Include: Nothing found for "^## meetings$"!
Include: Nothing found for "^## end-meetings$"!
Our primary goals are to:
Develop and release tested monthly builds of Android for Galaxy Nexus, Panda, Arndale, and Versatile Express.
Monthly Release Candidate: last Monday of Month, 16:00 UTC. See also Linaro releases and the monthly cycles.
Upstream development efforts.
Perform monthly toolchain benchmarking
- IRC: #linaro-android on irc.freenode.net
For android-build.linaro.org issues email email@example.com and/or file a bug at https://bugs.launchpad.net/linaro-android-infrastructure/+filebug
Android Kernel Meeting
- Add agenda items
Past Action items
- To be added.
- Action items that came out of this weeks meeting.
Android Team Meeting
- Actions from last meeting
Team issues discussion - Alexander and JamieBennett asking questions
- LEB/Image review - jserv
- Validation update - Jeremy
- Bug status - Patrik
Past Action items
- asac to invite TI to Android weekly meeting
- asac to chase TI on the mm bits
- patrikryd to try to push generic hook patch to upstream and see what happens
- jserv-- to send email to linaro-dev on toolchain work (previously sent to android-platform)
asac and JamieBennett to talk to james_w about keeping android blueprints on android team radar
- asac to think about validation meta blueprint and what to do with that with jamie and jeremychang
- patrikryd to sort out href8500 build situation
jserv-- to email JamieBennett and asac about what is needed from TI
- jeremychang to email plars (Paul Larson) to review Android changes to LAVA
- asac to invite TI to Android weekly meeting
jserv-- to bring https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-leb-gingerbread-pandaboard-gfx to a nice and round end
patrikryd to file a bug about killing u8500 build and put rational there; then kill build and get over it
- jeremychang to add work item to one of his blueprint to try mounting system/userdata/cache combined to / ... if that fails, we merge validation boot parititons
- asac and jeremychang to sync on validatoin partition layout and lamc
- patrikryd and jeremychang to file a bug for kernel boot with android toolchain problem
- patrikryd to nail down kernel build bug ... in worst case allow using the ubuntu cross toolchain for building kernel
Your Name <irc nick>
- 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.
- Your individual plans for the coming week(s).
- Your individual plans for the coming week(s).
Patrik Ryd <patrikryd>
- Almost done with the boot tarball blueprint
- kernel is now uImage
- ramdisk is now uInitrd
- u-boot included in tarball
- MLO included for beagle and panda
- boot panda
- push a patch to AOSP
- What do I need to prepare/do before LDS?
- public key error when pushing to AOSP
Jim Huang <jserv-->
- LEB: GFX + Multimedia accelerations for Pandaboard
- set up repositories for review: omap4-frameworks-base.git
- Pandaboard Dependency:
libaudio (from omapzoom): https://bugs.launchpad.net/linaro-android/+bug/757497
- liboverlay : hardware video overlay HAL
- libstagefrighthw : HAL for OMX renderer, loaded ok
- Linaro Toolchain for Android / NDK
Prepare repacked NDK: https://wiki.linaro.org/Platform/Android/Toolchain
- AOSP contributions:
libagl: Enable CLZ instruction generation for TextureObjectManager
- Hide private symbols using GCC visibility pragma
Hide internal symbol bionic_brk used by brk() and sbrk()
- Hide libm private symbols using GCC visibility pragma
- Add ARM optimized strcmp()
- LEB: GFX + Multimedia accelerations for Pandaboard
- Revise manifests for extra TI specific repositories
- No visible Android multimedia logs for reporting!
- No software available, so it is hard to perform multimedia tests.
Jeremy Chang <jeremychang>
- Move the common parts into device/linaro/common
- linaro-android-image-create ready to be reviewed and merged.
- basic validation and result submit on lava is done
https://wiki.linaro.org/JeremyChang/Sandbox/LavaAndroidValidation updated accordingly due to partiton layout changed
- Do more on lava.
- monkey integration on lava.
- 0xbench on lava.
- cts on lava.
- 0xbench application run needs be be monitored by an agent for lava
Android Kernel Meeting
To be added.
