drm/i915: fix panel fitting on LVDS on ILK+ v2
authorJesse Barnes <jbarnes@virtuousgeek.org>
Fri, 3 May 2013 20:26:37 +0000 (13:26 -0700)
committerDaniel Vetter <daniel.vetter@ffwll.ch>
Fri, 10 May 2013 19:56:29 +0000 (21:56 +0200)
This regression was introduced in:

commit b074cec8c652f2d273907a4b35239b4766c894ac
Author: Jesse Barnes <jbarnes@virtuousgeek.org>
Date:   Thu Apr 25 12:55:02 2013 -0700

    drm/i915: move PCH pfit controls into pipe_config

In refactoring this, it was only applied to eDP, which is incorrect.  In
fact, if we ever use the panel fitter to deal with overscan on HDMI,
we'll need to extend it again, so just drop the conditional altogether.

v2: drop check for eDP since we can use the fitter in any config (Daniel)

Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
drivers/gpu/drm/i915/intel_display.c

index d50994a229660f9f38c08e9c9828126f6a838ac9..5491a58e5b3773e16126b09452edcdb90c3c9de6 100644 (file)
@@ -3249,8 +3249,7 @@ static void ironlake_pfit_enable(struct intel_crtc *crtc)
        struct drm_i915_private *dev_priv = dev->dev_private;
        int pipe = crtc->pipe;
 
-       if (crtc->config.pch_pfit.size &&
-           intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_EDP)) {
+       if (crtc->config.pch_pfit.size) {
                /* Force use of hard-coded filter coefficients
                 * as some pre-programmed values are broken,
                 * e.g. x201.