Android Kernel Meeting
- Follow up on previous meetings actions
- John to ping AmitK on power mgmt interface
- Noritsuna to send John virtual battery driver patch
- John to cherry pick patches from how-to-porting into linaro+android git tree
- Linaro+Android 2.6.37 kernel update
- Questions on Noritsuna's schedule201102.pptx slides
- Linaro Android release schedule questions
- Discuss anything new for this week.
- Noritsuna to check with Mr.Sola and others to see if there is a better time/date where more people can join.
- John to email Mr.Sola about his power management interface work.
- Noritsuna to send John virtual battery driver patch.
- John to review git tree Mr.An has posted.
- Noritsuna to send John more details about what parts of ashmem he thinks need to be more generic.
- John to talk with Patrik about release schedule and code freezes
Android Team Meeting
- Actions from last meeting
- asac to follow up on linaro.android kernel with jstultz
- kjnabe to contribute framebuffer support back to jstultz git.linaro.org tree
- kjnabe to contribute beagle defconfig to jstultz git.linaro.org tree
- jserv-- and jeremychang to confirm that battery support works well and then backout our battery HACK from git.linaro.org
- patrikryd to fix the tarball generation and send patches for review
- jeremychang to bring up device/linaro/pandaboard
- patrikryd to explore with jserv-- and luse what options we have to ship toolchain
- luse to expand toolchain wiki page on distribution format options of toolchain and make a new wiki page on toolchain builder (or wrap in the toolchain main page)
JamieBennett to setup weekly reports for the team
- blueprint and spec for git infrastucture
- spec for git process, contrib. infrastucture and contrib. infrastructure
- Create tar balls when building Android
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).
- Any issues you currently face.
Android Kernel Meeting
<jstultz_vm> noritsuna: hey! <jstultz_vm> noritsuna: How are you doing? <noritsuna> We are researching about "Power Management(wake locks)", "Ashmem,pmem", "USB Composite driver". <jstultz_vm> noritsuna: Do you expect anyone else to join? <jstultz_vm> noritsuna: very cool! <noritsuna> Mr.Sola joined. <jstultz_vm> noritsuna: oh, is he with you? <jstultz_vm> noritsuna: or should we wait for him to join irc? <noritsuna> Mr.Sola is this site's owner(http://blog.sola-dolphin-1.net/). <noritsuna> He is professional of Android porting. <jstultz_vm> noritsuna: Ok. Very good! <jstultz_vm> noritsuna: So I posted the agenda here: https://wiki.linaro.org/Platform/Android/2011-02-16 <noritsuna> Sorry, Sola doesn't join this meeting. Because this time is his commuting time:-) <jstultz_vm> noritsuna: so lets start off with the previous items from our meeting two weeks ago. <jstultz_vm> noritsuna: Ok. Would it be better to move the meeting by an hour (earlier or later?) <noritsuna> I will hear it him. <jstultz_vm> noritsuna: Ok. Well, do let me know if the meeting time should be adjusted if there are others interested in joining. <jstultz_vm> noritsuna: So back to the agenda: <jstultz_vm> noritsuna: I talked briefly with AmitK about power management interfaces. <jstultz_vm> He wasn't sure if he understood what you were referencing in particular about consolidated power managment interfaces. <jstultz_vm> So I wanted to check and see if you could explain a bit more. <noritsuna> It's the Sola's work. but he hasn't finished it yet. <jstultz_vm> noritsuna: Ok. Should I send an email to Sola to try to better understand the details of what he's wanting to do? <noritsuna> He said "I will make Power Management Guideline for Android.". <noritsuna> It's first step. Maybe We will be able to discuss about Power Management by it. <jstultz_vm> noritsuna: Ok. I'll do that then. <jstultz_vm> noritsuna: So next on the list, the virtual battery driver patch. Did you have any time to look at sending that to me? <noritsuna> Sorry,,,I sent now. <jstultz_vm> noritsuna: Great! I appreciate it! <jstultz_vm> noritsuna: I'll make sure it gets included into the Linaro+Android kernel. <jstultz_vm> noritsuna: So third item from the last meeting: cherry picking the patches in your how-to-porting page. <jstultz_vm> noritsuna: I went through and tried to apply those changes. <jstultz_vm> noritsuna: Only 2 of the 4 patches actually applied, the others seem to have already been integrated into the 2.6.37 kernel (or a variant of it) <jstultz_vm> noritsuna: Once applied, I saw no difference with the changes then without. <jstultz_vm> noritsuna: So I wanted to check with you again to make sure that