Android Team Meeting
Started logging meeting in #linaro-meeting [09:04:14] <asac> https://wiki.linaro.org/Platform/Android/Meetings/2011-04-20 [09:04:14] * jserv-- here! [09:04:19] <asac> [LINK] https://wiki.linaro.org/Platform/Android/Meetings/2011-04-20 [09:04:30] <asac> welcome back jserv-- [09:04:45] <asac> [TOPIC] Actions from last meeting [09:05:05] <asac> * asac to invite TI to Android weekly meeting -> carried over [09:05:10] <asac> [ACTION] asac to invite TI to Android weekly meeting [09:05:18] <asac> (sorry didnt review the actions after coming back from travel) [09:05:42] <asac> * asac to chase TI on the mm bits -> i think this was about licensing; this is moving afaict [09:05:55] <asac> * jserv-- to create a the panda-gfx-integration-initial blueprint -> thats done, right? [09:06:06] <jserv--> asac, yes [09:06:14] <jserv--> asac, moved to multimedia as my focus now. [09:06:32] <asac> hmm [09:06:38] <asac> jserv--: where is that blueprint? [09:06:46] <asac> or did you make one blueprint for GFX + mm? [09:07:18] <jserv--> https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-board-panda-multimedia [09:07:29] <asac> right [09:07:34] <asac> but what about the initial gfx integration? [09:07:41] <jserv--> https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-board-panda-gfx-investigation [09:07:42] <asac> i know there are work items somewhere, but i cant find the blueprint now [09:07:43] <asac> hmm [09:07:57] <asac> yes, but taht doesnt cover the work for integrating GFX [09:08:13] <jserv--> didn't it? [09:08:20] <asac> no ... i only was about investigation [09:08:30] <jserv--> ok, the name is misleading. [09:08:33] <asac> the stuff about creating linaro-gingerbread branch and putting GFX there and so on [09:08:37] <asac> jserv--: look at the work items [09:08:43] <asac> Try gingerbread build for android and see if GFX still works there: DONE [09:08:45] <asac> thats the last item [09:08:51] <asac> there is nothing about MAKING IT HAPPEN [09:09:02] <asac> i know that something was done ... we should replay it [09:09:12] <jserv--> ok [09:09:33] <asac> patrikryd: the work on the manifest and getting the build going etc. where is that covered? [09:09:37] <jserv--> however, I can not rename blueprint. [09:09:47] <asac> jserv--: you shouldnt. that blueprint is done and is fine [09:10:01] <patrikryd> asac: for LEB? [09:10:02] <asac> we should have a new one with the work items you did PLUS the work items to ensure that there is a build coming out [09:10:05] <asac> patrikryd: yes [09:10:15] <asac> https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-evaluation-build-initial [09:10:17] <jserv--> asac, ok, I though there was no need to open another blueprint for GFX. [09:10:25] <asac> ah i see [09:10:33] <asac> that one kind of captured the hw-pack branch [09:10:50] <patrikryd> asac: Nowere. I only sent out the patch this morning to get comments. [09:10:52] <asac> jserv--: well. this was work and i think the fact that we didnt do a blueprint also caused confusion what branch name etc. [09:11:06] <asac> yeah [09:11:07] <asac> so [09:11:21] <asac> problem is that 11:10 < asac> https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-evaluation-build-initial [09:11:25] <asac> is a meta blueprint [09:11:34] <asac> having work items in there that are related to concrete subtasks isnt that good [09:11:38] <patrikryd> asac: Should have a blue print. 1) create manifest, 2) setup builds , ... [09:11:39] <asac> becaues we cannot set this to implemented [09:11:42] <asac> right [09:11:46] <asac> one sec [09:12:28] <asac> linaro-android-leb-gingerbread [09:13:25] <patrikryd> asac: is that on the wiki or launchpad? [09:14:32] <patrikryd> https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-leb-gingerbread [09:14:54] <asac> https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-leb-gingerbread [09:14:58] <asac> hehe [09:15:02] <asac> i just created it [09:15:06] <asac> [LINK] https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-leb-gingerbread [09:15:12] * jserv-- subscribes. [09:15:18] <patrikryd> ok [09:16:26] <asac> https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-leb-gingerbread-pandaboard-gfx [09:16:29] <asac> jserv--: https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-leb-gingerbread-pandaboard-gfx [09:16:32] <asac> that one is yours ... [09:16:40] <asac> just add the work items you did a-posteriori there and set to done [09:16:55] <asac> let me move the work items from the initial meta blueprint there [09:16:56] <asac> one sec [09:19:03] <asac> jserv--: Create OMAP4 enablement branch: TODO [09:19:05] <asac> what does that mean? [09:19:23] <asac> ok i approved it and added the work items from the initial bp ... and will remove it from there [09:19:40] <jserv--> asac, because we have to modify Android frameworks/base in order to make OMAP4 OMX work [09:19:50] <jserv--> asac, but it is TI specific. [09:20:08] <jserv--> check details: https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-board-panda-multimedia [09:20:35] <jserv--> MM is much complex. [09:20:52] <asac> yeah [09:20:59] <asac> so ... https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-evaluation-build-initial [09:21:11] <asac> is now a pure meta ... i added the gingerbread gfx and the leb-gingerbread as dependency [09:21:20] <jserv--> ok [09:21:24] <asac> jserv--: can i rename the multimedia to leb-gingerebread-pandaboard-mm ? [09:21:35] <asac> to be consistent? [09:21:51] <jserv--> asac, agree! [09:22:03] <asac> renamed to: [09:22:07] <asac> [LINK] https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-leb-gingerbread-pandaboard-multimedia [09:22:23] <asac> ok good ;) [09:22:30] <asac> jserv--: mm is for 11.05 i guess? [09:22:41] <asac> ok seems the milestone is already like that [09:22:54] <jserv--> yes [09:23:10] <asac> jserv--: can you finish the gfx work items so we can set that to implemented? i think its just ensuring that the branch is named as expected and that the stuff is in our manifest included [09:23:20] <asac> ok let me look at multimedia ;) [09:23:28] <jserv--> asac, yes, my action item. [09:23:33] <asac> jserv--: INPROCESS -> INPROGRESS ;) [09:23:47] <asac> [ACTION] jserv-- to bring 11:16 < asac> jserv--: https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-leb-gingerbread-pandaboard-gfx [09:23:50] <asac> oops [09:23:54] <asac> [ACTION] jserv-- to bring 11:16 < asac> jserv--: https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-leb-gingerbread-pandaboard-gfx to a nice and round end [09:23:57] <asac> damn [09:23:59] <asac> i hate me [09:24:02] <asac> anyway. ... guess thats clear [09:24:06] <patrikryd> jserv--: Can you put the code on a LEB-panda branch instead of hw-packs? [09:24:24] <asac> jserv--: INPROGRESS is the state ... otehrwise our parser doesnt see this as inprogress, but probably as normal TOD [09:24:27] <asac> O [09:24:39] <asac> jserv--: also work items must not be two lines ... that also confuses the parser [09:24:42] <jserv--> patrikryd, sure. I can git merge it into LEB-panda branch [09:24:45] <asac> https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-leb-gingerbread-pandaboard-multimedia -> let me clean this [09:26:10] <asac> jserv--: ok ... please put comments to the whiteboard on top of the work itesm like i did now [09:26:23] <asac> jserv--: we can link the bug you mentioned to the blueprint ... in that way it becomes a work item [09:26:26] <asac> but only if it blocks the blueprint [09:26:48] <jserv--> roger it. [09:26:56] <jserv--> https://wiki.linaro.org/Platform/Android/Meetings/2011-04-20#Jim_Huang_.3Cjserv--.3E <-- [09:27:08] <asac> jserv--: shouldnt we make "libaudio integration in gingerbread pandaboard leb" a blueprint? rather than a bug? [09:27:17] <jserv--> No interesting issue: No visible Android multimedia logs for reporting! [09:27:46] <jserv--> asac, I think libaudio itself is a bug since it doesn't work with omapzoom's OMX. [09:28:08] <asac> jserv--: lets first integrate libaudio in the pandaboard leb ... and then look if it makes sense to make this a consolidated solution for other boards as well [09:28:16] <asac> jserv--: ok [09:28:40] <asac> your call [09:28:55] <asac> if this is more than 2 days of work, make a blueprint with a few work items out of it is my guess [09:29:03] <jserv--> ok [09:29:05] <asac> especially if its about moving something to leb / generic branch [09:29:21] <asac> ok sorry for the detour [09:29:24] <asac> its me running the meeting ;) [09:29:27] <asac> we were at actions [09:29:40] <asac> * Team to review what they are working on and document what isn't approved/doesn't have a Blueprint/is not targeted to discuss in next weeks meeting [09:29:43] <asac> so yeah [09:29:56] <asac> this is just a reminder that we want to have blueprints or bugs for everything we do [09:30:05] <asac> and if its about new topics that we want to have a quick review of the direction and work item [09:30:21] <asac> my idea is that noone should have more than 3 open things to work on at the same time [09:30:28] <asac> jserv-- can probably take 4 ;) [09:30:34] <patrikryd> :) [09:31:15] <asac> so if you currently have 3 blueprints active (e.g. approved, not finished, milestoned against this or next month) and you think there is something more important to work on [09:31:35] <asac> talk to me and jamie and we will think about what of the current blueprints we either wrap up early or take off the CPU so you can take another blueprint [09:31:44] * jserv-- nods. [09:31:51] <patrikryd> ok [09:32:04] <asac> i want to do blueprint approval in this meeting, but if it blocks you you can always ping me or jamie anytime [09:32:32] <asac> its importnat that we manage the blueprint queue decently and deliver stuff like a machine (e.g. setting things to implemented) ;) [09:32:57] <asac> * asac and JamieBennett to talk to james_w about keeping android blueprints on android team radar [09:33:04] <asac> i havent talked to james ... [09:33:31] <asac> but i think now that we are close to know how exactly we want to work with those blueprints we can start thinking about how to present our work and track progress [09:33:43] <asac> i guess we will only have monthly burndowns because we dont plan in cycles [09:34:02] <patrikryd> Was that about being visible on status.l.o [09:34:04] <patrikryd> ? [09:34:08] <asac> basically 1. monthly burndown for current month; 2. work items already planned for next month ... and a list of project horizon [09:34:13] <asac> patrikryd: i think so [09:34:23] <asac> tricky part is that we had no TR to start with [09:34:55] <asac> so whatever we do now would be munged into tr-platform-all (which coveres