https://bugzilla.gnome.org/show_bug.cgi?id=744932
--- Comment #131 from Jonas Ådahl <[email protected]> ---
(In reply to Owen Taylor from comment #118)
> (In reply to Jonas Ådahl from comment #117)
>
> > The new version of the big-chunk-patch does this now. It could even more
> > refactoring, but in summary, drmModeSetCursor2() is now called once per
> > frame
> > only, and updating the cursor will queue a redraw which will cause calling
> > of
> > the whole stage to be redrawn (well, not really, because there may be no
> > damage) with all the drm ioctls with it.
>
> Have you tested in detail what happens when only moving the HW cursor? Do we
> *actually* not paint anything? Do we still page flip?
The side effect of doing it this way was that I needed to do a 1x1 pixel
damage. So given this, as per IRC discussion, lets wait with doing things like
this until we have better control over KMS interaction and can use atomic
modesetting. The attached version calls the ioctls in the same way as before,
with only the gbm_bo managing changed.
Anyhow, the new patches are attached, and I went through the ones that were
previously accepted and marked them as such.
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
wayland-bugs mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/wayland-bugs
