On Wed, Oct 31, 2018 at 8:03 AM Bastien Nocera <hadess@xxxxxxxxxx> wrote: > > Hey, > > On Wed, 2018-10-31 at 09:02 +0500, Mikhail Gavrilov wrote: > > With kernel 4.19 all time dualshock 4 has phantom triggering left > > analog stick. > > This is appear in the fact that in any game an up / forward movement > > may starts randomly any time and not ended even if I release the left > > analog stick. > > > > Of course I checked this dualshock with PS4, and there was no such > > problem. > > So I’m completely sure that this is not a hardware problem. > > Are you sure that it's not simply a difference in hysterisis? I would > expect most software using those values to ignore "close to center" > jitters. You might also want to make your script use the actual min/max > values for those analogue sticks, I don't think they're actually > between -32767 and 32767 as you use in the script (although I don't > have a joypad to verify this right now). The DS3 / DS4 axes are, so 0 to 255 with a center around 127/128. Scaling by these factors doesn't make sense, xbox controllers and others are 16-bit. I would need to check what the dead band is for our controllers, but probably at least 4 around the center. > > $ uname -r > > 4.19.0-1.fc30.x86_64 > > You say it's a regression, but in which kernel version was this > working? > > Cheers I would be curious as well which kernel version there would be a potential regression. Nothing in this area changed since our refactor around 4.12. For the sticks we more or less pass values from the hardware to userspace, there is no scaling or other operations on them. Thanks, Roderick
