EGL_MESA_screen_surface version 9

David Reveman davidr at novell.com
Sat May 21 05:44:38 PDT 2005


On Fri, 2005-05-20 at 20:12 -0400, Jon Smirl wrote:
> On 5/20/05, Adam Jackson <ajax at nwnk.net> wrote:
> > On Friday 20 May 2005 18:58, Jon Smirl wrote:
> > > Hardware cursor support is an integral part of a screen surface. It
> > > should probably be included.  Hardware cursor support is probably more
> > > important to embedded devices too.
> > 
> > Cursors can be implemented either as normal textured polys or as hardware
> > glyphs.  In the first case it's just more objects in the scene graph.  In the
> > second it's logically layered on top of the existance of a screen surface.
> > 
> > I don't think cursors belong in MESA_screen_surface.
> 
> I'm not looking to address software cursors, right now I'm only
> interested in the hardware one.
> 
> If this is going to be a separate spec we need to start it immediately
> since the current Xgl implementation assumes the existence of a
> hardware cursor. I can make Xgl directly access the fbdev hardware
> cursor support but I would much rather use an EGL version.

Xgl doesn't assume the existence of a hardware cursor. Xgl works fine
without a hardware cursor. Xglx is currently using a software cursor
when the X server it's running on top of doesn't support ARGB cursors
(XRenderCreateCursor).

> 
> How is EGL going to get a bitmap into the cursor? Use OpenGL command
> to create the surface and then pass it into the cursor create function
> which will copy it? Since I already have the working fbdev
> implementation it is pretty easy to implement cursor support as soon
> as we come up with an API.
> 

-David




More information about the dri-egl mailing list