Hi Roderick,
sorry for the late response, I've been busy in these days. I'll reply
to both emails in this one.
Thanks for testing. I added a similar patch to my hid-playstation tree
this morning (hid-sony will go away soon).
Ok perfect! Does hid-sony will continue to support old Dualshock
controllers or they will be all merged into hid-playstation?
I'm not entirely happy with the patches yet and need to do some
thinking. The issue is the value range, which is not correct. For the
accelerometer the numerator and denominator need to be 1 to match the
proper range. (It just happens that the scaling factors are the way
they are I think.)
The uncalibrated gyroscope values are 16-bit signed, but the hardware
limits are +/- 2000 degree per second. To have enough precision during
calibration the value range is extended (resolution per axis is set to
1024, so 1024 corresponds to 1 degree per second).
Can you tell me a bit more about what is the expected calibrated
output for both gyro and accelerometer?
Like unit of measurement (I hope deg/s and m/s^2) and data type?
Is "1024 -> 1dg/s" a mapping used only while calibrating or is the
expected conversion during normal usage?
When I heavily shake a DS4 in evtest, I can easily reach values like
15 million or more. However, the maximum value we have set on the
evdev node to about 2M. In other words there is something wrong. I
need to look closely at other internal code.
Definitely seems there is an issue.
Can you suggest me a tool to test the gyro/accel using linux?
Thanks,
Alain
