On Sun, Jan 04, 2026 at 05:42:09PM +0100, Landry Breuil wrote:
> Le Sun, Jan 04, 2026 at 03:02:45PM +0100, Matthieu Herrb a écrit :
> > On Sun, Jan 04, 2026 at 02:51:07PM +0100, Landry Breuil wrote:
> > > Le Sun, Jan 04, 2026 at 01:54:10PM +0100, Matthieu Herrb a écrit :
> > > > On Tue, Dec 09, 2025 at 10:31:55AM +0100, Landry Breuil wrote:
> > > the only remaining issue is the missing graphical rendering of the
> > > screens in the left pane, but that works fine once mesa is built with
> > > wayland support :)
> > 
> > At least under sway, the pointer motion becomes sluggish while wdisplays is
> > running, on my 2 testing machines. I couldn't find the reason so far,
> > but it may be linked to the attempts at rendering a miniature of the
> > screen.
> 
> but was it with my hack that links mesa with wayland ? otherwise i'm
> pretty sure there's no miniature rendering attempt (but i havent looked
> at the actual code..)
> 
no. so it's something else. looking at ktrace output, there seems to
be a lot of shm_open() activity, which we know is bad on OpenBSD because
it touches the filesystem at each access.

I did infer that those where reading windows contents from the
compositor, but it can be something else.

-- 
Matthieu Herrb

Reply via email to