drm/i915: Don't clobber M/N values during fastset check
authorVille Syrjälä <ville.syrjala@linux.intel.com>
Wed, 12 Jun 2019 17:24:23 +0000 (20:24 +0300)
committerVille Syrjälä <ville.syrjala@linux.intel.com>
Tue, 18 Jun 2019 15:28:51 +0000 (18:28 +0300)
We're now calling intel_pipe_config_compare(..., true) uncoditionally
which means we're always going clobber the calculated M/N values with
the old values if the fuzzy M/N check passes. That causes problems
because the fuzzy check allows for a huge difference in the values.

I'm actually tempted to just make the M/N checks exact, but that might
prevent fastboot from kicking in when people want it. So for now let's
overwrite the computed values with the old values only if decide to skip
the modeset.

v2: Copy has_drrs along with M/N M2/N2 values

Cc: stable@vger.kernel.org
Cc: Blubberbub@protonmail.com
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Hans de Goede <hdegoede@redhat.com>
Tested-by: Blubberbub@protonmail.com
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=110782
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=110675
Fixes: d19f958db23c ("drm/i915: Enable fastset for non-boot modesets.")
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190612172423.25231-1-ville.syrjala@linux.intel.com
Reviewed-by: Imre Deak <imre.deak@intel.com>
drivers/gpu/drm/i915/display/intel_display.c

index ad5cccf..853f2a3 100644 (file)
@@ -12300,9 +12300,6 @@ intel_compare_link_m_n(const struct intel_link_m_n *m_n,
                              m2_n2->gmch_m, m2_n2->gmch_n, !adjust) &&
            intel_compare_m_n(m_n->link_m, m_n->link_n,
                              m2_n2->link_m, m2_n2->link_n, !adjust)) {
-               if (adjust)
-                       *m2_n2 = *m_n;
-
                return true;
        }
 
@@ -13434,6 +13431,33 @@ static int calc_watermark_data(struct intel_atomic_state *state)
        return 0;
 }
 
+static void intel_crtc_check_fastset(struct intel_crtc_state *old_crtc_state,
+                                    struct intel_crtc_state *new_crtc_state)
+{
+       struct drm_i915_private *dev_priv =
+               to_i915(new_crtc_state->base.crtc->dev);
+
+       if (!intel_pipe_config_compare(dev_priv, old_crtc_state,
+                                      new_crtc_state, true))
+               return;
+
+       new_crtc_state->base.mode_changed = false;
+       new_crtc_state->update_pipe = true;
+
+       /*
+        * If we're not doing the full modeset we want to
+        * keep the current M/N values as they may be
+        * sufficiently different to the computed values
+        * to cause problems.
+        *
+        * FIXME: should really copy more fuzzy state here
+        */
+       new_crtc_state->fdi_m_n = old_crtc_state->fdi_m_n;
+       new_crtc_state->dp_m_n = old_crtc_state->dp_m_n;
+       new_crtc_state->dp_m2_n2 = old_crtc_state->dp_m2_n2;
+       new_crtc_state->has_drrs = old_crtc_state->has_drrs;
+}
+
 /**
  * intel_atomic_check - validate state object
  * @dev: drm device
@@ -13475,11 +13499,7 @@ static int intel_atomic_check(struct drm_device *dev,
                if (ret)
                        goto fail;
 
-               if (intel_pipe_config_compare(dev_priv, old_crtc_state,
-                                             new_crtc_state, true)) {
-                       new_crtc_state->base.mode_changed = false;
-                       new_crtc_state->update_pipe = true;
-               }
+               intel_crtc_check_fastset(old_crtc_state, new_crtc_state);
 
                if (needs_modeset(&new_crtc_state->base))
                        any_ms = true;