Email meeting log

[Administrative stuff]========================
* Late again. Apologies.

* Anton: Still waiting on some time options from you that would be good for a phone meeting.

* Anton: I went ahead and updated  your blueprints to move work-items over to the work-item area.

* Mounir:  Launchpad is telling me 12.04 isn't a valid work-item target. I'm having to put items to be done after 12.03 as targetting "backlog". Can you look into whats needed here?

[Lowmemory killer]===============================

* Anton: Thanks for updating the blueprint with your plans!

* Anton: I started looking at the ulmkd code, and it looks fairly clean. One small issue I noticed that you might want to address soon: You're not mlocking the daemon, which means we may have to fault in pages on notification, causing additional memory overhead. Some additional comments would be also helpful.    Also, any sense of how costly scanning proc to find the next task to kill is?

* Anton: You mentioned you had some "cgroups/vmevents bugfixes and enhacements".  Just to better understand them, are these required for or block the ulmkd work, or just good fixes useful to the community?

[Interactive cpu-freq gov]=======================

* No updates here. Might need to push this to the backlog


* Got some feedback from my iteration last week. Quickly merged in any changes I could and resent. So far no feedback on the lastest revision.

* Some IBM folks are already trying to integrate my range-tree code into the hugetlb code

* Need to research prio_tree code to see if it would be more appropriate, as well as implementing a  version using radix tree tags

[monotonic evdev]=================================

 * No updates. Not yet merged in 3.4, but hopefully will land next week.

[alarmtimer mending]==============================

* Alarm Driver landed upstream for 3.4 via staging.

* Android folks are still using my cleanup patches. Will likely push to the staging-next tree for 3.5


* In the current android-3.3 tree, Arve has significantly rewritten the wakelock implementation. Its much more compact and seems likely to be mergeable  for  staging.  I'll likely reach out to him in a week or so to see how the implementation is doing and if he would be ok pushing it into staging. I also want to get some time to test with it.


* I sent out some small changes to the script

* I spent some time to setup an initial tree for utilizing config framgment in the linaro tracking tree. Will hopefully announce it shortly and try to get it integrated for the 12.04 tree.

* Trying to reproduce a timekeeping issue that's cropping up occasionally on fedora systems

WorkingGroups/KernelArchived/AndroidUpstreamMeetings/2012-03-23 (last modified 2013-01-14 19:35:18)