everything that came late ... vlidation and android) [09:35:02] <asac> for that we just need to link the blueprints to that as the parent [09:35:14] <asac> from may on we have real TRs [09:35:33] <asac> * asac to think about validation meta blueprint and what to do with that with jamie and jeremychang [09:35:36] <asac> what was that about? [09:36:01] <asac> i think its about https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-validation ? [09:36:05] * patrikryd hopes that we will have time to discuss TRs, BPs and Budapest at the end of the meeting... [09:36:20] <jserv--> asac, IIRC, it was [09:36:33] <asac> patrikryd: yeah we can directly skip to that topic after actions are done i guess [09:36:36] <asac> jserv--: thanks [09:36:39] <asac> so ... [09:36:49] <asac> i think we should rename it to linaro-android-validation-initial ;) [09:36:54] <jeremychang> asac, ya. It's still meta. [09:36:55] <asac> and remove the cts an sikuli [09:37:00] <asac> those will come later [09:37:08] <asac> e.g. just have those below that that we will deliver for 11.05 [09:37:37] <asac> jeremychang: do you think if we remove cts and sikuli we can complete all blueprints below in may timeframe? [09:37:44] <asac> or should we also remove monkey? [09:38:07] <jeremychang> asac, cts run could be a time consuming long run. [09:38:30] <asac> jeremychang: ok ... so until we have a lava scheduler we wouldnt run this i guess [09:38:35] <asac> jeremychang: how long is "long run"? [09:39:00] <jeremychang> asac, Regardless of cts and sikuli, others should be ok to be integrated in Lava . [09:39:17] <asac> jeremychang: by 11.05? ;) [09:39:25] <jeremychang> asac, I mean in may timeframe. [09:39:27] <jeremychang> asac, yeah. [09:39:29] <asac> yeah [09:39:30] <asac> ok cool [09:39:46] <asac> so lets do this: i will rename to initial (e.g. for 11.05) and remove the dependencies for sikuli and cts [09:39:50] <asac> and lets put them on CPU in june [09:40:17] <jeremychang> asac, for complete cts plan run, it will take 5 hours and more at least. [09:40:26] <asac> kk [09:40:42] <asac> we need the lava scheduler first ... but once thats there having 5 or even 24h running tests should be ok [09:40:50] <jeremychang> asac, for 11.05. agreed.! [09:40:52] <asac> we want to stack up many boards ... but in order to use them we need the scheduler [09:41:49] <jeremychang> asac, not tried the scheduler yet. [09:42:41] <asac> jeremychang: scheduler is not your task ... validation team is late on that and will land late may only [09:42:44] <asac> https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-validation-initial [09:42:58] <asac> renamed, commented, dropped cts, sikuli and approved for 11.05 [09:42:59] <asac> thanks [09:43:12] <asac> * patrikryd to sort out href8500 build situation [09:43:20] <asac> patrikryd should we kill it or is there hope ;)? [09:43:47] <patrikryd> There is no hope and the Snowball will be sent next week... [09:43:58] <asac> patrikryd: i have ordered a snowball for you so we should rather skip u8500 imo [09:44:01] <asac> yeah [09:44:01] <asac> ok [09:44:02] <patrikryd> ... to a lucky few. [09:44:22] <asac> [ACTION]: patrikryd to file a bug about killing u8500 build and put rational there; then kill build and get over it ;) [09:44:31] <patrikryd> I'll write and bug report to remove it [09:44:37] <asac> patrikryd: you will get one [09:44:39] <patrikryd> ok [09:44:53] <patrikryd> I'll get one [09:45:10] <asac> patrikryd: you get one outside of our pool? i told them to give you one from our pool [09:45:11] <patrikryd> And so will Linaro... [09:45:22] <asac> if you get one without that i will ask them to send the board to someone else [09:45:26] <patrikryd> It's from the Linaro pool [09:45:29] <asac> ok [09:45:31] <asac> right [09:46:05] <asac> * jserv-- to email JamieBennett and asac about what is needed from TI [09:46:09] <asac> ok this was done afaict [09:46:14] <jserv--> my turn :( [09:46:15] <asac> * jeremychang to email plars (Paul Larson) to review Android changes to LAVA [09:46:20] <asac> jserv--: you did that, didnt you? [09:46:29] <asac> jserv--: iirc i replied that you should ask questions to the TI folks ;) [09:46:32] <asac> (and keep me CCed) [09:46:36] <jeremychang> asac, I did that for the initial. [09:46:43] <jserv--> asac, but I didn't handle well. I will explain in LEB report. [09:46:44] <asac> jeremychang: right. that is done [09:46:48] <asac> jserv--: sure [09:47:09] <asac> jserv--: you want me to carry over that actin then? for me its done [09:47:14] <jeremychang> asac, will push more to my branch and discuss with plars more further. [09:47:29] <asac> thanks jeremychang !! [09:47:35] <jserv--> asac, pass it :) [09:47:37] <asac> jeremychang: also lamc final fixes? [09:47:43] <asac> jserv--: agree!! [09:47:47] <asac> [TOPIC] Team issues discussion - Alexander and JamieBennett asking questions [09:47:52] <asac> i think i ask enough questions [09:48:01] <jeremychang> asac, right, lamc can be used to do deployment, [09:48:02] <asac> let me skip that ... only for high level summary the headlines: [09:48:21] <asac> * jserv-- submitted 6 optimizatoin patches to AOSP this week [09:48:22] <jeremychang> asac, I will do the rebase, then I think it could be merged. [09:48:43] <asac> * jserv-- landed experimental GFX support for pandea LEB [09:49:08] <jserv--> asac, it