On Wed, 2023-06-21 at 09:10 +0200, Javier Martinez Canillas wrote: > !! External Email > > [adding Zack Rusin again who seems to have fallen from the Cc list] > > Albert Esteve <aesteve@xxxxxxxxxx> writes: > > On 6/10/22 10:59, Daniel Vetter wrote: > > > On Fri, Jun 10, 2022 at 10:41:05AM +0200, Daniel Vetter wrote: > > [...] > > > > > - third issue: These special virtual display properties arent documented. > > > > Aside from hotspot there's also suggested X/Y and maybe other stuff. I > > > > have no idea what suggested X/Y does and what userspace should do with > > > > it. I think we need a new section for virtualized drivers which: > > > > - documents all the properties involved > > > > - documents the new cap for enabling virtual cursor planes > > > > - documents some of the key flows that compositors should implement for > > > > best experience > > > > - documents how exactly the user experience will degrade if compositors > > > > pretend it's just a normal kms driver (maybe put that into each of the > > > > special flows that a compositor ideally supports) > > > > - whatever other comments and gaps I've missed, I'm sure > > > > Simon/Pekka/others will chime in once the patch exists. > > What is missing for these patches to land? If I understood correctly is > just these properties documentation that Sima asked above ? > > > > Great bonus would be an igt which demonstrates these flows. With the > > > interactive debug breakpoints to wait for resizing or whatever this should > > > be all neatly possible. > > > -Daniel > > > > Hi all, > > > > We have been testing the v2 of this patch and it works correctly for us. > > > > First we tested with a patched Mutter, the mouse clicks were correct, > > and behavior was as expected. > > > > But I've also added an IGT test as suggested above: > > https://gitlab.freedesktop.org/aesteve/igt-gpu-tools/-/compare/master...modeset-cursor-hotspot-test?from_project_id=1274 > > > > I could post the IGT patch upstream once Zack's patches land. > > > > Zack, are you planning to re-spin the series soon? Otherwise Albert could > continue working on that if you prefer. Besides mutter I wanted to patch kwin as well, but I haven't been able to find the time to patch it as well. I can respin with discussed changes over the weekend if we're ok with getting this in without kde support from the start. z