Developer Platform Team Meeting

Agenda

  • Actions from last meeting - Steve chasing
  • Engineer progress reports - All team members
  • Work Item Progress and Team Issues discussion - Steve and JamieBennett questions to the team

  • Ubuntu and Linaro freeze schedule
  • Ideas for LDS demos for the platform team
  • Any other business (AOB)

Regrets

  • wookey
  • ppearse
  • aviksil
  • kunal

Past Action items

  • slangasek to handle merge of dpkg-cross (bug #736154)

People Present

  • slangasek
  • hrw
  • tgall_foo
  • fturgis
  • jcrigby
  • JamieBennett

  • lool
  • aviksil

Actions Recorded

  • tgall_foo to check with GrueMaster for recommendations on checking minimal audio functionality on armel

  • fturgis to ask linaro-dev whether systemtap support for probing kernel modules is wanted
  • slangasek to check why armel-cross-toolchain-base FTBFS wasn't caught in lucas's archive rebuild
  • slangasek to review/merge/upload armel-cross-toolchain-base 1.62


Engineers Reports

Your Name <irc nick>

Highlights

  • Short bullet points of work you've done that week which convey progress and highlights which can be used to report on how the team is progressing as a whole.

Plans

  • Your individual plans for the coming week(s).

Issues

  • Your individual plans for the coming week(s).

Wookey

Highlights

  • Collected Panda hardware: soldered up PSU, acquired USB dongle
  • Got Panda booting with snapshot+lmc
  • Helped Steve find out why xdeb never works for him - attempted a fix
  • Started grokking pkgconfig

Plans

  • Merge xdeb trunk into Debian
  • FFe for multistrap
  • Update dpkg :native patch
  • Try dev tools on panda and write up

Issues

  • 3 days of ARM Conf/induction will eat most of the week.

John Rigby <jcrigby>

Highlights

  • Got more feedback on u-boot patches for U8500. Not too bad. Submitted fixes addressing feedback and got some more feedback. Looks like after next submission the patches will be accepted.
  • Tested latest u-boot-stable on OMAP[34]. Still need to test on mx5x then will do a merge request. The FFE for this is device tree support on several platforms.
  • Wrestled with broken kernels. Next kernel package would be read except for the broken USB on omap3 bug. One day of struggle was caused by my breaking OMAP3 by applying a workaround for OMAP4.

Plans

  • Watch mailing list for a fix for the OMAP3 usb bug and release a kernel immediately there after.

Avik Sil <aviksil>

Highlights

Plans

  • Continue working on packaging lttng-modules

Marcin Juszkiewicz <hrw>

Highlights

  • Got armel cross toolchain to buildable state under Debian (gcc 4.6, linux 2.6.38).
  • Got armel cross toolchain to buildable state under Ubuntu again (linux-libc-dev multiarch issue).
  • Proposed 463+512MB memory scheme for Pandaboard - merged.

Plans

  • Mangle -cross packages in armel-cross-toolchain-base to contain copyright and changelog files.

Frederic Turgis <fturgis>

Highlights

  • Resumed some debug activites: kernel module, storing 64 bit values. Now understood, requires some dev + testing
  • Identified the need to focus on timestamping methods (as TI method based on 32kHz timer is not reusable as is) -> test tapset timing function

  • Leaving Linaro for next cycle (Rob Clark is allocated from my group so management taking me back) but going to Linaro UDS.
  • Not much feedback on uprobes/utrace

Plans

  • Decide if "kernel module" requires new package (or can wait for integration of earlier systemtap release)
  • timestamping
  • retry uprobes/utrace on its mailing list

Tom Gall <tgall_foo>

Highlights

  • 746491 / 733334 - The audio bugs. Pointed at new PA code outlined in bug 746023. Didn't work. alip on xm or beagle C4 with known good asound.state file. Didn't work ubuntu-desktop on xm or beagle C4 with known good asound.state file. Didn't work alsa mixer/xfce4-mixer both seem correctly setup, speaker-test, aplay, xmms2 the ubuntu-media player ... all yield nothing.