is by-product. (AOSP) [09:49:15] <asac> * jserv-- started to work on MM integration for panda LEB; but has some problems [09:49:44] <asac> * jeremychang submitted first iteration of lava integration [09:49:51] <asac> * patrikryd setup panda leb manifest [09:50:04] <asac> * asac presented android platform story to google at android builder summit [09:50:22] <asac> ok thats good enough for headlines ... if i missed something then tell me and i will add it [09:50:36] <asac> [TOPIC] LEB/Image review - jserv [09:51:18] * patrikryd created the panda leb manifest while eating breakfast this morning. :) [09:51:19] <asac> jserv--: ? [09:51:25] <asac> patrikryd: \o/ [09:51:35] <jserv--> The major focus of LEB development is targeting Pandaboard these week. [09:51:35] <asac> patrikryd: see how small things can produce big headlines ;) [09:51:41] <patrikryd> :) [09:52:04] <jserv--> GFX integration should be done, but multimedia (MM) is complex. [09:52:19] <asac> yeah ... thats why its targetted for may ;) ... and not april :-P [09:52:20] <jserv--> I want to highlight two major issues: [09:52:52] <jserv--> (1) There is No visible Android multimedia logs for reporting! (for stagefright) [09:53:00] <jserv--> (2) No software available, so it is hard to perform multimedia tests. [09:53:06] <asac> stop: what does "Logs" mean here? [09:53:09] <asac> commit logs? [09:53:16] * jserv-- is sorry that copy-n-paste doesn't work for unknown reasons. [09:53:17] <asac> or are you looking for logs on the console/logger? [09:53:28] <asac> jserv--: that didnt really look that bad [09:53:44] <asac> jserv--: ok ... so what about "no software available" [09:53:49] <jserv--> asac, I need stagefright specific logcat in order to analyze. [09:53:52] <asac> you say: no video playback application by default? [09:54:18] <jserv--> asac, but, stagefright doesn't generate enough logs to Android Log system. [09:54:35] <jserv--> asac, no, I can not use any UI program [09:54:35] <asac> ic [09:54:45] <asac> jserv--: why is that? [09:54:51] <jserv--> I attempt to use CTS to invoke UI operations [09:54:56] <jserv--> but no luck. [09:55:10] <asac> lets go step by step ... whats the problem with UI programs? [09:55:14] <asac> no pointer/mouse/keyboard? [09:55:33] <jserv--> asac, yes, Video application need pointer device. [09:55:52] <jserv--> asac, sometimes, keyboard can not focus on certain UI items. [09:55:57] <asac> cant we use monkey runner? [09:56:14] <jserv--> asac, that is my test. [09:56:28] <asac> ok, so this boils down to: we need a touch screen we can use for all our boards [09:56:49] <asac> can someone take an action to find a suitable touch screen that has linux driver support and at best is multitouch capable? [09:56:55] <jserv--> asac, or, we can rework software cursor patch into LEB-gingerbread. [09:56:57] <asac> i am happy to buy tat stuff and ship to android [09:56:58] <patrikryd> Or a hack for mouse like https://groups.google.com/d/topic/rowboat/wisbAU7LXV8/discussion [09:57:13] <jserv--> patrikryd, yes, I am considering. [09:57:19] <asac> jserv--: i wouldnt want to integrate that patch into LEB ... what we could do is a branch based off from LEB with MOUSE support [09:57:34] <asac> and have that as an unofficial testbuild running [09:57:48] <jserv--> asac, agree. so, I need to integrate this change and verify this first. [09:57:51] <asac> of course would mean some maintenance overhead (e.g. you have to regularly merge latest leb-gingerbread stuff down) [09:58:23] <asac> jserv--: that or we just buy a touchscreen [09:58:26] <jserv--> Besides the above issues, TI OMX (DSP) components were loaded in Pandaboard. [09:58:27] <asac> i would really like to have a touchscreen [09:58:29] <asac> REALLY [09:58:36] <jserv--> Audio was tested. [09:58:47] <asac> but noone can give me an answer to my question: what touchscreen should we buy [09:58:52] <jserv--> https://wiki.linaro.org/JimHuang/Sandbox/AndroidForPanda <-- [09:59:16] <jserv--> asac, I think we should not consider touchscreen at the moment. [09:59:26] <jserv--> asac, no hardware change required. [10:00:04] <asac> jserv--: well. given that google considers mouse bad ... and if a touchscreen would work without changing android, why wouldnt we want to have a "standard touchscreen" that we officially support for our "real" leb builds? [10:00:08] <jserv--> after merging TI's changes into git.linaro.org, the following are the pending tasks: [10:00:16] <jserv--> libaudio (from omapzoom) [10:00:17] <asac> as i said i dont want to integrate mouse stuff in the official leb builds (only in test builds) [10:00:21] <jserv--> liboverlay (from omapzoom) [10:00:22] <asac> go ahead [10:00:26] <jserv--> libstagefrighthw - hardware/omap3 [10:00:50] <jserv--> The minor problem is TI's directory hierarchy is the same as Gingerbread. [10:01:09] <jserv--> That is, HAL should live inside device/linaro/pandboard [10:01:15] <asac> right. that was expected ... which is why we said we most likely will have leb-gingerbread-pandaboard branches [10:01:24] <jserv--> instead of hardware/ti/omap3 <-- this is misleading. [10:01:28] <asac> hmm. ok [10:02:24] <jserv--> we discussed bootargs before, and we might hard code DSP + GFX enabled boot args first during deployment. [10:02:28] <patrikryd> jserv--: Why not branches? [10:02:33] <asac> jserv--: right. but hal