On Mon, 01 Jun 2026 00:39:43 +0200,
Kirill A. Korinsky <kirill@korins.ky> wrote:
> 
> 
> I think this hidms part is broader than the stated touchscreen fix.
> 
> The new released flag is true for every absolute HID device whose buttons
> transition from nonzero to zero; it is not tied to the synthetic zeroed
> report generated by the polled ihidev path, if I not mistaken.
> 
> Consequently, a real release report from an interrupt driven absolute
> device, with valid final X/Y coordinates, will have those coordinates
> suppressed, and the button release will be delivered at the previous stored
> position.
> 
> Not sure does it bring any issue, or just impossible case.
> 

From another hand, my daily driver had rare and anoyed bug when touchpad is
a kind of "stuck" and ignores clicks. Usually to "fix" that I need to use
multitouch click.

Since I start run your diff on this machine, I haven't encountered it.

Probably you had fixed it as side effect.

So, I'd like to OK it and suggest to go it in and see reports in bugs@ in
the case if it brokes somehting.

-- 
wbr, Kirill