20110217

February 17th, 2011, 15:00 UTC in #ubuntu-meeting.

Agenda

Action Items from last meeting

  • NCommander to talk to the release team on which team to track w.r.t. to bugs (co)

Standing Items

Meeting Outcome

[15:00] <NCommander> #startmeeting
[15:00] <MootBot> Meeting started at 09:00. The chair is NCommander.
[15:00] <MootBot> Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE]
[15:00]  * rsalveti waves
[15:00] <NCommander> [link] https://wiki.ubuntu.com/MobileTeam/Meeting/2011/20110217
[15:00] <MootBot> LINK received:  https://wiki.ubuntu.com/MobileTeam/Meeting/2011/20110217
[15:00] <NCommander> Who's here this morning
[15:01] <janimo> hello
[15:02]  * NCommander pokes GrueMaster ogra persia 
[15:02] <NCommander> and davidm
[15:02]  * GrueMaster is pre-caffeinated.
[15:02] <GrueMaster> davidm may be running late.
[15:03] <GrueMaster> (he mentioned it last night in a different channel).
[15:03] <NCommander> ah
[15:03]  * ogra pokes NCommander back
[15:03]  * NCommander is poked
[15:03] <NCommander> k
[15:03] <NCommander> let's go
[15:03] <NCommander> [topic] Action Item Review
[15:03] <MootBot> New Topic:  Action Item Review
[15:04] <NCommander> [topic] NCommander to talk to the release team on which team to track w.r.t. to bugs (co)
[15:04] <MootBot> New Topic:  NCommander to talk to the release team on which team to track w.r.t. to bugs (co)
[15:04] <NCommander> Done. Thanks ogra for bringing it up at the release meeting
[15:04] <NCommander> [topic] Standing Items
[15:04] <MootBot> New Topic:  Standing Items
[15:04] <ogra> i didnt
[15:04] <NCommander> ogra: I thought you said you did in #ubuntu-release
[15:04] <rsalveti> I had one action to talk with cooloney to work on the DVI patches, and he's already working on it
[15:04] <ogra> i just changed our report but that didnt mean that kate used the other tracker
[15:04]  * NCommander coughs
[15:05] <ogra> she does now after our conversation in #ubuntu-release, yes
[15:05] <NCommander> I also did talk with skaet_ on this the other day
[15:05] <ogra> i dodnt think she noticed it in the report
[15:05] <NCommander> ah
[15:05] <ogra> (which is why you should talk to her ;) )
[15:05] <NCommander> either way, its resolved
[15:05] <ogra> yes
[15:05] <ogra> :)
[15:06] <NCommander> [link] http://people.canonical.com/~platform/workitems/natty/ubuntu-armel.html
[15:06] <MootBot> LINK received:  http://people.canonical.com/~platform/workitems/natty/ubuntu-armel.html
[15:06] <NCommander> [link] http://people.canonical.com/~platform/workitems/natty/ubuntu-armel-natty-alpha-3.html
[15:06] <MootBot> LINK received:  http://people.canonical.com/~platform/workitems/natty/ubuntu-armel-natty-alpha-3.html
=== Nafallo_ is now known as Nafallo
[15:07] <NCommander> no comments on the workitem tracker?
[15:07] <ogra> we look pretty good i'd say
[15:08] <NCommander> cool
[15:08] <ogra> we need to move items next week though
[15:08]  * NCommander moves on
[15:08] <NCommander> [topic] Unity 2D Status
[15:08] <ogra> if there are any we dont make
[15:08] <MootBot> New Topic:  Unity 2D Status
[15:08] <rsalveti> also good :-)
[15:08] <ogra> its on the images since yesterday
[15:08] <ogra> i'll care for the new upstream next week
[15:08] <GrueMaster> And semi-works too.
[15:08] <ogra> the dash isnt fixed in that version
[15:09] <ogra> upstream has the fix already afaik
[15:09] <ogra> upstream is at 3.4 while we are at 3.2
[15:09] <GrueMaster> Is une-efl still seeded in the image or no?
[15:09] <ogra> no
[15:09] <ogra> its gone to universe already
[15:09] <rsalveti> oh, cool
[15:09] <GrueMaster> Ok.  Hadn't checked due to other high priority testing.
[15:10] <GrueMaster> It was useful in that unity-2d can't launch apps yet.
[15:10] <ogra> we're waiting for QT for NEON stuff
[15:10] <ogra> it launches apps for me
[15:10] <ogra> just doesnt show the dash
[15:10] <GrueMaster> Try launching a terminal.
[15:10] <ogra> well, shows it but doesnt get you search results
[15:10] <ogra> i dont have a terminal in the launcher
[15:11] <ogra> and the dash doesnt work
[15:11] <ogra> what i click on in the launcher works here
[15:11] <GrueMaster> Filemanager also crashes, but I think that is due to the icons bug.