should live in a leb-gingerbread branch then [10:02:45] <asac> i dont want our "generic" build to get the HALs for now [10:02:55] <jserv--> patrikryd, I am just afraid of licensing issues, and I did git force push often. [10:04:01] <asac> jserv--: can we build this multimedia HAL without having all the problematic binaries? [10:04:07] <asac> and just drop in the binaries after build? [10:04:11] <asac> e.g. as binary image overlay? [10:04:17] <jserv--> asac, only audio is verified [10:04:42] <jserv--> asac, I need liboverlay + libstagefrighthw to verify video playback. [10:05:13] <asac> jserv--: those are proprietary binaries? [10:05:21] <jserv--> asac, yes [10:05:37] <asac> jserv--: and those can be dropped in after the build? [10:05:42] <jserv--> asac, sorrry, liboverlay + libstagefrighthw are open source'd [10:05:46] <asac> ah [10:05:50] <asac> right [10:05:53] <jserv--> asac, but it requires binaries from TI [10:05:57] <asac> to build? [10:05:58] <asac> or to run=? [10:05:59] <jserv--> asac, some .so files [10:06:29] <asac> if we can make liboverlay and libstagefrighthw build without those binaries [10:07:05] <jserv--> asac, yes, we can build, but we need DSP loader to load TI OMX components. [10:07:10] <asac> sure [10:07:23] <asac> jserv--: thats good. so lets build everything, dont include the binaries in our git [10:07:26] <asac> and make a overlay tarball [10:07:34] <asac> that you unpack after installing the image to sdcard [10:07:39] <jserv--> asac, including GFX?! [10:07:41] <asac> the overlay tarball should be minimal [10:07:45] <asac> jserv--: no... GFX is fine [10:07:49] <asac> because license is ok [10:07:51] <jserv--> asac, got it. [10:07:52] <asac> audio isnt [10:07:55] <asac> mm isnt [10:07:56] <asac> i mean [10:08:07] <asac> jserv--: at best the build still boots and works without those binaries [10:08:21] <asac> jserv--: for now put the tarball in your people.linaro.org: home directory [10:08:31] <asac> i will sort out how to distribute that [10:08:33] <jserv--> but if there is no binary available, mediaserver can not run, and Android UI doesn't show. [10:08:38] <asac> jserv--: not in public_html please though ;) [10:08:52] <asac> jserv--: ok. that meas we need to work on it [10:08:58] <asac> jserv--: but its ok to have it first as a build [10:09:02] <jserv--> ok. [10:09:18] <asac> jserv--: i guess mediaserver can be fixed to use software codecs etc. if the binaries are not available [10:09:20] <jserv--> That is all work of LEB recently. [10:09:25] <asac> thanks! [10:09:30] <asac> good progress. thanks a lot [10:09:35] <asac> lets get the GFX enabled build out today [10:09:38] <asac> or tomorrow [10:09:43] <asac> and lets focus on multimedia after that [10:09:45] <jserv--> asac, we need extract fixes for mediaserver, and I can submit to AOSP later. [10:09:52] <jserv--> ok [10:10:05] <asac> i need to have a build i can give to george because he wants to show that to linaro board ... so this is really important ;) [10:10:28] <jserv--> By the way, another Pandaboard is arrived at 0xlab. [10:10:28] <asac> this build just needs GFX [10:10:41] <jserv--> so that, more developers can work on Pandaboard here. [10:10:42] <asac> jserv--: right. i think we sent it to you ... unless you ordered that on your own [10:10:51] <asac> ok [10:11:03] <jeremychang> for another tarball for these binaries, will that be related to the deployment doing in l-a-m-c ? [10:11:29] <asac> jeremychang: yes. we will add a --overlay option that just unpacks the tarball and maybe displays a license before that user has to accept [10:11:42] <asac> jeremychang: if you want to go ahead just add --overlay that just unpacks [10:11:55] <asac> dont want to work on the license side until we really know what we need [10:12:26] <jeremychang> asac, got it. [10:12:50] <asac> [TOPIC] Validation update - Jeremy [10:12:54] <jeremychang> asac, also lava does the deployment job [10:13:00] <asac> jeremychang: make it short and crisp we are well over time [10:13:23] <jeremychang> For lava work this week, I have not pushed to https://code.launchpad.net/~jeremychang/lava/android-support, [10:13:32] <jeremychang> I think I nearly made the basic validatin in lava done. [10:13:41] * patrikryd hopes that we will have time to discuss TRs, BPs and Budapest at the end of the meeting... [10:14:00] <asac> patrikryd: noted ... i am thinking about something else to allow you have lunch ;) ... wait a few seconds [10:14:11] <asac> jeremychang: cool. so what works and what doesnt? [10:14:24] <jeremychang> I will arrange and push my work in lava to my branch asap and have a further review by plars [10:14:32] <asac> so here is the thing [10:14:34] * patrikryd needs lunch as usual :) [10:14:56] <asac> plars is currently whining a bit that we want to keep the full partition layout for validation [10:15:03] <asac> seems we are still 1 partition ahead of what is kernel default [10:15:28] * plars is such a crybaby [10:15:28] <jeremychang> I still need to verify more about deployment and submit_result job for basic validation. [10:15:30] <plars> :) [10:15:36] <asac> hehe [10:15:55] <jeremychang> plars, :) [10:16:07] <asac> so question is ... can we make an exception and just have one big partition? for system/userdata/cache ? [10:16:13] <asac> and then one for sdcard? [10:16:21] <asac> what do you