intel_plane_color_copy_uapi_to_hw_state() is meant to copy color
pipeline state from the source plane (from_plane_state) to the
destination plane (plane_state). However, it was incorrectly iterating
the color_pipeline from the destination plane_state instead of the
source from_plane_state.
In non-joiner configurations this doesn't matter since both point to
the same plane. But in bigjoiner mode, the secondary pipe's plane has
no color_pipeline set from userspace (only the primary pipe's plane is
visible to userspace). This causes the while loop to never execute,
leaving the secondary plane's hw.ctm, hw.degamma_lut, and hw.gamma_lut
as NULL.
As a result, xelpd_load_plane_csc_matrix() skips programming the plane
CSC on the secondary joiner pipe (early return due to NULL blob),
causing a visible color split between the left and right halves of the
display when HDR color pipelines are active.
Fix this by reading color_pipeline from from_plane_state (the source/
primary plane) so that the HDR CTM and LUT blobs are properly copied to
the secondary joiner plane's hw state.
Fixes: a78f1b6baf4d ("drm/i915/color: Add framework to program CSC")
Signed-off-by: Vidya Srinivas <[email protected]>
---
drivers/gpu/drm/i915/display/intel_plane.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/i915/display/intel_plane.c
b/drivers/gpu/drm/i915/display/intel_plane.c
index c181a7d063ec..e403fe4a8a20 100644
--- a/drivers/gpu/drm/i915/display/intel_plane.c
+++ b/drivers/gpu/drm/i915/display/intel_plane.c
@@ -396,7 +396,7 @@ intel_plane_color_copy_uapi_to_hw_state(struct
intel_plane_state *plane_state,
bool changed = false;
int i = 0;
- iter_colorop = plane_state->uapi.color_pipeline;
+ iter_colorop = from_plane_state->uapi.color_pipeline;
while (iter_colorop) {
for_each_new_colorop_in_state(state, colorop,
new_colorop_state, i) {
--
2.45.2