[15:11] <ogra> hmm, works for me too
[15:12] <GrueMaster> (can't tell as I have no terminal. )  :P
[15:12] <ogra> i have issues with it on my x86 lappie, but not on the panda
[15:12] <ogra> but as i said. new upstrea pending that will solve the dash issue
[15:13] <GrueMaster> Open the file manager and go to /usr/share/applications.  It will crash trying to display the .desktop icons.
[15:13] <GrueMaster> (which was the alternate way to launch the terminal)
[15:14] <ogra> ah, i didnt try that
[15:14] <ogra> well, overall unity-2d looks ok apart from known bugs
[15:15] <GrueMaster> yep
[15:15] <rsalveti> that's fine
[15:15] <ogra> NCommander, move (if there arent any other issues)
[15:15] <rsalveti> now we need to blog about it :-)
[15:15] <rsalveti> or wait alpha-3
[15:15] <NCommander> [topic] Kernel Status (cooloney, rsalveti)
[15:15] <MootBot> New Topic:  Kernel Status (cooloney, rsalveti)
[15:15] <ogra> rsalveti, davidm alrteady did ;)
[15:15] <ogra> see planet
[15:15] <rsalveti> new kernel for omap3, based on 38-rc5, should work quite same way for us
[15:16] <rsalveti> ogra: nice, will check
[15:16] <rsalveti> for omap 4 cooloney is still working on the DVI patches
[15:16] <rsalveti> once he's done he'll update the ti-omap4-dev branch
[15:16] <ogra> did anyone test the ppa kernel yet ?
[15:16] <rsalveti> janimo?
[15:16] <janimo> I ran it today
[15:17] <janimo> and booted, but reverted to .35 after that
[15:17] <rsalveti> should work, as it's basically upstream
[15:17] <ogra> what didnt work ? ;)
[15:17] <janimo> I tried turning off proc1
[15:17] <janimo> and the command hung
[15:17] <janimo> even if the board did not
[15:17] <ogra> no PM in that kernel
[15:17] <janimo> so no CPU hitplug?
[15:17] <ogra> so thats not surprising
[15:17] <janimo> the sysfs file is exposed though
[15:17] <rsalveti> janimo: don't expect pm to work :-)
[15:17] <ogra> i would think hotplug is a PM feature
[15:17] <janimo> it should not if it does not work
[15:17] <rsalveti> well, pm is there
[15:17] <GrueMaster> That's the same on .35 kernel.
[15:17] <rsalveti> just doesn't work :-)
[15:18] <ogra> GrueMaster, on the natty .35 one ?
[15:18] <janimo> also I saw tiwlann errors in dmesg but maybe I did not nstall some firmware pkg
[15:18] <ogra> there it should work
[15:18] <janimo> did not test much else
[15:18] <rsalveti> janimo: could be
[15:18] <GrueMaster> Mmm.  May have been mav.
[15:18] <janimo> as I wanted to build mono and did not want to risk an oops or something
[15:18] <ogra> natty .35 has the PM patches
[15:18] <rsalveti> natty should be fine
[15:18] <GrueMaster> I hit it testing f-spot last night.
[15:18] <rsalveti> I tested yesterday while testing mono
[15:18] <ogra> .38 wont have or get them though
[15:18] <janimo> rsalveti, so tiwlan on panda should work without TI PPA now?
[15:18] <rsalveti> put second cpu off and still going fine
[15:19] <janimo> that would be nice
[15:19] <rsalveti> janimo: yup, just need the right firmware
[15:19] <janimo> hmm, so a restriced package just as on x86?
[15:20] <janimo> without needed to fetch from the TI ppa?
[15:20] <ogra> likely that way
[15:20] <persia> Is the firmware merged into linux-firmware?
[15:20] <rsalveti> yeah
[15:20] <ogra> persia, not yet
[15:20] <janimo> anyway we'll discuss in #arm  later
[15:20] <ogra> and i think the codebase isnt fully merged either yet
[15:20] <rsalveti> but no other news from TI
[15:20] <rsalveti> nor Linaro, about newer changes for OMAP 4
[15:20] <ogra> we have to wait for TI's "go"
[15:20] <janimo> do TI make a new code drop in the PPA sometime?
[15:21] <janimo> ok
[15:21] <rsalveti> once we have the "go" we can update the branch and publish it to the ppa
[15:21] <ogra> janimo, for natty, yes, but note they are not bound to release schedule
[15:21] <janimo> sure
[15:21] <ogra> its a PPA after all
[15:21] <ogra> just needs to be ready by release
[15:21] <janimo> I see tiwlan firmware on x86 so maybe I just need to get that pkg on the panda too
[15:21] <janimo> if it is built against this kernel version
[15:21] <ogra> try it :)
[15:21] <rsalveti> current schedule ways that we'll have a working 38 kernel days before the final release hehe :-)