think? [10:16:40] <jeremychang> asac, right. [10:16:48] <asac> plars: thinking about it ... i think we can fix it by doing what i always wanted ... share the same vfat partition [10:16:51] <asac> for boot [10:16:59] <asac> e.g. we have eliminated the roottfs [10:17:11] <asac> if we now eliminate the boot and just reuse the one that is already there all should work [10:17:28] <plars> asac: on freescale we need an extra partition besides the typicall boot/rootfs layout [10:17:32] <asac> right [10:17:33] <asac> ok [10:17:38] <jeremychang> asac, at least we need to remove one more partition after ramdisk introduced. [10:17:39] <asac> jeremychang: can we do that then? [10:17:53] <asac> jeremychang: sure, but we still have 1 partition too much [10:18:02] <asac> jeremychang: in short term we have no choice but to kill all for lava [10:18:23] <asac> i think: boot + combined system/userdata/cache partition ... and sdcard [10:18:26] <jeremychang> asac, so system/userdata/cache become a one? [10:18:42] <asac> i think so though not sure if that works [10:18:48] <asac> because root partition is ramdisk [10:18:56] <asac> what happens if you mount a partition to / ? [10:19:02] <asac> i guess that hides everything from ramdisk? [10:19:28] <jeremychang> asac, I am ok for this, and make the size as added together. [10:19:34] <asac> plars: we dont need the extra partition btw [10:19:47] <asac> plars: we can have just an offset ... which is the official instruction anyway [10:19:56] <asac> the non-fs data is just a thing that ubuntu invented kind off [10:20:02] <plars> ok [10:20:17] <asac> jeremychang: ok, please try what happens if we mount system etc. to / [10:20:27] <asac> if that causes problems we should think aobut sharing the boot partitoin in lava [10:20:51] <asac> we definitly wont go back to non-ramdisk rootfs [10:20:54] <asac> that would be stupid [10:21:01] * asac so happy that we made this move [10:21:02] <asac> ok [10:21:05] <asac> thanks jeremychang [10:21:41] <jeremychang> asac, though how about the partition layout work in l-a-m-c? [10:21:43] <asac> [ACTION] jeremychang to add work item to one of his blueprint to try mounting system/userdata/cache combined to / ... if that fails, we merge validation boot parititons [10:21:50] <asac> jeremychang: that should stay as planned [10:22:22] <asac> jeremychang: --boot --system --userdata ... each getting tarball ... and add --overlay that just unpacks a tarball on / level afterards [10:22:42] <asac> and we create a /sdcard partition of course [10:22:45] <asac> jeremychang: make sense? [10:22:57] <jeremychang> asac, ok. only for used by lava. [10:23:09] <asac> right ... lava gets this hack [10:23:26] <asac> if not we will poke plars until he gives in and admits that its ok to share a boot partition [10:23:38] <patrikryd> combined /, /system, /data -> some errors when running CTS. [10:23:46] <asac> thats fine for now [10:23:50] <asac> we will survive [10:23:59] <patrikryd> agree [10:23:59] <asac> we can fix the boot partition later anyway [10:24:09] <asac> but dont want to put any load on validation team [10:24:12] <asac> to make this happen [10:24:13] <asac> for now [10:24:20] <asac> so we can deliver stuff running by 11.05 [10:24:21] <jeremychang> asac, ok, basically we make the sdcard for lava extra manually. [10:24:46] <asac> jeremychang: lets talk about that offline after the meeting [10:24:57] <asac> [ACTION] asac and jeremychang to sync on validatoin partition layout and lamc [10:25:05] <jeremychang> patrikryd, what kind of the error could happen when running CTS ? [10:25:08] <asac> [TOPIC] Bug status - Patrik [10:25:20] <jeremychang> asac, ok. [10:25:24] <asac> jeremychang: probably CTS test testing that the partitions are read only and have the right mount flags [10:25:31] <asac> e.g. noexec for data/cache etc. [10:25:52] <patrikryd> jeremychang: Nothing important. There are test cases to check that you can not write to /system .... [10:26:11] <patrikryd> We have 9 open bugs. [10:26:17] <jeremychang> asac, yeah. system partition should be ro by default while data and cache not. [10:26:24] <asac> patrikryd: anything important? panda booting with UI now? [10:26:27] <patrikryd> Only the (old) bug 747911 is new [10:26:28] <ubot2> Launchpad bug 747911 in linaro-android "linaro-generic[panda]: init.omap4.sh not executable" [High,New] https://launchpad.net/bugs/747911 [10:26:38] <asac> patrikryd: do our images work now? e.g. is the refactoring done for ramdisk etc.? [10:26:46] <asac> patrikryd: right. that one should be easy to fix, no? [10:27:07] <patrikryd> asac: will check it once I have Panda with UI. [10:27:12] <asac> otherwise we could fix init.sh to not run ./.... but rather sh /etc/..../init.omap4.s [10:27:43] <asac> patrikryd: yeah. 747911 causes the UI to not display for me for sure [10:27:53] <asac> so remember to manually fix that when testing ;) [10:28:15] <patrikryd> My problem is earlier. [10:28:43] <asac> can you file a bug with symptoms? [10:28:46] <asac> is it about mounting? [10:28:51] <patrikryd> According to jeremychang the problem I have is probably related to building kernel with standard Android toolchain. [10:29:49] <patrikryd> Will build kernel with linaro toolchain and see if it helps [10:29:58] <asac> thats such a mistery [10:30:11] <asac> patrikryd: couldnt yu just use TARGET_TOOLS_PREFIX? [10:30:25] <patrikryd> I