diff --git a/drivers/gpu/drm/i915/display/intel_atomic.c b/drivers/gpu/drm/i915/display/intel_atomic.c
index ea24a45dab86f52a1bb4fe6b73bbc85e0eae9947..fd0026fc3618d6575ddee5e6994e79b4205d435c 100644
--- a/drivers/gpu/drm/i915/display/intel_atomic.c
+++ b/drivers/gpu/drm/i915/display/intel_atomic.c
@@ -209,6 +209,7 @@ intel_crtc_duplicate_state(struct drm_crtc *crtc)
 	crtc_state->update_wm_pre = false;
 	crtc_state->update_wm_post = false;
 	crtc_state->fifo_changed = false;
+	crtc_state->preload_luts = false;
 	crtc_state->wm.need_postvbl_update = false;
 	crtc_state->fb_bits = 0;
 	crtc_state->update_planes = 0;
diff --git a/drivers/gpu/drm/i915/display/intel_color.c b/drivers/gpu/drm/i915/display/intel_color.c
index 5008a3fca4842fa64542362ff605cb1d6de7f62a..3980e8b50c288999ac7a9bf95a2f49148e688772 100644
--- a/drivers/gpu/drm/i915/display/intel_color.c
+++ b/drivers/gpu/drm/i915/display/intel_color.c
@@ -1022,6 +1022,55 @@ void intel_color_commit(const struct intel_crtc_state *crtc_state)
 	dev_priv->display.color_commit(crtc_state);
 }
 
+static bool intel_can_preload_luts(const struct intel_crtc_state *new_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->uapi.crtc);
+	struct intel_atomic_state *state =
+		to_intel_atomic_state(new_crtc_state->uapi.state);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+
+	return !old_crtc_state->hw.gamma_lut &&
+		!old_crtc_state->hw.degamma_lut;
+}
+
+static bool chv_can_preload_luts(const struct intel_crtc_state *new_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->uapi.crtc);
+	struct intel_atomic_state *state =
+		to_intel_atomic_state(new_crtc_state->uapi.state);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+
+	/*
+	 * CGM_PIPE_MODE is itself single buffered. We'd have to
+	 * somehow split it out from chv_load_luts() if we wanted
+	 * the ability to preload the CGM LUTs/CSC without tearing.
+	 */
+	if (old_crtc_state->cgm_mode || new_crtc_state->cgm_mode)
+		return false;
+
+	return !old_crtc_state->hw.gamma_lut;
+}
+
+static bool glk_can_preload_luts(const struct intel_crtc_state *new_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->uapi.crtc);
+	struct intel_atomic_state *state =
+		to_intel_atomic_state(new_crtc_state->uapi.state);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+
+	/*
+	 * The hardware degamma is active whenever the pipe
+	 * CSC is active. Thus even if the old state has no
+	 * software degamma we need to avoid clobbering the
+	 * linear hardware degamma mid scanout.
+	 */
+	return !old_crtc_state->csc_enable &&
+		!old_crtc_state->hw.gamma_lut;
+}
+
 int intel_color_check(struct intel_crtc_state *crtc_state)
 {
 	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
@@ -1165,6 +1214,8 @@ static int i9xx_color_check(struct intel_crtc_state *crtc_state)
 	if (ret)
 		return ret;
 
+	crtc_state->preload_luts = intel_can_preload_luts(crtc_state);
+
 	return 0;
 }
 
@@ -1217,6 +1268,8 @@ static int chv_color_check(struct intel_crtc_state *crtc_state)
 	if (ret)
 		return ret;
 
+	crtc_state->preload_luts = chv_can_preload_luts(crtc_state);
+
 	return 0;
 }
 
@@ -1271,6 +1324,8 @@ static int ilk_color_check(struct intel_crtc_state *crtc_state)
 	if (ret)
 		return ret;
 
+	crtc_state->preload_luts = intel_can_preload_luts(crtc_state);
+
 	return 0;
 }
 
@@ -1328,6 +1383,8 @@ static int ivb_color_check(struct intel_crtc_state *crtc_state)
 	if (ret)
 		return ret;
 
+	crtc_state->preload_luts = intel_can_preload_luts(crtc_state);
+
 	return 0;
 }
 
@@ -1366,6 +1423,8 @@ static int glk_color_check(struct intel_crtc_state *crtc_state)
 	if (ret)
 		return ret;
 
+	crtc_state->preload_luts = glk_can_preload_luts(crtc_state);
+
 	return 0;
 }
 
@@ -1415,6 +1474,8 @@ static int icl_color_check(struct intel_crtc_state *crtc_state)
 
 	crtc_state->csc_mode = icl_csc_mode(crtc_state);
 
+	crtc_state->preload_luts = intel_can_preload_luts(crtc_state);
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c
index 8e14e5b0507c4e073f854d02d889abff88363934..551de2baa56935efd57cfcb8c5ecfbafccf78fb2 100644
--- a/drivers/gpu/drm/i915/display/intel_display.c
+++ b/drivers/gpu/drm/i915/display/intel_display.c
@@ -14329,6 +14329,11 @@ static void intel_update_crtc(struct intel_crtc *crtc,
 		/* vblanks work again, re-enable pipe CRC. */
 		intel_crtc_enable_pipe_crc(crtc);
 	} else {
+		if (new_crtc_state->preload_luts &&
+		    (new_crtc_state->uapi.color_mgmt_changed ||
+		     new_crtc_state->update_pipe))
+			intel_color_load_luts(new_crtc_state);
+
 		intel_pre_plane_update(old_crtc_state, new_crtc_state);
 
 		if (new_crtc_state->update_pipe)
@@ -14842,6 +14847,7 @@ static void intel_atomic_commit_tail(struct intel_atomic_state *state)
 	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
 		if (new_crtc_state->hw.active &&
 		    !needs_modeset(new_crtc_state) &&
+		    !new_crtc_state->preload_luts &&
 		    (new_crtc_state->uapi.color_mgmt_changed ||
 		     new_crtc_state->update_pipe))
 			intel_color_load_luts(new_crtc_state);
diff --git a/drivers/gpu/drm/i915/display/intel_display_types.h b/drivers/gpu/drm/i915/display/intel_display_types.h
index fadd9853f96679961352f891ff3cdde96b0d9489..83ea04149b7757ea399420ce50837f962180893d 100644
--- a/drivers/gpu/drm/i915/display/intel_display_types.h
+++ b/drivers/gpu/drm/i915/display/intel_display_types.h
@@ -821,6 +821,7 @@ struct intel_crtc_state {
 	bool disable_cxsr;
 	bool update_wm_pre, update_wm_post; /* watermarks are updated */
 	bool fifo_changed; /* FIFO split is changed */
+	bool preload_luts;
 
 	/* Pipe source size (ie. panel fitter input size)
 	 * All planes will be positioned inside this space,