    • Q: What is the multimedia team using?
    • Q: is there any sort of raw interface to at least validate audio is working with as few software layers as possible?
  • Not reported yet but I am occasionally seeing something odd with ALIP, like double screen savers, double desktops.... wondering if the xubuntu-default setting update might have a problem yet.
  • Graphical issue with ubuntu-desktop with refresh/painting. Affects all boards. 753878

  • First set of nano diet changes out for review. This is the minimal set, doesn't need live-build changes nor does it need busybox. Need to discussion with slangasek. The "rest" of the changes as well as the kernel module changes maybe should be immediately after 11.05 is shipped?

Plans

  • Focus on the cross gdb ddeb tool for the rest of this week.

Steve Langasek <slangasek>

Highlights

  • thanks to the Ubuntu release team's lenience, landed a change to the linux-libc-dev package to make it coinstallable under multiarch
    • this lets us make further progress on multiarch -dev packages this cycle, since linux-libc-dev is at the root of the -dev dep tree
    • coordinated uploads of libc packages in Ubuntu (klibc, eglibc, newlib) to adapt to this change
    • the change broke biarch package builds, so fixed up a symlink in gcc-multilib for compatibility
    • libc6-dev won't be coinstallable in natty - it has some compat links in place because upstream sources like emacs23 needs to find crt files under /usr/lib - will prepare a ppa upload to disable this so we have something readily available to base further multiarch -dev work on
  • also helping to push multiarch forward in Debian; making some progress vs. http://wiki.debian.org/Multiarch/Bootstrapping

Plans

  • last merges into Ubuntu archive for Ubuntu 11.04 release
    • upload last round of armel cross-toolchain packages
  • review cross-build merge requests from Peter - many of which will now need to be shelved until oneiric opens due to Ubuntu freeze, but reviewing can still happen

Issues

  • flashplugin+i386 nspluginwrapper segfaults when running under arm, interfering with my plan for a demo at LDS