can [10:30:29] <asac> patrikryd: i mean for me the build form our builders clearly worked in that it showed stuff on serial [10:30:42] <asac> and that uses the android toolchain [10:30:47] <asac> oh right [10:30:54] <asac> we have this painful -Werror hack still :/ [10:30:54] <patrikryd> But I don't want want to use it for the rest of android (at the moment) [10:31:00] <asac> i understand [10:31:08] <asac> patrikryd: just build the kernel manually [10:31:13] <asac> patrikryd: maybe thats what i did to be honest [10:31:22] <patrikryd> will try that after lunch [10:31:22] <asac> patrikryd: ssh into our dev box [10:31:31] <asac> i used the ubuntu cross toolchain iirc [10:31:39] <asac> now i remember ;) [10:31:47] <asac> ok good ... [10:32:10] <asac> great [10:32:11] <asac> so [10:32:13] <asac> thanks [10:32:17] <asac> patrikryd: do we have a bug on that? [10:32:18] <asac> jeremychang: ? [10:32:22] <asac> if not please file one [10:32:29] <jserv--> so, should we pick changes in toolchain-preview branch to leb now? [10:32:32] <asac> [ACTION] patrikryd and jeremychang to file a bug for kernel boot with android toolchain problem [10:32:46] <asac> jserv--: no ... i will not take -Werror disabling to leb [10:32:58] <asac> at least not until i have aDother dream that inspires me [10:33:04] <jserv--> asac, which means, no single toolchain solution? [10:33:27] <patrikryd> jserv-- [10:33:44] <asac> jserv--: not the official LEB ... maybe a leb testbuild for benchmarking [10:33:48] <jeremychang> asac, nod. file a bug for this. [10:33:51] <patrikryd> jserv--: How hard do you think it will be to solve the problems on the toolchain branch? [10:33:52] <asac> otherwise we should fix it for real [10:34:10] <asac> jserv-- doesnt want to fix the build and has other stuff to do [10:34:20] <asac> i will see if zach wants to take a stab at that [10:34:23] <asac> for may [10:34:43] <asac> s/doesnt want/shouldnt/ [10:34:50] <patrikryd> jserv--: just asking for an estimate of the amount of work. [10:34:59] <jserv--> asac, but, if we don't test regularly, we might not be able to figure out potential problems. [10:35:00] <asac> i think would be 1 week [10:35:14] <jserv--> patrikryd, did you mean 'real' fixes? [10:35:21] <asac> jserv--: right. we can have a testbuild where we merge our stuff on top of the hacked tooclhain branch [10:35:24] <patrikryd> jserv--: yes [10:36:00] <jserv--> patrikryd, I don't think it is worth doing since Google did it in internal tree already. [10:36:07] <patrikryd> jserv--: I know that they would be hard to get to AOSP. [10:36:21] <patrikryd> jserv--: I know... [10:36:35] <patrikryd> jserv--: just wanted a guess on the amount of work. :) [10:36:41] <jserv--> patrikryd, we just need the integrated development flow, I think. [10:36:45] <asac> jserv--: just an estimate would be good ... dont be scared ... it wont be you [10:36:49] <jserv--> single toolchain, single kernel tree. [10:37:12] <jserv--> asac, I don't think disabling Werror is wrong. [10:37:20] <asac> jserv--: thats not the question ;) [10:37:34] <jserv--> it makes us concentrate on other important thing. [10:37:47] <asac> yes, we already had that discussion ;) [10:38:05] <jserv--> ok, please ignore me. [10:38:06] <asac> and as i said you have other stuff to do ... but understanding how much effort a clean solution would be would be great [10:38:31] <patrikryd> If we solve the problems maybe more people starts using our toolchain. [10:38:35] * jserv-- would like to git cherry-pick from Honeycomb tree instead. :( [10:38:41] <asac> i know [10:38:42] <asac> i know [10:38:56] <asac> for now we can set up a test build and merge our leb on top of the -Werror fix [10:39:06] <asac> but lets first get the leb working please [10:39:15] <asac> ok ... we are 40 minutes over time ;) [10:39:17] * patrikryd knows that you are not allowed to do that even if you have the source code. :) [10:40:00] <asac> [ACTION] patrikryd to nail down kernel build bug ... in worst case allow using the ubuntu cross toolchain for building kernel [10:40:04] <asac> [TOPIC] AOB [10:40:09] <asac> blueprints, topics etc. [10:40:19] <asac> can we have a nother meeting tomorrow 0900 UTC? [10:40:21] <asac> about that? [10:40:31] <asac> i really this will take too long now [10:41:15] <asac> looking at leaves calendar seems everyone is still here [10:41:18] <asac> tomorrow [10:41:19] <asac> so lets do that [10:41:25] <asac> unless someone screams [10:41:27] <patrikryd> 0900 UTC -> ??? CET? [10:42:20] <asac> patrikryd: 0900 UTC is the same time we had today [10:42:22] <asac> 1100 [10:42:26] <asac> your and my time [10:42:32] <jserv--> ok [10:42:33] <patrikryd> ok [10:42:40] <jeremychang> ok. [10:42:43] <asac> ok added invite [10:42:46] <jeremychang> time goes by so fast. [10:42:48] <asac> sent out to you gues [10:42:50] <asac> guys [10:42:52] <asac> yeah ;) [10:42:58] <asac> i know ... i know :/ [10:43:01] <asac> ok thanks everyone [10:43:05] <asac> sorry for the long meetings ;) [10:43:22] <asac> but due to timezone its the most effective time we have [10:43:32] <asac> where we all can talk at the same time and everyone is focussed and so on [10:43:40] <asac> cu tomorrow latest! [10:43:42] <asac> enjoy [10:43:45] <asac> #endmeeting Meeting ended.
Platform/Android/Meetings/2011-04-20 (last modified 2011-04-26 11:59:50)