- Team Work - Summary
- Team Work - Detailed
Place: #linaro-meeting on irc.freenode.net
Time: Wednesday at 1300 UTC
[TOPIC] review actions from last meeting [TOPIC] review team work [TOPIC] Milestone Update (asac) [TOPIC] progress reports [PROGRESS REPORT] Graphics WG (Jesse) [PROGRESS REPORT] Multimedia WG (asac) [PROGRESS REPORT] Distribution (tgall_foo, kunal) [TOPIC] AOB
[TOPIC] Your Topic - Your Name
Actions from Last week
* fgu file bug against linaro-image-tools on proxy problems * asac and kunal to work on way to autobuild all linaro kernel flavours daily * Tech leads to check the teams specs and push completion
Completed Actions from Last week with Status
* All (especially gfx wg members) to check Jesse's slides * Done!
<Mootbot-UK> Meeting started at 14:03. The chair is asac.
[TOPIC] review actions from last meeting
[ACTION] fgu file bug against linaro-image-tools on proxy problems
<asac> i think that was done, right fgu ?
<fgu> yes https://bugs.launchpad.net/bugs/673570
[ACTION] asac and kunal to work on way to autobuild all linaro kernel flavours daily
<asac> -> we are working on that ... takes a bit longer
<asac> but no action required. the meego story seems to move again. we are working getting OBS setup on meego side and also have an effort started with linux-foundation about unblocking toolchain delivery into meego
[ACTION] Tech leads to check the teams specs and push completion
<asac> -> graphics WG is in good shape
<asac> -> mm WG is improving .... but the main aspects are flashed out ... work items and some refinements have to happen still
[TOPIC] review team work
<asac> so first, please add your work _before_ this meeting
<asac> the idea of this topic is that we can go through it and extract a executive summary for the team weekly report
<asac> with 50%+ of people not doing it before this doesnt make sense as we might the really interesting gems
- * tgall_foo suspects there will be massive lock contention in the wiki rightafter the meeting
<asac> well. do it by today EOD please
<asac> lets skip the team work review this time due to lack of content
<tgall_foo> where EOD == asac's EOD
[ACTION] asac to send around reminder that submitting that info is critical and an essential part of assignee/linaro work
[TOPIC] Milestone Update (asac)
<asac> so currently our tracking infrastructure is still worked on
<asac> so we cannot really look at it... also the blueprints are not finished, so the infrastructure wouldnt really help there
<tgall_foo> when do we expect to see burn down charts for natty ?
<asac> anyway, everyone can take a look at their specs... if you scheduled something against natty-alpha-1 and you already see that you wont be able to make it
<asac> tgall_foo: its work in progress ... JamieBennett is fighting with the added flexibility we did this cycle with lots of projects hosting our blueprints etc.
<jazh> natty-alpha-1 should be Dec 2?
<asac> those milestones are usually Thursday
<alf_> btw, all please make sure to read https://wiki.linaro.org/Platform/UserPlatforms/BlueprintSpecFaq and make sure the syntax you use is *exactly* as described there
<asac> if your work item involves delivery to the archive they should go up before Tuesdays though
<asac> good reminder ...
<asac> i alredy sent around info on that i think ... but i can only reemphasize: not using that syntax will be devastating to our tracker tools, so please review your current whiteboard and fix it accordingly
[ACTION] everyone to update blueprints to adhere WorkItem syntax
[TOPIC] progress reports
[PROGRESS REPORT] Graphics WG (alf_ in for Jesse)
<alf_> I guess I can speak for myself and then jammy can give us an update, too
<asac> sounds good
<alf_> so, for cairo the progress is good, mostly infrastructure work (setting up repos, studying the code)
<alf_> and there is also a gl enabled cairo package in a PPA (check report)
<alf_> good news is that after discussion with upstream we don't have to worry about GL 1.x so the work is going to be a bit easier
<asac> nice ... \o/
<alf_> for graphics benchmarks, I finally managed to get some benchmarks to run with spandex and I am working on that atm
<alf_> also, the glmark2-es2 program now works correctly (without artifacts)
<alf_> in general, it all looks good for natty-alpha-1
- * alf_ hands mic to jammy
<asac> alf_: one question ... did you have time to checkout qt traces?
<alf_> asac: no
<asac> e.g. how that works and moreso: what traces are already available in the git?
<jazh> ok, this week I fixed some bugs in fsl driver to make texture_from_pixmap_glesv2 (from alf) work on imx51
<jazh> and about the compositor spec, I have make meego handset UX run on imx51
<jazh> meegotouch-compositor is used on handset UX
<asac> jazh: how did you do that? e.g. what did you start from? did you use linaro rootfs?
<jazh> it was done based on N900 image
- james_w jazh
<tgall_foo> jazh, so you didn't create any linaro debs then ?
<jazh> to use linaro rootfs, we may need to package meego core/meego touch framework/... in linaro archive
<asac> would be interesting to see what the minimal of upstream sources we need to package to work on that
<asac> e.g. is it really the full touch framework? or rather just 3 packages?
