On Fri, May 24, 2013 at 11:20:21AM +0200, Daniel Vetter wrote:
> On Wed, May 08, 2013 at 12:55:17PM +0300, ville.syrjala at linux.intel.com wrote:
> > From: Ville Syrj?l? <ville.syrjala at linux.intel.com>
> >
> > Check if the CRTC framebuffer matches the fbdev helper's framebuffer,
> > and if it does, doen't enable cursors/sprites.
> >
> > Signed-off-by: Ville Syrj?l? <ville.syrjala at linux.intel.com>
>
> This one here had conflicts. But it also feels redundant with patch 3,
> since that one should make sure that we won't ever have a cursor/sprite
> stuck around forever.
>
> Of course it's a bit ugly since the cursor might flash shortly, but for
> now I'll brush that off with our lack of atomic modesetting.
>
> Or did I miss something subtly?
My initial idea was to avoid the flashing.
However, now that I think about it a bit more, we do call set_config
from the fb_helper set_par/pan_display hooks, which could then end up
enabling sprites/cursors w/ fbdev. So either we need to keep this patch,
or maybe provide custom set_par/pan_display to turn off
sprites/cursors as well.
> -Daniel
> > ---
> > drivers/gpu/drm/i915/intel_display.c | 15 +++++++++++++++
> > 1 file changed, 15 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
> > index 94d6604..cfe2803 100644
> > --- a/drivers/gpu/drm/i915/intel_display.c
> > +++ b/drivers/gpu/drm/i915/intel_display.c
> > @@ -3264,12 +3264,23 @@ static void ironlake_pfit_enable(struct intel_crtc *crtc)
> > }
> > }
> >
> > +static bool fbdev_active_on_crtc(const struct drm_crtc *crtc)
> > +{
> > + const struct drm_i915_private *dev_priv = crtc->dev->dev_private;
> > +
> > + return dev_priv->fbdev && dev_priv->fbdev->helper.fb == crtc->fb;
> > +}
> > +
> > static void intel_enable_planes(struct drm_crtc *crtc)
> > {
> > struct drm_device *dev = crtc->dev;
> > enum pipe pipe = to_intel_crtc(crtc)->pipe;
> > struct intel_plane *intel_plane;
> >
> > + /* don't enable sprite planes for fbdev */
> > + if (fbdev_active_on_crtc(crtc))
> > + return;
> > +
> > list_for_each_entry(intel_plane, &dev->mode_config.plane_list, base.head)
> > if (intel_plane->pipe == pipe)
> > intel_plane_restore(&intel_plane->base);
> > @@ -6522,6 +6533,10 @@ static void intel_crtc_update_cursor(struct drm_crtc *crtc,
> > u32 base, pos;
> > bool visible;
> >
> > + /* don't enable cursors for fbdev */
> > + if (on && fbdev_active_on_crtc(crtc))
> > + return;
> > +
> > pos = 0;
> >
> > if (on && crtc->enabled && crtc->fb) {
> > --
> > 1.8.1.5
> >
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx at lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
>
> --
> Daniel Vetter
> Software Engineer, Intel Corporation
> +41 (0) 79 365 57 48 - http://blog.ffwll.ch
--
Ville Syrj?l?
Intel OTC
