • Another bug that was discovered happens when inserting a flash drive, a dbus error occures and often two file manager instances are launched. Unsure what package to file this against.

Next week

  • RC testing!
  • Tie up any loose ends. Check filed bugs for updated status.

David Sugar

This Week

  • some bug triages
  • some spec/Natty speculation
  • redid install of my Beagle
  • Hydrogen update released
  • gcl issues with thumb2

Next Week

  • Test install rc
  • patch for plt-scheme
  • some curren arm ftbfs's assigned to me
  • some bug triages

ARM/Meeting/2010/20100928 (last edited 2011-07-28 17:59:15 by davidm)