<jazh> and I haven't get too much time to check chromium os yet for chromium-wm
<rajeev> what do u mean by meego core
<tgall_foo> asac, yeah I think that would be worthwhile .. specially since for the graphical images we want meego-touch to be part of it
<asac> right. so someone should at least review what is missing on the stack up to the compositor and make a list so we can inject that into our platform efforts
<jazh> maybe i need some help from tom_gall and alf for meego packaging in linaor
<asac> jazh: you dont need to do the packaging ... just give output what is needed
<asac> platform specialists can then do it assuming its not like a huge set of packages missing
<asac> but well we can take this directly ...
<jazh> about the gl proxy blurprint, I have written a draft design document
[ACTION] kunal and tgall_foo to assemble list and estimate effort to get meegotouch-compositor into linaro graphics WG images
<asac> (i think its not that many packages ... but lets see)
<asac> jazh: yeah... that was really detailed ... can you share the URL here for the proxy spec for the audience?
<jazh> and I have sent it out for review, I'd like to get some feedback first before move to implement
<tgall_foo> yeah I don't thnk it'll be too bad
<asac> jazh: sure ... please share URL here anyway (even if not finished)
<asac> ok thanks jazh ... anything else to update us on?
<asac> ok lets move on (hope i am not offline ;))
<jazh> asac, tgall_foo: my network had some problem just before, sorry for that
[PROGRESS REPORT] Multimedia WG (asac)
<asac> so ... plan review for mm WG was done yesterday
<asac> final slides are here:
<asac> there are some concerns about codec optimization and we need to reshuffle things
<asac> (sorry i got twenty requests from all sides)
<asac> i will try my best to accomodate the project reshuffling requests i got wrt codec optimization work ... atm vp6 and aac tasks are a bit understaffed there
<asac> we will reach out and get another NEON expert on board for that
<asac> so going through a few projects ... i think the pulseaudio story is moving now. i reviewed the code and i have a good understanding what could be done. mandeep is tasked to make this happen, but now i am pretty sure its doable
<asac> for the smp optimization work we are a bit blocked on a9 boards ... we are looking at getting panda boards there for engineers, but there is no estimated time yet
<kiko> asac, do you want a versatile express?
<tgall_foo> kiko, are there a9 based versatile expresses ?
<asac> however, i think we can continue on refactoring and focus on obvious work to be done and do the fine tuning when the hw situation is solved
<kiko> tgall_foo, "the" VE we have is A9-based.
<kiko> quad-A9 in fact
<asac> kiko: thanks kiko for the offer. i will talk to the engineers that struggle most to get hardware for their SMP work
<asac> quad a9? wow
<asac> thats nice ... so 1080p for vp8 there (just kidding)
<tgall_foo> at least!
<asac> kiko: is the smp support out of the box in our 10.11 release? do you know out of your head?
<asac> ok ... so moving on to other topics.
<kiko> asac, yes, it is -- talk to mattman for information
[ACTION] asac to follow up on vexpress a9 quad core lead and talk to mattman about status
<asac> so the validation story ... i am not super bullish about the current status of the instrumented player and metrics application blueprint; i gave some more details on what we really want to achieve there and would like to ask sudipj_ and vishalr to look at the slides and adjust the specs for it
[ACTION] asac to work with vishalr and sudipj_ to get the validation story right
<asac> on the other topic ... gstreamer-omax + zero copy i think we know what we are aiming for
<asac> we want to not go for xv ... nor v4l2 there, but rather go for the one and only way to do a gsteglsink
<asac> this will probably be exciting stuff, but unfortunately relies on us getting fully enabled mali images
<gsachin_> atleast we can have something inside STE for development
<asac> gsachin_: kan_hu and jian can help on that part; they are gstreamer masters
<asac> and i think with them and also with benjiG we should be able to get something out of it.
<asac> having this deliverable is far more exciting than many other things, so lets make this happen
<gsachin_> yes right
<asac> also on gstreamer-omax, i think consolidating TI and STE tree is something we cannot start right away
<benjiG> does that means you have fix the way to handle correctly hw buffers ?
<asac> as a) we are waiting on something we can work against first - aka 0.11 branch is not even out yet
<asac> benjiG: i think thats part of the zero copy thing, yes
<asac> benjiG: we have to actively engage with the kernel WG there
<asac> and drive a solution forward that is acceptable upstream
<asac> so i would like benjiG and someone else interested to attend the kernel WG calls to track progress on this. also talk to paulmck and lool about how we can get something rather sooner than later
<asac> ok ... i think thats mainly it for the multimedia update ...
[PROGRESS REPORT] Distribution (tgall_foo, kunal)
<tgall_foo> ok ... asac I think is a little more up to date on where there are with the natty seeds ...
<asac> right. status: natty seeds are set up; config branches pushed ... now i myself have to enable the images for it
<tgall_foo> long story short getting the daily builds going for natty is an area of focus in advance of alpha1 next week
<tgall_foo> asac, do you know if anyone has successfully been able to build a natty image with live helper ?