you feel they are necessary on top of the 2.6.37 based Linaro+Android tree. <jstultz_vm> (Maybe explaining how I can reproduce the problems they fix) <noritsuna> Mr.An finished 2.6.37's porting. He uploaded it to gitorius repository. Please wait a minute. I looking for this site's URL. <noritsuna> git://gitorious.org/kj-android/jstultz-linux-37-linaro-android.git <jstultz_vm> noritsuna: ah! <noritsuna> This site's kernel is 2.6.37 for Android. It's your repository base. <jstultz_vm> noritsuna: let me take a look at that. <jstultz_vm> noritsuna: Ok. I'm glad to see Mr.An is working with my tree. <jstultz_vm> noritsuna: I'll review it for changes, and see if there is anything I need to merge back in. <jstultz_vm> noritsuna: Meanwhile, do let Mr.An know that if he runs into any problems, to please let me know. I'd be happy to help solve any issues found in my tree. <noritsuna> OK. <jstultz_vm> noritsuna: So on that topic, I wanted to let you know that I've updated my Linaro+Android tree to include the latest bits from both kernel trees. <jstultz_vm> I tested it and it seems to be running well. <jstultz_vm> So I also wanted to ask about your schedule201102.pptx slides you sent me. <jstultz_vm> I was curious about your ashmem work listed there. <jstultz_vm> noritsuna: Could you describe it in any more detail? Are you trying to just push the ashmem work into mainline? Or are you modifying it? <noritsuna> No. This has strongly relation with hardware. And this implement is each boards. Therefore we have to port it each board of kernel. We want to make interface for ashmem to boards. <jstultz_vm> Not sure I'm understanding. Do you mean that the ashmem driver has to be ported for each board, because it isn't generic enough? And you're trying to create a more generic interface? <noritsuna> Yes. <jstultz_vm> noritsuna: I guess I'm not sure I see how it is not already generic? What portions have to be ported to each board? Do you have an example patch? <jstultz_vm> noritsuna: from my earlier review, I don't recall seeing much arch specific code there. <noritsuna> Sorry, I don't have example patch. <jstultz_vm> noritsuna: Ok. Well, I'd be interested in hearing more about what parts need to be ported to each board. Mind sending me an email about it? <jstultz_vm> I want to cover the last few bits of the agenda before our hour is up. <jstultz_vm> I replied to your schedule question, and hopefully Patrik will provide more details. <jstultz_vm> Feel free to reply to that email if you have any further questions. Clearly the Linaro Android Platform is a little behind the regular Linaro releases (as there hasn't been a alpha posted yet), but I think the plan is to try to catch up. <noritsuna> Oh,,, Each boards don't have specific code. If I reseach this week and I don't find specific code, I will do other work. <jstultz_vm> noritsuna: Are there any other issues you want to bring up for this week? <noritsuna> No. <jstultz_vm> noritsuna: Ok. Well that's all then! Thank you so much for joining the meeting! <noritsuna> I think so about Android-Linaro's schedule. If Linaro's common kernel isn't fixed, our job is hard. <noritsuna> Thx! <jstultz_vm> noritsuna: So what do you mean specifically by "Fixed"? <noritsuna> =code freeze <jstultz_vm> noritsuna: Ok. We have to offset a code freeze with resolving issues we find (and staying current with the Linaro kernel) <jstultz_vm> noritsuna: Finding the right balance will be needed. <jstultz_vm> noritsuna: Hopefully as the Linaro standard kernel slows down, the Linaro+Android kernel should also have fewer changes. <jstultz_vm> noritsuna: I think a hard freeze should be done by 2011-04-14. <jstultz_vm> noritsuna: (as that is when the Linaro tree is frozen) <noritsuna> Sure. We think that porting Android isn't hard work. But We don't understand when we shuold start to port yet. Please teach us:-) <jstultz_vm> noritsuna: Ok. I'll talk with Patrik and try to see when we will have a initial release from which you can start working with. I think it depends on how integrated into our development cycle you want to be. If you can't handle much change, you probably should wait till closer to the end of the release cycle. <jstultz_vm> But if you start earlier, we can also integrate more of your changes back in before the final release. <noritsuna> We think to port first time on the end of March(for beta freeze). <jstultz_vm> noritsuna: Ok. I'll talk more with Patrik and make sure our schedule is clear for you. <jstultz_vm> noritsuna: Thanks again for joining! I'll see you next week! <noritsuna> Thx! <noritsuna> Hi, see you next week!