IRC logs

Started logging meeting in #linaro-meeting
[15:05:04] <slangasek> [LINK] https://wiki.linaro.org/Platform/DevPlatform/Meetings/2011-04-13
[15:05:34] <slangasek> [TOPIC] Actions from last meeting
[15:05:44] <slangasek> * slangasek to handle merge of dpkg-cross (bug #736154)
[15:05:46] <slangasek> done
[15:05:52] <slangasek> that was short :)
[15:06:08] <hrw> ;D
[15:06:56] <slangasek> if there are other merges into Ubuntu that are holding things up, please shout as we're very close to the Ubuntu release now and should get any remaining changes in sooner rather than later
[15:07:17] <slangasek> so even if you've asked me for the merge already, a friendly reminder would not be amiss
[15:07:29] <slangasek> [TOPIC] Engineer progress reports
[15:07:35] <slangasek> $ echo $(shuf -e hrw jcrigby dmart wookey slangasek ppearse aviksil tgall_foo kunal fturgis)
[15:07:38] <slangasek> tgall_foo dmart fturgis slangasek hrw wookey jcrigby ppearse aviksil kunal
[15:07:43] <slangasek> tgall_foo: what's the news?
[15:08:23] <tgall_foo> firstly there are a couple of items out for review involving the nano build, they are tied to that "nano" needs to go on a diet bug I had opened
[15:08:44] <slangasek> yes, on my radar
[15:08:46] <tgall_foo> This is the minimal set, doesn't need live-build changes nor does it need busybox. Need to discussion with slangasek. The "rest" of the changes as well as the kernel module changes maybe should be immediately after 11.05 is shipped?
[15:09:07] <tgall_foo> the next set would include busybox ... something we can talk about offline
[15:09:26] <slangasek> kernel modules - yes, I think we should have a session at LDS to work out exactly what we want there and get an agreed plan on how to proceed
[15:09:28] <tgall_foo> bugs where a big focus again this week
[15:09:32] <tgall_foo> 746491/733334 - The audio bugs. Pointed at new PA code outlined in bug 746023. Didn't work.
[15:09:33] <tgall_foo> alip on xm or beagle C4 with known good asound.state file. Didn't work
[15:09:33] <tgall_foo> ubuntu-desktop on xm or beagle C4 with known good asound.state file. Didn't work
[15:09:33] <tgall_foo> alsa mixer/xfce4-mixer both seem correctly setup, speaker-test, aplay, xmms2
[15:09:33] <tgall_foo> the ubuntu-media player ... all yield nothing.
[15:09:33] <tgall_foo> Q: What is the multimedia team using?
[15:09:35] <tgall_foo> Q: Is there a minimal audio interface to validate sound works with as few audio layers involved as possible?
[15:09:57] <tgall_foo> Not reported yet but I am occasionally seeing something odd with ALIP, like double screen
[15:09:57] <tgall_foo> savers, double desktops.... wondering if the xubuntu-default setting update might have
[15:09:57] <tgall_foo> a problem yet.
[15:09:57] <tgall_foo> Graphical issue with ubuntu-desktop with refresh/painting. Affects all boards. bug 753878
[15:10:06] <tgall_foo> and that's all I have to report
[15:10:39] <slangasek> tgall_foo: I suggest you check with GrueMaster on IRC regarding minimal audio interfaces; he's the Ubuntu ARM QA guy and has alsa upstream background
[15:10:46] <slangasek> (and he's UTC-7 so should be easy to catch)
[15:10:50] <tgall_foo> ah excellent
[15:11:10] <slangasek> can you also follow up with the multimedia team to figure out what they are using (... if anything)?
[15:11:25] <tgall_foo> yup that's my plan ... just throwing those two questions out as ones that I want to get to the bottom of
[15:11:47] <slangasek> [ACTION] tgall_foo to check with GrueMaster for recommendations on checking minimal audio functionality on armel
[15:11:53] <slangasek> got it :)
[15:11:53] <hrw> tgall_foo: looks like ALIP image needs tests - can make it tomorrow (if will find my second usb mouse)
[15:12:07] <slangasek> tgall_foo: anything else?
[15:12:12] <tgall_foo> slangasek, nope
[15:12:32] <hrw> tgall_foo: but on panda only - beagle C3 has -ENOMEM problem
[15:13:00] <slangasek> tgall_foo: ok, thanks
[15:13:01] <slangasek> fturgis:
[15:13:31] <fturgis> Resumed some debug activites: kernel module, storing 64 bit values in memory... Now understood, requires some dev + testing
[15:13:48] <tgall_foo> hrw, sounds good
[15:13:48] <fturgis> thanks to nicolas pitre
[15:14:10] <fturgis> need to focus on testing timestamping methods -> writing custom tests to test tapset timing function
[15:14:24] <fturgis> Not much feedback on uprobes/utrace
