Op 04-02-10 09:24, Samu Onkalo schreef: > Lis3 accelerometer chip family changes for power management, > click and threshold event handling. > > Patch set adds interrupt handlers for click/tap events and threshold > based events. Actual configuration which events are enabled > is done via platform data. All the features cannot be used in parallel. > Interrupts are implemented only for 8 bit device, since I'm not familiar > with other devices and I don't have suitable testing environment. > > Changes: > > lis3: Add missing constants for 8bit device > This is quite clear. Some click feature related register definitions > were missing. > > lis3: Separate configuration function for 8 bit device > Move platformdata based configurations for 8 bit device to > separate function to keep common part little bit more readable. > > lis3: Introduce platform data for second ff / wu unit > 8 bit device has two freefall / wakeup detection blocks. Add possibility > to configure also the second unit. Change hipass filter configuration > to platform data. Change is compatible with existing platform data. > > lis3: Power control for the chip > This kind of feature has been in the driver earlier. It was removed > because saving was so small in laptop environment. However, in smaller > devices, even a small saving need to be implemented. When driver detects > that no-one is really interested about the acceleration, chip is powered down. > Input device, freefall device and sysfs are controlling this. By default, > chip is powered on to keep functionality similar to current implementation. > > lis3: Add skeletons for interrupt handlers > Interrupt handlers are added in two patches to keep changes cleaner. > This first patch adds two dummy threaded interrupt handlers for 8 bit device. > > lis3: Interrupt handlers for 8bit wakeup and click events > This patch adds content to dummy handlers. Depending on the chip configuration, > either click or ff/wu handling is called. For click event, BTN input event is > sent separately for each axes. For threshold event, coordinates are updated > immediatelly to input device. This allows input device to be used either in > polled mode and / or interrupt driven mode. Polling can stopped from userspace > by via input device sysfs. > > Patch set applies to 2.6.33-RC6 tree. > Tested with 2.6.32 environment for omap3 Hello, Sorry, I haven't had time to fully review the patches. Nevertheless, I've tested them on my HP laptop with a 12bit device, and can confirm it's all fine. The only hiccup is in patch 6: it declares the joystick with 3 buttons, although on this hardware it's impossible to have this feature. So, it would be nice if input_set_capability() could be conditionned. I also fully agree with the comments from others that it would be better if it could use the special runtime power-management framework from Rafael. BTW, what is exactly the usecase for the "active" sysfs attribute? To me, it looks like it just prevents userspace from using the position attribute, while the joystick interface is still accessible, so it is not very effective! So, for the first four patches, here is my Acked-by: Éric Piel <eric.piel@xxxxxxxxxxxxxxxx> Cheers, Eric _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
