[PATCH 0/6] drm: Add mouse cursor hotspot support to atomic KMS

[Daniel Vetter]

On Fri, Jun 10, 2022 at 08:54:03AM +0000, Simon Ser wrote:
> I agree with what others have replied, just adding a few more details.
> 
> On Thursday, June 9th, 2022 at 21:39, Zack Rusin <zackr at vmware.com> wrote:
> 
> > virtualized drivers send drm_kms_helper_hotplug_event which sends a HOTPLUG=1
> > event with a changed preferred width/height
> 
> (Note: and the "hotplug_mode_update" property is set to 1.)
> 
> > suggested_x and suggested_y properties
> 
> These come with their own set of issues. They are poorly defined, but it seems
> like they describe a position in physical pixel coordinates. Compositors don't
> use physical pixel coordinates to organize their outputs, instead they use
> logical coordinates. For instance, a HiDPI 4k screen with a scale of 2 will
> take up 1920x1080 logical pixels. There is no way to convert physical pixel
> coordinates to logical pixel coordinates in the general case, because there's
> no "global scale factor". So suggested_x/y are incompatible with the way
> compositors work.

I dropped a request for a proper doc section that explains all the
virtualized kms driver stuff. I think we should also put in a
"limitations" part there and just spec that any kind of scaling is a no-go
on these (and that drivers better validate this is the case).
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch