=== Highlights === * LWN covered my Android Upstreaming Summary here: https://lwn.net/Articles/514901/
* Implemented byte-range management for volatile ranges.
* After talking with Thomas last week, I took a first pass at implementing refined-jiffies, which should allow CLOCK_TICK_RATE to be managed dynamically rather then statically. This helps the unified zImage work.
* Generated a patch implementing Tixy's suggested fix for the evdev issue he was seeing. Sent it to Rafael for review and then submitted it to lkml. Its queued for 3.7
* Got feedback from Mozilla developers, clarifying what their needs are with respect to the volatile ranges patch
* Got weekly Android upstreaming email going again
* Replied to questions related to my Android Upstreaming Summary
* Followed up on time patches that should be backported to -stable. Generated patch queues for each of the -stable trees.
* Did some further work prepping time changes for 3.7
* Investigated time regression in 3.6
* Replied to clock TAI question on the ptp list.
=== Plans === * Make first pass at sending SIGBUS on accessing purged volatile pages
* Finish testing of refined jiffies and send patches out
* Finish testing -stable backports, and send patches out
=== Issues === * NA
On Fri, Sep 7, 2012 at 11:19 PM, John Stultz johnstul@us.ibm.com wrote:
- Got feedback from Mozilla developers, clarifying what their needs are with
respect to the volatile ranges patch
Maybe we should poke developers of e.g. the GNOME application "Shotwell" or KDE's "Photobook"? IIRC the place where I first saw this type of volatile range usecases was in that type of app that needs to keep a gazillion of thumbnails around.
Windows even consider the operation so costly that it saves the thumbnails to files named "Thumb.db" in the image dirs for later reuse if the file is untouched, I bet some of these apps do that too.
Yours, Linus Walleij
linaro-kernel@lists.linaro.org