Android Team Meeting
Started logging meeting in #linaro-meeting [10:40:16] <JamieBennett> ah, of course it hadn't started [10:40:21] <patrik_> #startmeeting [10:40:22] <JamieBennett> :( [10:40:25] <asac> seems to not work :( [10:40:27] <asac> nevermind [10:40:31] <Daviey> hmm [10:40:41] <asac> yay!!! [10:40:43] <Daviey> \o/ [10:40:46] <asac> here we go [10:41:34] <patrik_> [TOPIC] review actions from last meeting [10:41:44] <jserv--> noritsuna, http://gitorious.org/kj-android/jstultz-linux-37-linaro-android/commit/6fbf84404a7ee1f0a53e889c9752d7b3127fdbec # virtual_battery.c was written by Author: Masashi YOKOTA [10:41:54] <jserv--> noritsuna, but submitted by sola? [10:42:12] <patrik_> the next action was... [10:42:19] <patrik_> "jserv-- and jeremychang to confirm that battery support works well and then backout our battery HACK from git.linaro.org" [10:43:02] <jserv--> yes. [10:43:20] <jserv--> I think we could cherry-pick first. [10:44:01] <jserv--> next? [10:44:35] <jeremychang> ready to try when receiving the patch. [10:44:47] <asac> jeremychang: the kernel patch is above [10:45:20] <asac> since you work on beagle, i wonder if we should just do the beagle defconfig patch on our own [10:45:29] <jeremychang> asac, ok. [10:45:29] <patrik_> ok so we keep the action then? [10:45:38] <jserv--> patrik_, yes [10:45:50] <asac> yeah carry forward i guess [10:45:57] <patrik_> [ACTION] "jserv-- and jeremychang to confirm that battery support works well and then backout our battery HACK from git.linaro.org" [10:46:21] <noritsuna> jserv--: Yes. This is virtual battery patch. [10:46:34] <patrik_> next one was.... [10:46:39] <patrik_> "patrik_ to explore options to get artifacts on sdcard - either dd ext4 .img or backport tarball support" [10:47:12] <jserv--> noritsuna, I meant, what is the relationship between Masashi YOKOTA and sola? [10:47:15] <patrik_> asac wanted tarball, so that is what I'm working on. [10:47:55] <noritsuna> jserv--:Sorry, I don't know "Masashi YOKOTA ". [10:47:58] <patrik_> I still think dd would work, but I think asac said that tar ball fits better with the other linaro tools. [10:48:20] <jserv--> noritsuna, therefore, how can you say, "it was written by sola"? [10:48:41] <asac> yeah, tarball gives more flexibility wrt sizing etc. imo [10:48:41] <jserv--> noritsuna, I'm just curious about it. ok, let's talk offline. [10:49:00] <patrik_> I will continye today on the tarball support. [10:49:25] <patrik_> I have cherrypicked from android master, but there are bug I have to fix. [10:50:12] <asac> patrik_: will we also get a tarball for the rootfs? [10:50:28] <patrik_> [ACTION] patrikryd to fix the tarball generation and send patches for review. [10:50:29] <asac> thats the one piece i couldnt find [10:50:41] <asac> ok [10:50:53] <patrik_> that is where the current bug is. [10:51:02] <asac> ok super [10:51:27] <patrik_> Some variable is not set during the build for some reason. :( [10:51:52] <patrik_> Next action was... [10:52:06] <patrik_> "jeremychang to bring up device/linaro/pandaboard" [10:52:51] <patrik_> jeremychang: Any progress? [10:53:08] <jeremychang> patrik_, I have not give pandaboard a try. [10:53:47] <jeremychang> patrik_, as I know, Matt is using that one. I will try to borrow from Matt. [10:54:04] <jserv--> pandaboard is occupied here. [10:54:22] <asac> jserv--: what are you guys doing? [10:54:26] <asac> jserv--: thats a linaro board [10:54:35] <asac> if matt works on linaro topics thats good [10:54:45] <asac> if not, its not his anymore ;) [10:54:49] <asac> but yours [10:54:50] <jserv--> asac, Matt is trying to modify boot loader. [10:55:07] <jserv--> asac, in order to make Linux kernel boot. [10:55:17] <asac> u-boot doesnt work? [10:55:34] <jserv--> asac, we prefer to use our own Qi boot loader. [10:55:38] <asac> well [10:55:42] <asac> thats not on topic atm [10:55:53] <jserv--> asac, anyway, we are trying to boot. [10:55:56] <asac> its