On Mon, Oct 29, 2018 at 6:18 AM Jiri Kosina <jikos@kernel.org> wrote:
>
> Benjamin indicated that Peter probably has found the issue in the code
> (failure to properly reset on direction change) that might be causing
> this.

So honestly, once I looked at that hid_scroll_counter_handle_scroll()
function, that's the first thing I tried - get rid of the "half-way
threshold" thing, and reset on direction changes.

It fixes the instability, and I don't see the "back-and-forth"
movements and I don't get the "move the mouse and it generates mouse
wheel events" any more.

It basically makes the wheel _work_ for me.

I'm not entirely convinced it's as good as it used to be, though. It
still feels like it might be a bit over-sensitive, but that may be
because now I'm just looking for it..

Patch I'm using attached. I'm inclined to just commit it, but if
somebody has a better idea, I can test alternatives too.

                    Linus