[15:14:42] <fturgis> we also didn't decide about kernel module thing
[15:14:58] <fturgis> :-( Leaving Linaro for next cycle. Rob Clark is allocated from my group so management is taking me back. But I am going to Linaro UDS as we want to keep contact on tooling
[15:15:29] <fturgis> that's it
[15:15:52] <slangasek> it's nice to know you'll still be in touch :)
[15:16:10] <fturgis> so I can keep my panda board
[15:16:16] <slangasek> what else could we do to get you the feedback you're looking for on utrace? Is the problem that no one is using this?
[15:16:40] <fturgis> I think so, it seems to be too much Redhat related, and not upstreamable for Linus
[15:17:33] <fturgis> I will try to get some feedback again then I think userspace tracing will have to be done differently
[15:17:41] <slangasek> ok
[15:17:55] <slangasek> can you remind me what the kernel module issue is that needs decided?
[15:18:36] <fturgis> we can't probe kernel module, solved in systemtap v1.4. Shall we backport in current package or is it not so important ? I never probe modules on my side
[15:20:18] <slangasek> ah. if you don't use it and no one's demanding it, I guess we don't need to do a possibly-risky backport
[15:20:43] <fturgis> I can ask linaro-dev who may want to probe kernel modules ?
[15:20:50] <slangasek> good idea
[15:21:03] <fturgis> let's do it this way
[15:21:06] <slangasek> [ACTION] fturgis to ask linaro-dev whether systemtap support for probing kernel modules is wanted
[15:21:14] <slangasek> thanks
[15:21:37] <slangasek> and I'm next
[15:22:21] <slangasek> it's been annual review season in Canonical, so some of my time has been taken away from Linaro for that
[15:22:40] <slangasek> otherwise, have been helping to shake out remaining bugs from multiarch in general for the Ubuntu archive
[15:23:25] <slangasek> we landed a change to the linux-libc-dev package thanks to an exceptionally lenient freeze exception on the part of the Ubuntu release team; that went ok except for the bit where it broke biarch package builds, so had to fix up a symlink in gcc-multilib
[15:23:54] <hrw> slangasek: and eglibc, klibc, newlib, armel-cross-toolchain-base got broke too
[15:24:02] <hrw> all fixed
[15:24:12] <slangasek> but the upshot is that you can have linux-libc-dev coinstalled for more than one arch on your system now, which is good because that's at the root of our -dev dep tree
[15:25:14] <slangasek> libc6-dev won't be coinstallable in natty - it has some compat links in place because upstream source like emacs23 needs to find crt files under /usr/lib - but I'll be preparing a ppa upload to disable this so we have something readily available to base further multiarch -dev work on
[15:25:40] <slangasek> hrw: yes, the first three were at least coordinated; the last one I still don't know why it didn't get caught in the rebuild check and need to look into it further
[15:26:00] <slangasek> [ACTION] slangasek to check why armel-cross-toolchain-base FTBFS wasn't caught in lucas's archive rebuild
[15:26:21] <hrw> slangasek: he prbably rebuilt main only
[15:26:33] <slangasek> hrw: I know he rebuilt more than main
[15:26:49] <hrw> ok
[15:26:51] <slangasek> anyway, I'll follow up
[15:27:12] <slangasek> we don't want to leave other packages broken in Ubuntu from this (though at this point it's extremely unlikely there are any other issues)
[15:27:47] <slangasek> also helping to push multiarch forward in Debian; making some progress vs. http://wiki.debian.org/Multiarch/Bootstrapping
[15:28:12] <slangasek> I think that's it for me
[15:28:13] <slangasek> questions?
[15:29:14] <slangasek> hrw: your turn then :)
[15:29:23] <hrw> ok
[15:29:43] <hrw> updated armel-cross-toolchain-base to build under Debian again (gcc 4.6, linux 2.6.38)
[15:29:54] <hrw> - one gcc-4.6 patch sent for merge
[15:30:15] <hrw> - fixed linux-libc-dev multiarch issue under ubuntu
[15:30:37] <hrw> - adding copyright, changelog into -cross packages (dpkg-cross drops them)
[15:30:45] <hrw> - fixing md5sums of -cross packages
[15:31:03] <hrw> fixed few ubuntu packages
[15:31:28] <hrw> again wasted time trying unity - next time in 6 months or more
[15:31:32] <hrw> thats all
[15:32:17] <slangasek> nice work on the armel package build failure fixing
[15:32:23] <slangasek> are you in the porting jam again today?
[15:32:44] <hrw> no - today spent time getting actb working
[15:32:51] <hrw> want to get it into natty
[15:33:03] <slangasek> yep!