ok if you share the board, but it shouldnt block general pandaboard [10:56:20] <JamieBennett> Is there any technical reason not to use uboot? [10:56:29] <JamieBennett> and keep things consistent? [10:56:31] <asac> they want to make things faster [10:56:36] <asac> imo thats fine, but its not high priority [10:56:36] <jserv--> JamieBennett, no. we just try to boot. [10:56:42] <JamieBennett> OK [10:56:42] <asac> jserv--: you can use our uboot [10:56:46] <asac> that works [10:56:50] <asac> (afaik) [10:56:58] <asac> otherwise you can use the uboot from pandroid [10:57:03] <jserv--> asac, not all the time. not sure why. [10:57:11] <asac> http://code.google.com/p/pandroid/downloads/detail?name=u-boot.bin [10:57:32] <asac> jserv--: you refer to the "panda boards sometimes shut down"? [10:57:34] <asac> bug? [10:57:34] <jserv--> how about keep this item? [10:57:41] <jserv--> asac, maybe. [10:58:02] <asac> jserv--: what is the serial number you have? [10:58:16] <asac> sure we have to keep it [10:58:21] <asac> pandaboard device tree was added [10:58:25] <asac> needs to be fixed to boot though [10:58:28] <asac> same for beagleboard [10:58:33] <asac> which requires install stuff in flash [10:58:48] <jserv--> ...A150 [10:58:54] <asac> yeah [10:59:08] <asac> we have the feeling that the even numbers have the bug that makes panda shutdown [10:59:24] <JamieBennett> https://wiki.linaro.org/KenWerner/Sandbox/708883 [10:59:40] <asac> jserv--: can you try our panda images? [10:59:47] <asac> headless that is [10:59:53] <jserv--> asac, ok, we will do that. [11:00:15] <asac> superbe [11:00:29] <JamieBennett> jserv--: there is a link on that page above to a test image [11:00:38] <JamieBennett> and how to install it [11:00:45] <jserv--> JamieBennett, thanks :) [11:01:05] <asac> ok lets move on [11:01:14] <jeremychang> asac, we will do that. [11:01:23] <patrik_> So we keep the action? [11:01:26] <asac> yes [11:01:41] <patrik_> [ACTION] "jeremychang to bring up device/linaro/pandaboard" [11:01:55] <patrik_> next action was... [11:02:08] <asac> jserv--: really try the MLO and u-boot.bin from here: http://code.google.com/p/pandroid/ [11:02:16] <patrik_> "patrikryd to setup toolchain/gcc toolchain/binutils and toolchain/manifest on git.linaro.org" [11:02:16] <asac> thats a "save" starting point [11:02:37] <jserv--> asac, cool. get it. [11:02:54] <patrik_> Sorry! Forgot that one. Will create the gits today. [11:03:09] <asac> kk ... question: is that the way we want to do things? [11:03:12] <jserv--> patrik_, https://wiki.linaro.org/Platform/Android/Toolchain # [11:03:13] <asac> jserv--: ? [11:03:26] <asac> do we want to host our toolchain/manifest etc. on git.linar.org [11:03:33] <asac> that gives you a toolchain checkout that works? [11:03:43] <asac> or will the toolchain-builder make this void [11:03:43] <asac> ? [11:03:49] <jserv--> patrik_, once you set up, we can point manifest to git.linaro.org [11:04:07] <asac> ok cool [11:04:25] <asac> do we want to put the our gcc in as gcc/gcc-linaro-4.5-2011.11 [11:04:27] <jserv--> the wiki page only described the "old" way. [11:04:32] <asac> or rather as gcc-linaro-4.5 ? [11:04:43] <asac> (and just update as we get new tarballs/snapshots)? [11:05:06] <jserv--> asac, I think obtaining from snapshot is fine. [11:05:07] <patrik_> I think we want to replace the binary toolchain in android with the a binary version of the lianro toolchain. [11:05:19] <asac> thats a separate topic [11:05:28] <asac> i dont think we want that ... what we want is to ship a binary tarball [11:05:29] <asac> or NDK [11:05:35] <asac> you can use TARGET_TOOLS_PREFIX [11:05:46] <asac> maybe at some point we commit stuff to prebuilt [11:06:01] <asac> what are the benefits of prebuilt? [11:06:08] <asac> i guess [11:06:20] <asac> 1. easier to ship a good working toolchain alongside android [11:06:29] <asac> 2. easier to make our toolchain the default for our build [11:07:18] <asac> [ACTION] patrik_ to explore with jserv-- and luse what options we have to ship toolchain [11:07:19] <patrik_> ok. [11:07:25] <jserv--> how about only replace prebuilt toolchain near Linaro 11.05 release? [11:07:44] <asac> yeah that sounds good. but i definitly want to also ship a binary toolchain tarball or NDK [11:07:51] <jserv--> asac, ok [11:07:53] <asac> so its consumable by folks that dont want to use our tree [11:07:55] <patrik_> Do we need all the tools gits at linaro.org? [11:08:15] <asac> patrik_: i think same as for platform we just need what we change for now [11:08:18] <patrik_> [ACTION] patrik_ to explore with jserv-- and luse what options we have to ship toolchain [11:08:24] <asac> so i assume for now we have manifest and toolchain/gcc [11:08:28] <luse> we will keep our own binutils? [11:08:46] <asac> luse: linaro does not release its own binutils [11:09:04] <asac> luse: it sends patches upstream and binutils releases often enough so it comes back quickly [11:09:04] <jserv--> asac, so, point to Google's binutils in manifest? [11:09:06] <luse> asas, but we create git repo for binutils ? [11:09:13] <asac> jserv--: thats what i would think [11:09:25] <jserv--> asac, ok, I agree. [11:09:28] <asac> luse: if we dont change it we can keep using googles [11:09:42] <luse> asac, ok [11:09:59] <jserv--> however, it implies potential bugs by Google engineers. [11:10:07] <asac> we are currently thinking if there is a benefit to mirror all platform/ (and similarly all toolchain/) in git.linaro.org ... but so far i am not convinced :) [11:10:10] <patrik_> ok. I'll create a tools manifest and the git for gcc to start with. [11:10:15] <asac> yeah [11:10:51] <asac> jserv--: thats true. and if we find a need that we need our own binutils, we should add that to git.linaro.org ... but not before i guess [11:11:15] <asac> jserv--: or if we want to move ahead and use a more recent one that would also be valid case [11:11:30] <jserv--> asac, ok, let's investigate it. [11:11:32] <asac> but my understanding is that more recent binutils causes the elf problem [11:11:34] <asac> ;) [11:11:41] <patrik_> The next actionf from last meeting was "luse to investigate how to build android toolchain all-static" [11:11:55] <luse> All-static toolchain can use LDFLAGS=-static to make components [11:12:10] <luse> but not recommend [11:12:17] <asac> i think that has to be qualified further ... it was about defining how we really want to ship our toolchain [11:12:23] <asac> in a way suitable for most [11:12:35] <asac> also considering all-static ;) [11:14:03] <jserv--> luse, or, how about adding "--enable-static" option in builder configure script? [11:14:03] <asac> luse: the wiki page should explore better how we want to distribute our toolchain [11:14:18] <asac> so that it can be used by majority of our potential users [11:15:02] <luse> asac, ok [11:15:29] <asac> [ACTION] luse to expand toolchain wiki page on distribution format options of toolchain and make a new wiki page on toolchain builder (or wrap in the toolchain main page) [11:15:29] <jserv--> next? [11:15:40] <patrik_> [ACTION] luse to expand toolchain wiki page on distribution format options of toolchain and make a new wiki page on toolchain builder (or wrap in the toolchain main page) [11:15:50] <asac> ok [11:15:51] <patrik_> we had one more... [11:15:52] <luse> jserv--, --enable-static will confuse the meaning disable-shared [11:15:52] <asac> lets move on [11:16:09] <jserv--> luse, ahh, yes, we need explicit one. [11:16:27] <patrik_> " asac jserv-- jeremychang and plars to sort out testing framework story for android" [11:16:34] <asac> thats ongoing [11:16:39] <asac> doesnt need to be carried forward [11:16:48] <patrik_> ok [11:16:51] <asac> jeremychang wrote a firs tpatch for making abrek agent based [11:16:57] <asac> check out great email from jserv-- on linaro-dev [11:17:10] <asac> i have one