[15:22] <rsalveti> *says
[15:22] <ogra> if it works we can just pull it over
[15:22] <rsalveti> we'll see :-)
[15:22] <rsalveti> move?
[15:22] <ogra> well, as long as we have sound and display we can drop .35 imho
[15:22] <NCommander> [topic] QA Status (GrueMaster)
[15:22] <MootBot> New Topic:  QA Status (GrueMaster)
[15:22] <rsalveti> ogra: probably
[15:23] <ogra> the other features will just stream in
=== yofel_ is now known as yofel
[15:23] <GrueMaster> This week has been busy with a flurry of kernel updates from the kernel team on older releases.
[15:23] <ogra> and network :)
[15:23] <ogra> oh, we moved =
[15:23] <persia> And USB, please.
[15:23] <GrueMaster> sigh.
[15:23] <ogra> persia, network expects USb
[15:24] <GrueMaster> Kernel updates for Lucid and Maverick on Dove have a serious regression in that they corrupt the console output.
[15:24] <rsalveti> ouch
[15:25] <GrueMaster> Also working on a revised script for pulling a bug list from LP that better suits the teams needs.  WIP.
[15:26] <GrueMaster> May not be able to get everything as some search query capabilities appear to be missing from lplib.
[15:27] <ogra> file bugs ;)
[15:27] <GrueMaster> Linaro team has me hopping on some old (karmic) bugs.  Apparently they can't reproduce the failures on natty with newer platforms.  Go figure.
[15:27] <ogra> heh
[15:27] <GrueMaster> ogra: planned on it once I can dive back in.  Kernel testing preempted this for the moment.
[15:28] <GrueMaster> That's it for QA.
[15:28] <ogra> great
[15:29] <NCommander> [topic] ARM Porting/FTBFS status (NCommander, janimo)
[15:29] <MootBot> New Topic:  ARM Porting/FTBFS status (NCommander, janimo)
[15:29] <janimo> triaged/tested and closed a few bugs in yesterday's porting jam
[15:29] <NCommander> Did some debugging work on mono, made some interesting progress
[15:30] <janimo> fixed one more alignment related FTBFS during the week
[15:30] <NCommander> janimo: any chance you looked at kdebindings? I'm stumped on that build failure, and a second set of eyes might help
[15:30] <janimo> NCommander, ok, will take a look
[15:30] <ogra> the mono smp finding was awesome, kudos to rsalveti and GrueMaster !!
[15:30] <janimo> yes
[15:31] <NCommander> ogra: unfortunately, even adding memory barriers to mono doesn't quite fix it :-(
[15:31] <NCommander> If anything, it breaks worse now
[15:31] <rsalveti> urgh
[15:31] <janimo> I think both banshee bugs have the same root and should be dupes
[15:31] <GrueMaster> It will be a major step in the right direction.
[15:31] <rsalveti> NCommander: did you talk with upstream already?
[15:31] <NCommander> rsalveti: yeah, and we cooked a patch
[15:31] <janimo> NCommander, did they have a patch in 2.8 or 2.10 already?
[15:31] <NCommander> http://pastebin.com/aRZ0R2HK - for reference
[15:31] <MootBot> LINK received:  http://pastebin.com/aRZ0R2HK - for reference
[15:32] <ogra> NCommander, it works with nosmp
[15:32] <NCommander> janimo: no charges to the code in 2.8 or 2.10, I am working against 2.10 for the moment with intergration with upstream
[15:32] <NCommander> ogra: yes, I know :-)
[15:32] <ogra> rthats a massive progress
[15:32] <rsalveti> we can just say, turn one cpu off in case you want to play music with banshee :-)
[15:33] <janimo> NCommander, is that for armv6?
[15:33] <ogra> rsalveti, if PM works ;)
[15:33] <NCommander> janimo: and great, we're continuing to work on it :-/
[15:33] <rsalveti> ogra: haha, true
[15:33] <janimo> on ARMv7 memory barrier is DMB I think no coprocessor instrucitons involved
[15:33] <ogra> but yeah, gross hack but would work
[15:34] <NCommander> janimo: we decided for the sake of upstream to do it with the MCR instruction so they don't have to specifically decide at runtime which ARM variant we're on :-/
[15:34] <janimo> NCommander, from what I've seen in the code they test for armv7 in a few places already, but it's ok
[15:35] <NCommander> janimo: I think I figured out why our code is busted though
[15:35] <janimo> what seems interesting is that on their CI site (monkeywrench) 2.8/arm/ubuntu 9.04 passes tests