[15:33:13] <slangasek> happy to sponsor that when you have it ready
[15:33:15] <hrw> checked porting jam list yesterday and did not found something interesting
[15:33:48] <slangasek> ah, are we at the bottom of the barrel then :)
[15:33:56] <slangasek> I found a binutils segfault yesterday, if that's interesting :)
[15:34:04] <hrw> report bug?
[15:34:18] <slangasek> (qemu-linaro FTBFS because ld segfaults when building an alpha emulator, heehee)
[15:34:28] <slangasek> bug is reported but I didn't put it in the porting queue
[15:34:39] <slangasek> anyway, off-topic :)
[15:34:40] <slangasek> jcrigby:
[15:35:08] <jcrigby> u8500 u-boot appears to be headed upstream
[15:35:27] <slangasek> \o/
[15:35:45] <jcrigby> went through 4 revs of patches and last feedback was pretty minimal
[15:35:57] <hrw> nice
[15:36:05] <jcrigby> spent some time wrestling with kernel
[15:36:30] <jcrigby> one day last week trying to find workaround for upstream problems but no luck
[15:36:38] <jcrigby> that was the omap3 usb
[15:36:48] <jcrigby> looks like we just have to wait for upstream
[15:37:25] <slangasek> where do we stand with the beagle video underrun?
[15:37:33] <jcrigby> yesterday was spent figuring out a self inflicted problem, a workaround for omap4 broke omap3 display
[15:37:35] <slangasek> (and with the next kernel upload?)
[15:38:22] <jcrigby> next kernel would be ready but for the beagle video underrun, and the omap3 usb problem
[15:39:05] <jcrigby> the video underrun can be fixed by not entering ondemand frequency scaling
[15:39:22] <jcrigby> I supposed I could hack the min freq in a sauce patch
[15:39:35] <jcrigby> suppose
[15:39:47] <slangasek> the omap3 usb is a regression vs. the current package?
[15:39:58] <jcrigby> yes
[15:40:00] <JamieBennett> jcrigby: whats the upstream status on the pm problem with video underrun?
[15:40:06] <lool> Will there be an u-boot-linaro upload with 2011.03?
[15:40:16] <jcrigby> lool, thanks
[15:40:35] <jcrigby> yes it is uploaded, I need to do a merge request
[15:40:43] <lool> Cool
[15:40:53] <jcrigby> the changelog mentions fdt support as reason for FFE
[15:40:55] <lool> Probably in the queue then I guess
[15:41:15] <jcrigby> I have not actually done the merge req, I wanted to do more testing
[15:41:20] <jcrigby> that testing is now done
[15:41:25] <slangasek> I don't see u-boot-linaro in the unapproved queue
[15:41:48] <slangasek> so I guess "uploaded" didn't mean "to the archive"
[15:41:53] <jcrigby> sorry,
[15:41:57] <lool> pushed? :-0
[15:42:01] <jcrigby> yes it is pushed
[15:42:15] <hrw> slangasek: ok, looks like I will be able to request review today
[15:42:29] <aviksil> just logged in from New Delhi!
[15:42:37] <slangasek> hrw: great :)
[15:42:42] <slangasek> aviksil: hi there! Have your visa? :)
[15:42:43] <jcrigby> that is it for me
[15:42:58] <aviksil> slangasek: the interview is tomorrow
[15:43:01] <slangasek> jcrigby: ok, thanks
[15:43:14] <jcrigby> JamieBennett, sorry I have not seen any comment on C4 underrun
[15:43:16] <slangasek> aviksil: ok. care to give a status update, as long as you're here?
[15:43:29] <aviksil> slangasek: sure
[15:43:30] <JamieBennett> jcrigby: OK, so we will need to hack it for now it seems
[15:43:53] <jcrigby> JamieBennett, I will try a kernel hack today and let you know if that works
[15:44:03] <slangasek> jcrigby: please follow up to the underrun bug report to clarify the current status, the PM team seemed to be under the impression this was now fixed
[15:44:14] <aviksil> proposed a patch for oprofile to run on omap4
[15:44:16] <jcrigby> will do
[15:44:44] <aviksil> slangasek: should I sent the patch to other mailing lists?
[15:44:45] <slangasek> aviksil: any feedback on that patch?
[15:44:52] <aviksil> nothing so far
[15:45:29] <slangasek> do you have a link to the exact mail? I'll ask npitre for a review
[15:45:39] <aviksil> other than that, working on packaging lttng-modules
[15:46:06] <aviksil> slangasek: https://bugs.launchpad.net/ubuntu/+source/oprofile/+bug/702999
[15:46:16] <slangasek> ok
[15:46:20] <aviksil> slangasek: comment #11
[15:46:22] <slangasek> thanks
[15:46:39] <aviksil> slangasek: sent you a mail regarding the packaging
[15:47:04] <aviksil> slangasek: will discuss with you regarding this offline
[15:47:08] <slangasek> yes, sounds good
[15:47:18] <aviksil> that's all from me
