Sort of a mix of things from this and the previous week.
=== Highlights === * Chased down a timekeeping bug in mult selection, pushed fix to Ingo
* Pinged Michael Marek on status of merge_config.sh fragment tool. No response yet. Will send patch to akpm next week if I don't hear anything.
* Worked on madvise VOLATILE ashmem alternative. Got first draft of the patch that seems to be working. Pinged Robert Love at Google to see if there were any unit tests available (sadly there aren't). Gotten some internal review from VM folks and need to address some style issues, then will hopefully sent to Robert for input and possibly publicly later.
* Pinged Rafael on his wakelock related plans, after the Kernel Summit discussions. At his request I split out wakelock patches (credits to Andy Green for his Androidization tracking tree making this very easy) into an independent branch and submitted.
* Generated background slide for Android Upstreaming work.
* Enabled eCryptFS in Linaro+Android kernel defconfigs and released an the initial 11.11 linaro-android kernel.
* Tried to chase down issues relating to 11.10 LEB release (from the linaro web page) being based on a 3.0 kernel, despite 3.1 branches being available. Seems there was a bit of miscommunication last cycle. Pressed to make sure this gets resolved early for 11.11.
* Google got the Android kernel/common.git tree re-published. Latest branch is still 3.0 based, but there are some new patches there.
* Took some time to work on sched_clock() mult overflows that are catching some systems at ~208 days of uptime.
=== Plans === * Work on additional madvise VOLATILE test cases, submit patch to rlove and possibly lkml.
* Submit merge_config.sh to akpm
* Submit a number of timekeeping cleanups for 3.3 to Thomas/Ingo
* Try to submit sched_clock rework for discussion.
* Tag 11.11. final linaro+android kernel
=== Issues === * None