[15:35]  * NCommander is now more awake and relooking at the patch
[15:35] <davidm> G;day NCommander ogra janimo rsalveti GrueMaster persia
[15:35] <ogra> how about discussing that offline ?
[15:35] <ogra> seems to get a bit to much into detail for the meeting
[15:35] <GrueMaster> janimo: 9.04 was armv5 iirc.
[15:36] <janimo> GrueMaster, oh that would explain a lot of things
[15:36] <ogra> yes
[15:36] <ogra> 9.04 was essentially debian recompiled
[15:36] <janimo> they don;t have more arm boards to set up more CI boxes. too bad
[15:37] <GrueMaster> They can order panda's.  :P
[15:38] <ogra> move ??
[15:38] <NCommander> [topic] ARM Image Status (ogra, NCommander)
[15:38] <MootBot> New Topic:  ARM Image Status (ogra, NCommander)
[15:39] <ogra> failed today (as all arches did)
[15:39] <GrueMaster> yep
[15:39] <ogra> thanks to computer-janitor
[15:39] <ogra> lool fixed that afaik
[15:41] <GrueMaster> Zzzzzz
[15:41] <NCommander> [topic] AOB
[15:41] <MootBot> New Topic:  AOB
[15:41] <ogra> flash-kernel has issues with panda with the new kernel
[15:41]  * NCommander groans
[15:42] <ogra> specifically with linaro using a unified omap flavour
[15:42] <NCommander> ogra: ugh, they changed omap4 to use 'omap'?
[15:42] <ogra> but also with the kernel reporting omap4 boards differently
[15:42]  * NCommander notes that breaks a ton of sanity checks ...
[15:42] <ogra> NCommander, yes
[15:42] <ogra> that was on their roadmap for quite some time
[15:43] <GrueMaster> How will that even be possible?
[15:43] <ogra> ??
[15:43] <ogra> it apparently works
[15:43] <GrueMaster> Can one kernel run both platforms?
[15:43] <ogra> ask linaro kernel devs how they did it
[15:43] <ogra> sure
[15:43] <GrueMaster> ok
[15:43] <rsalveti> sure, at least one omap kernel for all
[15:43] <rsalveti> that's the upstream direction
[15:43] <ogra> well, i guess it wont run on omap2
[15:44] <rsalveti> hehe, true
[15:44] <rsalveti> omap 3 and omap 4
[15:44] <ogra> but who has the hw to test that anyway
[15:44] <GrueMaster> Oh, boohoo.
[15:44] <rsalveti> nobody cares for omap 2 anymore
[15:44] <ogra> anyway, just wanted to mention it
[15:44] <ogra> aob seemed appropriate
[15:44] <GrueMaster> Interesting.
[15:44] <rsalveti> and soon just one u-boot too
[15:44] <ogra> we need to look into that
[15:44] <ogra> and most likely reqork some of the subarch detection
[15:44] <rsalveti> x-loader should still be different for the moment, as it needs the hardware id
[15:44] <GrueMaster> ...and in the darkness, bind them.
[15:45] <ogra> (as we need to do for blaze anyway to fix the open bug)
[15:45] <janimo> these are all post natty right? unified uboot and kernel
[15:45] <ogra> janimo, fixing flash-kernel is natty
[15:45] <ogra> bootloaders arent
[15:45] <rsalveti> janimo: unified probably
[15:45] <janimo> right, I mean omap3/4 unification
[15:45] <janimo> binaries
[15:45] <ogra> and will likely require HW changes anyway
[15:45] <ogra> the rom code wont just handle unified x-loader
[15:46] <rsalveti> nops
[15:46] <ogra> janimo, unified kernels are reality in linaro
[15:46] <rsalveti> but only one u-boot and kernel is cool enough already
[15:46] <ogra> and flash-kernel currently breajs for them
[15:46] <ogra> 'breaks
[15:46] <rsalveti> lool seems to be fixing it
[15:46] <ogra> lool fixes upstream
[15:47] <ogra> we need to fix it downstream for the moment
[15:47] <ogra> debian doesnt support omap4 yet
[15:47] <ogra> and i dont even thing omap3 atm
[15:47] <persia> So we should have unified kernels in "linux" for natty+1?
[15:48] <rsalveti> at least for omap
[15:48] <ogra> yes
[15:48] <persia> Cool.
[15:48] <ogra> possibly +2 who knows
[15:48] <ogra> in any case linaro has it today
[15:48] <ogra> anyway, thats all from me
[15:49] <rsalveti> I'm also done
[15:49] <NCommander> anything else?
[15:49] <rsalveti> move! :-)
[15:52] <NCommander> #endmeeting

Action Items

Weekly Reports

Michael Casadevall (NCommander)

  • Mono debugging continues
  • Patch pilot work on Wednesday

Minutes

  • UNE 2D demoted to universe
  • Mono debugging continues
  • mono bug determined to be SMP specific
  • tiwlan driver available

ARM/Meeting/2011/20110217 (last edited 2011-07-28 17:57:38 by davidm)