Android Kernel Sync w/ OESF team

Agenda

  • Introductions and background
    • Current work with Android Kernel
  • Goals for Linaro Android images
    • Stable image for centralized base for SoC vendors
    • Experimental image for demonstrating Android upstreaming effort
  • 2.6.37 Android + Linaro kernel testing
  • Work items to explore
    • Investigate what is required to implement Binder API using DBUS
    • Look into minimal set of Android kernel patches required to get Android environment running stable.
    • Other suggestions?

IRC Log

* jstultz_vm has changed the topic to: Agenda: https://wiki.linaro.org/Platform/Android/2011-02-04
* noritsuna (~noritsuna@8.98.151.202.cc9.ne.jp) has joined #linaro.android
<jstultz_vm> noritsuna:  hello!
<noritsuna> Hello. Sorry, I'm late.
<jstultz_vm> noritsuna:  Don't worry about it! :)
<jstultz_vm> We'll give a few more minutes to give a chance for others to join.
<noritsuna> I got message from Mr.Watanabe. He is having emergency meeting now;-( He can't join this meeting.
<noritsuna> But I got his work's report.
<jstultz_vm> noritsuna:  ok. are you expecting anyone else?
<jstultz_vm> noritsuna:  if not then we might as well go ahead.
<noritsuna> I think to explan this message to you.
<jstultz_vm> So I posted the agenda in the channel topic header.
<jstultz_vm> Just wanted to cover some introductions, and see what you have been working on.
<jstultz_vm> We already met at the Dallas summit, but for the sake of others reading the IRC logs later, I'll introduce myself again.
<jstultz_vm> I'm John Stultz, and I'm working on Android kernel upstreaming. I've done a fair amount of work in the kernel related to timekeeping, realtime and x86 system enablement
<jstultz_vm> I'm fairly new to Android, but picking it up (with help from noritsuna!).
<jstultz_vm> noritsuna:  Do you want to say a few words about your background?
<noritsuna> OK. I'm Noritsuna. I'm OESF's coordinator. OESF deliver Embedded Android's distribution to Emebedded companies.
<noritsuna> I
<noritsuna> My background is SIP(simple initiation protocl) engineer.
<noritsuna> I developed NTT's softphone.
<noritsuna> NTT is the most famous career in Japan.
<jstultz_vm> Oh neat! I use a sip based voip service. Its very useful!
<noritsuna> Mr.Watanabe & Mr.An are Android's SIer. (This company name is ISB)
<noritsuna> They developed NEC's device(http://av.watch.impress.co.jp/docs/news/20101125_409295.html).
<noritsuna> Sorry, It is Japanese article;-( 
<jstultz_vm> noritsuna:  Thanks for covering their introductions too.  Next up I wanted to cover some of the discussions I've had with Alexander and Patrik about some of the Linaro Android platform goals.
<jstultz_vm> So basically, there are two goals for the Linaro Android Platform:
<noritsuna> OK.
<jstultz_vm> 1) Be a stable image where a number of Linaro partners can centralize their development on Android, gaining features resulting from the Linaro kernel and toolchain efforts.
<jstultz_vm> 2) Provide an experimental image where we can demonstrate the effort of the Android kernel upstreaming/consolidation effort
<jstultz_vm> The first is the higher priority initially, but as the upstreaming effort moves forward, the second will have lots of interesting development work items.
<jstultz_vm> Currently, Patrik is working on getting the images for #1 created, and will be including my 2.6.37 Android + Linaro kernel.
<jstultz_vm> At some later time we'll start looking into the second, which may include changes to Android's userland to adapt it to utilize possible alternative features pushed into mainline.
<jstultz_vm> So for instance, if my posix alarm-timers work goes into mainline, we may want to look at adapting the Android userland to make use of that interface instead of the /dev/alarm device.
<jstultz_vm> Is that division of goals make sense to you?
<jstultz_vm> Does that division, not "is that", sorry.
<jstultz_vm> noritsuna:  So do you want to talk about some of your current Android kernel work?
<noritsuna> We think to develop Power management interface for Android-kernel. Because Android has some power state.
<jstultz_vm> noritsuna:  Is this connected to the wakelocks and early suspend? Or is this something similar to power-top in the regular Linux environment?
<noritsuna> It is the wakelocks and early suspend.
<jstultz_vm> noritsuna:  Interesting!
<noritsuna> however we don't have basically android-kernel. we are making basically android-kernel for Android2.3. We make it by your repository base.
<jstultz_vm> noritsuna:  So, have you had a chance to look at or test my Android+Linaro git tree?
<noritsuna> Yes. We can run it on Android2.3. But it have many error. we are fixing now.
<jstultz_vm> noritsuna:  What sort of errors? I'd be very interested to hear, as I'd like to fix them as well.
<noritsuna> We will patch to it. ex. http://pylone.jp/blog/virtual_battery .
<noritsuna> We patched page flipping & .DSS.
<jstultz_vm> noritsuna:  Ok. I think the page flipping was mentioned on your "how-to-porting" wiki page, right?
<jstultz_vm> https://github.com/OESF/Linaro-Android_LinaroSprint2011Q1/wiki/How-to-porting
<jstultz_vm> The tree patches being:
<jstultz_vm> s/The tree/The three
<noritsuna> Yes.
<jstultz_vm>  ARM: Adding regulator supply for vdds_sdi.
<jstultz_vm> ARM: OMAP: Beagle: support twl gpio differences on xM
<jstultz_vm> ARM: OMAP3: beagle: use GPIO2 on the xM A3 to turn DVI on
<jstultz_vm> And the page flipping was: B virtual size fixes: android needs vysize=ysize * 2
<jstultz_vm> Ok. I didn't seem to have any problems in my testing, so I assumed those were not needed.  I'll cherry pick those into my tree and do some additional testing.
<jstultz_vm> Are there other issues that still need to be resolved after those patches are applied?
<noritsuna> Nothing.
<jstultz_vm> Ok, great.
<jstultz_vm> noritsuna:  Do you want to submit the virutal battery driver to our tree as a patch?
<noritsuna> Yes. we want to sumit permission.
<jstultz_vm> noritsuna:  That would be great. If you send it as a patch with a proper "Signed-off-by" line, I'll merge it into our tree.
<jstultz_vm> noritsuna:  the other items I wanted to cover were some of the consolidation/upstreaming work I'm thinking about.
<noritsuna> OK! 
<jstultz_vm> noritsuna:  Alexander said you were interested in some of the consolidation work, so I wanted to share and see if you had any interest in helping with those items.
<jstultz_vm> So one item would be investigating the Binder API to see what would be required in order to implement the same API using dbus.
<jstultz_vm> The other item I'm thinking about is trying to find the minimal set of patches to the linux kernel that allows the Android environment to function properly (although not optimized).
<noritsuna> Yes! we think to make springboard for discussion of Android-kernel(about Power management, how to develop) as next job.
<jstultz_vm> In my initial testing, using only the Binder + Ashmem patches lets the Android GUI boot ok. Although some application crashes occur, so its likely some other patches are needed too.
<jstultz_vm> Once we have that minimal set, we can focus on getting that functionality up stream so Android can run on an unmodified Linux kernel (although some features like early suspend might not work).
<jstultz_vm> noritsuna:  what are you thinking about with regards to power management ?
<noritsuna> We think to standardize power management function(interface) of each SoC's Android-kernel.
<jstultz_vm> noritsuna:  what would this interface look like?
<noritsuna> Sorry, we have no idea now;-(
<jstultz_vm> noritsuna:  No problem, I'm just interested in hearing about it.
<noritsuna> We will analize it next time.
<jstultz_vm> noritsuna:  So since we're getting to the last few minutes of the meeting. I wanted to ask if there was anything else you wanted to talk about, or if there was anything you need from me for next week?
<noritsuna> We will start to work at full capacity next week.  Becuase Mr.Sola(http://blog.sola-dolphin-1.net/) will join this group next week. (He caught flu this week. He could't join us this week;-()
<jstultz_vm> noritsuna:  I'll also ask AmitK who is part of our Linaro power-management team to see what his thoughts are on standardized power management interfaces.
<noritsuna> OK! When I finish to make  springboard for discussion of Android-kernel, we will want to discuss with AmitK:-)
<jstultz_vm> noritsuna:  Sure.
<noritsuna> And if we have some problems, we want to do meeting with you:-)
<jstultz_vm> noritsuna:  Yes. The next meeting will be on Weds.
<noritsuna> Yes! Thanks!
<jstultz_vm> noritsuna:  And feel free to email me if you have any questions before then.
<jstultz_vm> noritsuna:  so just to cover action items:
<jstultz_vm> I will try to cherry pick the patches you pointed out in the wiki, test them and post them to the git repo.
<jstultz_vm> I will also contact Amit to talk about power managment interfaces.
<jstultz_vm> And if you could, please send the virtual battery driver as a patch and I will try to merge it into our git tree.
<jstultz_vm> And that's all.
<jstultz_vm> noritsuna:  It was great talking with you! See you again on Weds.

Platform/Android/Meetings/2011-02-04 (last modified 2011-02-16 14:31:06)