[android-porting] Re: Sensors

2009-01-15 Thread Ken Schultz
Mathias, I am looking to use the cupcake version of sensors.h for the sensor HAL layer. >In "theory", all you need to do is implement a sensor HAL module >properly (as documented in sensors.h. pay extreme attention to the >definition of the axis and units). Unfortunately, in the current git >rep

[android-porting] sensors HAL/single input device for multiple kernel drivers

2009-01-26 Thread Ken Schultz
Hi all, I am in the process of writing a sensors HAL based on the new cupcake module structure. One restriction of the sensors HAL is that there can only be a single input device used for all sensors. Can anyone provide any assistance and/or example on how this might be done with multiple kerne

[android-porting] sensors HAL limitation

2009-01-29 Thread Ken Schultz
It seems the the sensors HAL can only reference a single /dev/input/ event#, but if you have multiple sensor drivers how would you have them all reference the same event node? Is this even possible? I was under the impression each input driver should have its own event node. --~--~-~--~-

[android-porting] Re: sensors HAL limitation

2009-01-29 Thread Ken Schultz
event > node. Drivers can use the same strategy and push events into the same > event node. > > Mathias > > On Thu, Jan 29, 2009 at 4:17 PM, Ken Schultz wrote: > > > It seems the the sensors HAL can only reference a single /dev/input/ > > event#, but if you have m

[android-porting] Re: sensors HAL limitation

2009-01-29 Thread Ken Schultz
Correction. The other drivers would register to the input driver but never write to their own event node. Only the main driver would write to its own event node. On Jan 29, 6:43 pm, Ken Schultz wrote: > So essentially the other drivers would register to the input driver > but never wr

[android-porting] Re: sensors HAL limitation

2009-01-29 Thread Ken Schultz
imple here? On Jan 29, 6:45 pm, Ken Schultz wrote: > Correction.  The other drivers would register to the input driver but > never write to their own event node.  Only the main driver would write > to its own event node. > > On Jan 29, 6:43 pm, Ken Schultz wrote: > > > So

[android-porting] Re: sensors HAL limitation

2009-01-30 Thread Ken Schultz
Any details/example how this is done? Ken On Jan 29, 10:07 pm, pramod gurav wrote: > On Fri, Jan 30, 2009 at 6:46 AM, Ken Schultz wrote: > > > The problem I'm running into is how to do this in kernel space.  I > > would suspect you would need to share the input_dev str

[android-porting] best touchscreen resolution (pixels/mm)

2009-02-11 Thread Ken Schultz
Does anyone know the best touchscreen resolution for Android? Most likely there would be a sweet spot for this. The higher the resolution the more data which could cause a bottleneck for data flow and the low the resolution the worse the accuracy is. I would imagine this to be in the units pixe

[android-porting] sensors HAL

2009-03-05 Thread Ken Schultz
Is there a recommended way to pass information between a sensors_control_context_t struct and a sensors_data_context_t struct? (i.e. active_sensors) For devices without a hardware eCompass there needs to be a way in the polling function to determine if the sensor is activated, so it can inject da