comment i want to drop for that... otherwise i seems we are on track ;) [11:18:11] <patrik_> [TOPIC] AOB [11:18:17] <JamieBennett> \o [11:18:21] <JamieBennett> So ... [11:18:41] <JamieBennett> In the rest of Platform we have weekly reports, detailing what the teams are doing, problems, risks e.t.c [11:18:49] <JamieBennett> I'd like to get that started with the Android team [11:18:56] <asac> yeah [11:19:06] <asac> we should start submitting weekly status reports [11:19:08] <JamieBennett> https://wiki.linaro.org/Status shows the rest of the reports [11:19:14] <JamieBennett> and the format [11:19:32] <asac> what i did in the past was to have everyone on team submit their weekly work/status update to the meeting wiki page [11:19:47] <JamieBennett> I'll email the Android team with instructions on what we would like to see and lets start producing the reports [11:19:51] <asac> and then have a TOPIC slot that reviewed those items to extract one or two most exciting news [11:20:18] <JamieBennett> asac: yes, that works or emailing me and I'll collate them/ask questions [11:20:22] <asac> JamieBennett: can you set up a weekly meeting template? and establish the process of submitting weekly status reports? [11:20:29] <JamieBennett> asac: will do [11:20:31] <asac> JamieBennett: i want to do it the way i said [11:20:35] <JamieBennett> OK [11:20:44] <asac> weekly reports ... first line of report should be "most exciting" news [11:20:54] <asac> then a slot in meeting to quickly review what was submitted [11:20:56] <jserv--> and keep progress posted. :) [11:21:07] <asac> and extract two lines of hot topic [11:21:07] <JamieBennett> [ACTION] JamieBennett to setup weekly reports for the team [11:21:12] <asac> thanks [11:21:14] <asac> jserv--: yeah [11:21:22] <patrik_> [ACTION] JamieBennett to setup weekly reports for the team [11:21:42] <asac> JamieBennett: so what i liked about other meetings was the ability to for engineers to put their reports on the wiki page _before_ the meeting themselves [11:21:46] <asac> that worked quite well [11:21:59] <JamieBennett> yes, then we can discuss them, indeed [11:22:06] <asac> nobody has to run after folks. if folks fail to submit they just are off the plate and officially did nothing in that week ;) [11:22:17] <JamieBennett> :) [11:22:27] <JamieBennett> I'll send instructions around today [11:22:28] <JamieBennett> \ [11:22:48] <asac> JamieBennett: i think besides jserv--, jeremychang, patrik_, there should be john stultz reporting here was well if possible and see if community folks like noritsuna et al want to report progress as well [11:23:12] <JamieBennett> OK [11:23:13] <jeremychang> same time I will try to see how 0xbench and abrek can work together. [11:23:27] <asac> yeah thats good ;) [11:23:38] <asac> would be cool to get that rocking [11:23:52] <asac> i had something for AOB ... but now i forgot it :( [11:24:06] <patrik_> I have a question... [11:24:29] <asac> shoot [11:24:39] <patrik_> we have to patches for review from Jeremychang. About ABD and ethernet. [11:24:56] <jserv--> s/ABD/adb/ ? [11:24:57] <asac> abd? [11:24:58] <patrik_> question to all. Do we need it? [11:24:59] <asac> ah [11:25:22] <asac> well ... we need network support or USB gadget support if we want to use the agent approach for LAVA [11:25:33] <jserv--> patrik_, if these patches are not landed, we have no efficient way to communicated with LAVA> [11:25:35] <asac> i think gadget support is even less reliable [11:25:48] <asac> and we shouldnt put that in as a requirement for lava [11:25:48] <jserv--> patrik_, you can say, it is a bug in adb. [11:25:52] <asac> what is adb about? [11:26:09] <asac> ethernet is just about bringin up networking, but what about adb? what is that patch about? [11:26:24] <jserv--> adb uses USB or TCP/IP. [11:26:36] <asac> yes [11:26:40] <jserv--> and the current implementation doesn't really listen to both. [11:26:44] <asac> ah [11:27:01] <jserv--> adb daemon only listens to the first one (USB). [11:27:19] <jserv--> and then, if your board lacks of USB gadget, it fails. [11:27:24] <asac> right [11:27:24] <asac> ok [11:27:25] <jeremychang> unless we have the property service.adb.tcp.port set [11:27:35] <asac> so thats just a setting thing, right? [11:27:41] <jeremychang> adb can listen on ethernet. [11:27:41] <jserv--> asac, exactly. [11:27:55] <asac> ok i think its ok to take a patch for settings [11:28:02] <asac> in the long run we have to review what kind of settings we take patches for [11:28:07] <asac> and how we can make this rather a build option [11:28:15] <asac> (which might be suitable for upstrema submission) [11:28:36] <jserv--> asac, ok, fixing in upstream is great. [11:28:46] <patrik_> there is also a patch on adb.c [11:28:55] <asac> whats that patch about? [11:29:03] * asac tries to find the mail [11:30:09] <jeremychang> patrik_, the patch is for making adbd listen on both ethernet and usb if possible. [11:30:11] <patrik_> subject of mail was "[linaro-android][PATCH] system/core: Make adbd listen on network and usb gadget concurrently" [11:30:48] <asac> http://lists.linaro.org/pipermail/linaro-dev/2011-February/002526.html [11:30:54] <asac> i dont see a patch to adb.c [11:30:59] <patrik_> jeremychang: Have you tried to get it into aosp? [11:31:10] <jeremychang> Sorry I have no much comment in the mail. [11:31:38] <jserv--> patrik_, since OTG is not fully supported in all dev boards, we need TCP/IP for adb. Thus, jeremychang attempted to make it really work for both cases. [11:31:47] <asac> http://lists.linaro.org/pipermail/linaro-dev/2011-February/002541.html [11:31:49] <asac> ok found it [11:31:57] <jeremychang> patrik_, not yet but I will do this. [11:32:38] <asac> jeremychang: we have problem to take patches from you guys until contract is signed because we need copyright assignment in order to forward stuff [11:32:48] <patrik_> Sounds useful. [11:33:18] <jserv--> asac, I see. [11:33:25] <patrik_> Is there a legal problem? [11:33:26] <asac> i am happy to take it if you are willing to resubmit with @linaro.org once you have that address [11:33:32] <asac> patrik_: no, just slow [11:34:04] <jserv--> any other question? [11:34:15] <asac> are you willing to resubmit ;)? [11:34:26] <jeremychang> I am using @linaro.org already. [11:34:36] <asac> jeremychang: ok cool. can you resend that patch then? [11:35:02] <jeremychang> asac, I will resubmit. [11:35:04] <asac> the config changes are not important ... just code changes scare me [11:35:05] <asac> thanks [11:35:14] <asac> ok thats all [11:35:18] <asac> i had something, but it fell off my plate [11:35:20] <asac> o [11:35:21] <asac> h [11:35:34] * jserv-- waits for dinner. [11:35:34] <asac> so yeah, sorry for being late here. next time dont wait for me (though i dont expect to be late) [11:35:50] * patrik_ needs lunch [11:35:58] <asac> jserv--: one answer: is mount ext3 .... a valid way to do it? [11:35:58] <asac> ;) [11:36:04] <asac> then you can go [11:36:42] <jserv--> asac, I'll provide the sample for you later. [11:36:46] <asac> kk [11:36:47] <patrik_> is it for androids init.rc? [11:36:48] <asac> ok thanks [11:36:50] <asac> yes [11:36:54] <asac> fixing the mounts ;) [11:37:14] <asac> i think we should start with ext3 for now ... and maybe think about btrfs later or so [11:37:27] <patrik_> I usually have ext3. [11:37:33] <jserv--> good idea. [11:37:40] <asac> jserv--: ok no need. i can sort this out with patrik_ ;) [11:37:44] <asac> enjoy dinner, lunch etc. [11:37:44] <patrik_> Gingerbread came with ext4 (or yaffs2) [11:37:57] <asac> thanks [11:38:01] <asac> please #endmeeting ;) [11:38:06] <asac> if you forget that noone can start a new meeting [11:38:08] <asac> :-P [11:38:09] <jserv--> ok, see you. [11:38:13] <patrik_> #endmeeting Meeting ended.
Platform/Android/Meetings/2011-02-16 (last modified 2011-02-17 22:12:23)