[PATCH] drm/i915: Don't zero out the Y plane's watermarks

* [PATCH] drm/i915: Don't zero out the Y plane's watermarks
@ 2021-03-27  0:59 ` Ville Syrjala
  0 siblings, 0 replies; 7+ messages in thread
From: Ville Syrjala @ 2021-03-27  0:59 UTC (permalink / raw)
  To: intel-gfx; +Cc: stable, Stanislav Lisovskiy

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Don't zero out the watermarks for the Y plane since we've already
computed them when computing the UV plane's watermarks (since the
UV plane always appears before ethe Y plane when iterating through
the planes).

This leads to allocating no DDB for the Y plane since .min_ddb_alloc
also gets zeroed. And that of course leads to underruns when scanning
out planar formats.

We really need to re-enable the pre-merge pixel format tests or else
I'll just keep breaking this stuff...

Cc: stable@vger.kernel.org
Cc: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Fixes: dbf71381d733 ("drm/i915: Nuke intel_atomic_crtc_state_for_each_plane_state() from skl+ wm code")
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/intel_pm.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index b2aede2be89d..49c19acdb7c6 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -5511,12 +5511,12 @@ static int icl_build_plane_wm(struct intel_crtc_state *crtc_state,
 	struct skl_plane_wm *wm = &crtc_state->wm.skl.raw.planes[plane_id];
 	int ret;
 
-	memset(wm, 0, sizeof(*wm));
-
 	/* Watermarks calculated in master */
 	if (plane_state->planar_slave)
 		return 0;
 
+	memset(wm, 0, sizeof(*wm));
+
 	if (plane_state->planar_linked_plane) {
 		const struct drm_framebuffer *fb = plane_state->hw.fb;
 		enum plane_id y_plane_id = plane_state->planar_linked_plane->id;
-- 
2.26.2


^ permalink raw reply related	[flat|nested] 7+ messages in thread