[15:47:21] <slangasek> aviksil: oh, if this patch hasn't been sent to *any* mailing list yet, please send it to linaro-dev for review
[15:47:53] <aviksil> slangasek: ok, i'll do so
[15:48:12] <slangasek> please fix it up to apply cleanly to the linaro kernel tree first, if any changes are needed there
[15:48:24] <aviksil> slangasek: sure
[15:48:35] <slangasek> thanks
[15:48:46] <aviksil> slangasek: it's based on linaro-2.6.38 only
[15:49:22] <slangasek> that's what we want for now, sure
[15:50:47] <slangasek> [TOPIC] work item progress
[15:50:54] <slangasek> [LINK] http://status.linaro.org/linaro-foundations.html
[15:51:08] <slangasek> we seem to be slipping a bit regarding the trendline
[15:51:20] <hrw> slangasek: should I add actb to bug 750585?
[15:51:42] <slangasek> some of this is because this report includes any blueprints that we have workitems on even if the blueprints themselves belong to another team...
[15:51:53] <hrw> WIs...
[15:52:01] <slangasek> but if you have work items that you think aren't getting done this cycle, talk to me and we'll postpone them
[15:52:08] <slangasek> hrw: you can add it to the bug, sure
[15:52:19] * tgall_foo looks through the list
[15:52:49] <JamieBennett> tgall_foo: don't forget you get a personal burndown page with all your WI's on too :)
[15:53:03] <slangasek> yes, everyone please look through your personal WI lists... not going to try to go through them all in this meeting
[15:53:09] <slangasek> [TOPIC] Ubuntu and Linaro freeze schedule
[15:53:41] <slangasek> JamieBennett has just sent mail to linaro-dev detailing the remaining timeline between now and the release
[15:53:45] <slangasek> [LINK] http://lists.linaro.org/pipermail/linaro-dev/2011-April/003993.html
[15:54:05] <slangasek> so I wanted to call attention to this, since it affects any work we're doing that involves packages that are included on images
[15:54:14] <slangasek> or that are uploaded to the Ubuntu archive generally
[15:54:34] <slangasek> any questions/concerns there?
[15:54:48] <JamieBennett> If you have any questions after the meeting, feel free to ping me :)
[15:54:58] <JamieBennett> (or slangasek)
[15:55:38] <slangasek> [TOPIC] Ideas for LDS demos for the platform team
[15:55:58] <slangasek> so we're going to have a demo night at LDS for Linaro projects
[15:56:06] <hrw> I am happy with state of my packages. all important changes went into archive and rest is bugfixes which waits for review
[15:56:06] <tgall_foo> well on the positive side, I suspect most/all demos running linaro will be running our images :-)
[15:56:27] <hrw> [ACTION] slangasek to review/merge/upload armel-cross-toolchain-base 1.62
[15:56:45] <slangasek> I thought I had a good one that I was going to put forward for Developer Platform
[15:57:03] <slangasek> the bad news is that adobe flashplugin + i386 nspluginwrapper segfaults when running under qemu on armel ;(
[15:57:17] <JamieBennett> cross-compile ? QEMU ?
[15:57:29] <hrw> slangasek: try amd64 one ;D
[15:57:49] <slangasek> (oh, I guess the good news is that I have all the packages needed to install flashplugin + i386 nspluginwrapper on an armel natty system using multiarch ;)
[15:57:57] <slangasek> hrw: heh, Ubuntu doesn't even *ship* the amd64 one
[15:58:01] <slangasek> [ACTION] slangasek to review/merge/upload armel-cross-toolchain-base 1.62
[15:58:11] <slangasek> JamieBennett: how do you turn that into a flashy demo?
[15:58:20] <slangasek> that's the question :)
[15:58:31] <JamieBennett> QEMU demo with a demo of how to test the images ;)
[15:58:32] <slangasek> we've got lots of cool output, how do we make it eye-catching :)
[15:58:39] * JamieBennett nods
[15:58:41] <JamieBennett> needs though
[15:58:42] <JamieBennett> t
[15:59:01] <hrw> slangasek: alip image with qemu showing another alip image?
[15:59:12] <slangasek> haha, sick :)
[15:59:23] <JamieBennett> alip image with qemu showing another alip image cross-compiling QEMU ;)
[15:59:44] <hrw> will we have MX53 quickstart boards at LDS?
[15:59:53] <slangasek> not that I know of
[16:00:23] * hrw wants MX6 dual/quad 
[16:00:48] <slangasek> anyway, please give that some thought and if you come up with something you'd like to run a demo of, talk to me
[16:00:52] <JamieBennett> hrw: there will be iMX53 at UDS
[16:01:00] <slangasek> [TOPIC] AOB
[16:01:05] <slangasek> anything else?
[16:01:44] <slangasek> #endmeeting
Meeting ended.

Platform/DevPlatform/Meetings/2011-04-13 (last modified 2011-04-18 07:41:19)