On Mon, May 09, 2011 at 08:37:55AM -0700, Andy Ross wrote:

How would that work?  If it's an extra sysfs node, it would require an
extra file descriptor in the reader.  And if it's an event inside the
input stream, that would require the reader implement handling to throttle
the polling rate.  The point here was to avoid having to poll the
(noisy) accelerometer stream just for gross orientation changes.

Yes, so the accelerometer driver should (in-kernel) know that a coarse 
orientation event has occured and then send an appropriate uevent to 
userspace indiciating that it has new data.

http://build.meego.com/package/files?package=sensorfw-pegatron&project=Trunk

Yeah, that's absolutely dreadful. Don't do that.

Does it really need to be "fixed"?  If this were a whole class of
hardware, then I guess I'd agree.  But right now it's one quirky
device.

I'm going to NAK anything that reports "Coarse orientation change" to 
userspace without providing any context.

-- 
Matthew Garrett | mjg59@srcf.ucam.org