<tgall_foo> lastI checked that was busted
<tgall_foo> (last week)
<asac> tgall_foo: no... but thats not a problem. we have to get the daily images in place
<asac> so we can work on fixing things as we go
<asac> its normal that before alpha-1, image builds often cause problems. headless might work early
<tgall_foo> the linaro-multimedia-engineering and linaro-graphical-engineering seeds are assembled at least in the form of an early 1st draft
<asac> but full UI is often a problem even before alpha-2
<asac> yay. seeds are making good progress.
<tgall_foo> asac, well it's a problem in so much that without a working live-helper it's hard to have any confidence in what might be committed for the dailies ... least that's my concern
<asac> tgall_foo: why are we using the natty live-helkper
<tgall_foo> kunal, anything you'd like to cover?
<asac> i would have assumed we use the maverick live-helper
<tgall_foo> asac, I am
<asac> i mean: use maverick live-helper to build natty images
<tgall_foo> but one ought to be able to build a natty image with the natty live helper and natty packages :-)_
<asac> well ... kunal has a first version of bellagio packaged in his ppa
<asac> well done
<tgall_foo> awesome kunal !
<kunal> yes that's done 1st draft
<asac> next step is to use the patch system to integrate the changes done as part of the bellagio spec by vishalr
<kunal> and plus the integrate review comments
<asac> there are a few more refinements we discussed on the packaging, but the 1st one is always the hardest :-P
<asac> so right. once you incorporated those changes, and added the patch, let me know and i wil lcopy it to a place where we get an armel build from
<asac> then we can integrate it in our images and have a good base to work against to make hardware codecs for STE and others that want to use bellagio happen
<asac> ok thanks for the update kunal and tgall_foo
<asac> platform is often a bit confusing in beginning of cycle, but usually things become much busier later in cycle
<asac> using this time to look what it takes to get the meegotouch compositor in is good imo
<tgall_foo> continue to work through some of the ogre3d howto's ... have spent a little time understanding what's missing from the gles2 port
<Mootbot-UK> Meeting finished at 15:00.
* asac to send around reminder that submitting that info is critical and an essential part of assignee/linaro work * everyone to update blueprints to adhere WorkItem syntax * kunal and tgall_foo to assemble list and estimate effort to get meegotouch-compositor into linaro graphics WG images * asac to follow up on vexpress a9 quad core lead and talk to mattman about status * asac to work with vishalr and sudipj_ to get the validation story right
Team Work - Summary
Team summary goes here : https://wiki.linaro.org/Platform/UserPlatforms/WeeklyReport
Team Work - Detailed
Alexandros Frantzis (alf)
- Good progress on discussions with upstream.
- Provided cairo gl enabled packages in ppa:afrantzis/cairo-gl
- Set up a cairo branch at git.linaro.org.
- Graphics Benchmarks
- Investigated and fixed visual artifacts in glmark2-es2.
- Started investigating spandex.
- Ported our clutter-1.0 ubuntu package patches to clutter 1.4.
Benjamin Gaignard (benjiG)
- get latest hwmem patches: TODO review them and use documentation to produce spec
Frederick Gueydon (fgu)
Jammy Zhou (jazh)
- Get feedback from Unity team that there is no plan to add gles2 support to compiz/nux
- Make texture_from_pixmap_glesv2 work with fsl driver after fix some bugs in the driver
- Make Meego Handset UX work on imx51
- Finished draft document for detailed design of GL proxy, and send it out for review to get some feedbacks
Jesse Barker (JesseBarker)
Kan Hu (Kan)
- Studying FFmpeg aac encoder
- Draft work items for Optimization AAC encoder
Kunal Goel (kunal)
- Debian packaging done for Omxil Bellagio
Li Jian (jian)
Mandeep Kumar (mandk)
Rony Nandy (ronynandy)
- [SUMMARY] Specs updation,Interaction with tools team,Debugging ffmpeg for zero fps for some vp8 vectors
- Specs updation
- Interaction with tools team regarding OpenCL and OpenMP support.OpenCL support is there.Other interaction
- regarding compiler vectoriser related benchmarking.
- Debugging ffmpeg for zero fps for some vp8 streams.
- Monday-sick leave
Sachin Gupta (Sachin)
- Investigated empathy videochat application for camerabin usage
- Inestigated mali code drop from arm
- understood TI's solution for zerocopy between gstomx and xvideo
- requested further information from Rob regarding buffer sharing and detection between v4l2 sink and gstomx
Sudip Jain (sudipj)
- Investigated perf tool to extract system-metrics
- Investigated gst player data
Tom Gall (tgall_foo)
- Summary : good progress in advance of the alpha1 for linaro-development, linaro-multimedia-engineering and linaro-graphical-engineering seeds
- linaro-development seed proposal merged
- Initial linaro-multimedia-engineering seed assembled, built and booted (plenty yet to do)
- Initial linaro-graphical-engineering seed assembled, built and booted (plenty yet to do)
- working through the ogre3d tutorials
- debugged where in ogre3d simple GLES2 tutorials come to a crashing halt and established a working OpenGL base to compare to.
Vishal Raj (vraj)
Platform/UserPlatforms/WeeklyStatus/2010-11-24 (last modified 2010-11-29 19:14:05)